EDITED BY

JAN
FAGERBERG
T H E O X F O R D H A N D B O O K OF

INNOVATION
Edited by

JAN F A G E R B E R G
DAVID C. MOWERY
AND

R I C H A R D R. NELSON

OXFORD
U N IV E R S IT Y PR ESS
 P r e f a c e

In a famous poem, “ The Blind Men and the Elephant,” John Godfrey Saxe (1816-87)
described what may happen when different observers approach the same phenom­
enon from rather di fferent starting points. In the poem Saxe lets one of the blind men
approach the elephant’s side. The man finds it to be “ very like a wall.” Another fits
around its leg and concludes that it resembles a tree. And so on. They end up
disputing “ loud and long.” Saxe drew the following moral:
So oft in theologic wars,
The disputants, I ween,
Rail on in utter ignorance
Of what each others mean,
And prate about an Elephant
Not one of them has seen!

The point is, of course, that each “ disputant” has a valid insight, but needs to
combine it with the insights o f others to reach a holistic understanding. If we
substitute “ innovation” for the elephant and the “ social scientists from different
disciplines” for the blind men, we come close to understanding the motives that led
to the creation o f this handbook. Innovation is a multifaceted phenomenon that
cannot be easily squeezed into a particular branch of the social sciences or the
humanities. Consequently, the rapidly increasing literature on innovation is char­
acterized by a multitude of perspectives based on— or cutting across— existing
disciplines and specializations. There is a danger, however, that scholars studying
innovation do it from starting points so different that they become unable to— or
not interested in— communicating with each other, preventing the development o f a
more complete understanding of the phenomenon.
The purpose o f this volume is to contribute to a holistic understanding of
innovation. The volume includes twenty-one carefully selected and designed con­
tributions, each focusing on a specific aspect o f innovation, as well as an introduc­
tory essay that sets the stage for the chapters that follow. The authors are leading
academic experts on their specific topics, and include economists, geographers,
historians, psychologists, and sociologists. Some contributors have engineering
degrees in addition to their social science degree. Each chapter can be read separately,
but most readers will benefit from reading the introductory essay first. Readers
interested in pursuing further study on specific topics will find suggestions for
¥1 PREFACE

additional reading (m arked with asterisks) in the reference list at the end o f each
chapter.
As with all books there is a history behind it. In fact there are several. There is a
long history, related to h ow innovation studies have evolved over the years. M any o f
the contributions presented here, C hapter 1 in particular, give elem ents o f that story.
The shorter history begins in the m id-1990s w ith the big im petus to innovation
research in Europe provided b y the “ Fram ew ork” program m es o f the European
C om m ission. H aving participated actively in this research for som e tim e, several o f
the contributors to this volum e becam e interested in establishing a netw ork that
could support
X JT
discussion and evaluation o f its results. F o r this xpu rp
x
ose Jan Fager-
K-r

berg organized in 1999, with the support o f the N orw egian Research C ouncil, an
international netw ork for innovation studies that m et occasionally to discuss
selected topics within innovation research. The m eetings o f this group led to a
proposal for a book reflecting o u r current know ledge on innovation. O xford
University Press was contacted and w elcom ed the idea. E con om ic su p port from
the European C om m ission and the N orw egian Research C ou n cil m ade it possible
for the contributors to meet twice to exchange ideas and com m ent on each other
contributions, greatly enhancing the quality and consistency o f the volum e.
One o f the central participants in the netw ork that led to this volu m e was Keith
Pavitt, Professor at SPR U (U niversity o f Sussex) and editor o f Research Policy, the
leading journal in the field. W ith a background in both engineering and econom ics,
Keith was one o f the pioneers in cross-disciplin ary research on innovation. C h arac­
terized by a “ fact-finding” approach and a lack o f respect for received “ grand
theories” not supported by solid evidence, he influenced generations o f younger
researchers and helped put innovation studies on its current “ issue-driven,” em pir­
ically oriented track. Keith enthusiastically supported this b oo k initiative, very
quickly (before anybody else) circulated a full draft o f a chapter and participated
actively in the discussions during the first w orkshop in Lisbon in N o vem ber 2002. He
died unexpectedly shortly afterwards. The editors and contributors dedicate this
book to his m em ory.

J.F., D .M ., R.N .
Oslo, Berkeley; and New York
January 2004
 A c k n o w l e d g e m e n t s

W ithout financial su pport from the N orw egian Research C ouncil (projects
131468/510 and 139867/510) and the European C om m ission (the T E A R I project—
H PSE-C T-2002-60052) this b ook w ould not have been realized. We thank Trygve
Lande and Helge Rynning from N orw egian Research C ouncil and N icholaus Kas~
trinos from the European C om m ission for their cooperation. The Centre for
Technology, Innovation, and Culture (T IK ), U niversity o f Oslo gave Jan Fagerberg
a leave o f absence to start w orking with this project, which he spent at ISEG ,
Technical U niversity o f Lisbon. H e w ould also like to thank ISE G and the G ulben-
kian Foundation for helping to m ake this possible. Sim ilarly, D avid M ow ery would
like to thank the D ivision o f Research o f the H arvard Business School for a Bow er
Fellowship during 20 0 3-4 that aided his w ork on this volum e. M anuel G odinh o o f
ISEG helped organize the first w orkshop in Lisbon in N ovem ber 2002, and Bart
Verspagen o f E C IS (U niversity o f Eindhoven) sim ilarly assisted in organizing the
second w orkshop in R oerm ond in June 2003. In addition to the contributors several
people participated in these w orkshops and contributed to the progress o f the work,
we would particularly like to m ention Fulvio Castellacci, Joao Caraca, M aureen
McKelvey, Sandro M endon^a, R ichard Stankiewicz, and M ona W ibe. D uring the
final phases o f preparing the m anuscript for publication, M ike Hobday, C hris
Freeman, Ian M iles, and Susan Lees provided invaluable assistance in editing,
proofreading, and preparing Keith P a v itfs chapter for publication. D u rin g the
final phase Charles M cC an n provided valuable advice to the non-English/A m erican
authors. O var A ndreas Johansson at T IK w as a very efficient and helpful project
assistant. At O xford U niversity Press D avid M usson and M atthew D erbyshire were
inspiring and patient partners.
 C o n t e n t s

List o f Figures xii

List o f Tables xi ii

List o f Boxes xv

List o f Contributors xvii

1. In n o v a tio n : A G u id e to th e L ite ra tu re i
Jan Fa g er b er g

PARTI IN N O V A T IO N IN T H E M A K IN G

In tr o d u c tio n 28

2. T h e In n o v a tiv e F ir m 29
W illiam Lazonick

3. N e tw o rk s o f In n o v a to r s 56
W a l t e r W. P o w e l l a n d S t i n e G r oda l

4. In n o v a tio n P ro c e sse s 86
Keith Pa v i t t

5. Organizational Innovation 115

A lice Lam

6. M e a s u r in g In n o v a tio n 148
Ke it h Sm it h
X CONTENTS

P A R T II T H E S Y S T E M I C N A T U R E
OF IN N O V A T IO N

In tr o d u c tio n 18 0

7. S y ste m s o f In n o v a tio n : P e rsp e c tiv e s a n d C h a lle n g e s 18 1

C h a r l e s E d q u is t

8. U n iv e rsitie s in N a tio n a l In n o v a tio n S y ste m s 209

D a v i d C. M o w e r y a n d B h a v e n N . S a m p a t

9. F in a n c e a n d In n o v a tio n 240
M a r y O ’S u l l i v a n

10 . In n o v a tio n a n d In te lle c tu a l P r o p e r ty R ig h ts 266

Ove G ranstrand

11 . T h e G e o g r a p h y o f In n o v a tio n : R e g io n a l In n o v a t io n S y ste m s 291

B jorn A s h e i m a n d M e r ic S. G e r t l e r

12 . G lo b a liz a tio n o f In n o v a tio n : T h e R o le o f M u ltin a tio n a l

E n te rp rise s 318
Ra j n e e s h N a r u l a a n d A n t o n e l l o Z a n f e i

PART III H O W I N N O V A T I O N D I F F E R S

Introduction 348

13. Innovation through T im e 349

K r i s t i n e B r u l a n d a n d D a v i d C. M o w e r y

14 . S e c to ra l S y stem s: H o w a n d W h y In n o v a tio n D iffe r s a c ro ss S e c to rs 380

Franco M a ler b a

15. In n o v a tio n in “ L o w -T e c h ” In d u strie s 407

N ick vo n T u n z e l m a n n a n d V i r g i n i a A c h a
 CONTENTS ХІ

1 6. In n o v a tio n in S e rv ic e s 433
Ian M iles

17. In n o v a tio n a n d D iffu s io n 459

B r o n w y n H, H a l l

PART IV INNOVATION AND P ER FOR MA NC E

In tr o d u c tio n 486

18. In n o v a tio n a n d E c o n o m ic G r o w th 487

Ba r t V e r sp a g e n

19. In n o v a tio n a n d C a tc h in g -u p 5 14
J a n F a g e r b e r g a n d M a n u e l M . G od in ho

20. In n o v a tio n a n d C o m p e titiv e n e s s 543

J ohn C a n t w e l l

21. In n o v a tio n a n d E m p lo y m e n t 568

M a r io P ia n t a

22. S c ie n c e , T e c h n o lo g y , a n d In n o v a tio n P o lic y 599

B e n g t - a k e L u n d v a l l a n d S u s a n a B orras
 L i s t of F i g u r e s

1.1 Scholarly articles w ith “ In n o v a tio n :' in the title, 19 55-2004 2

3.1 Strong and weak ties 61
3.2 Structural holes and redundant ties 62
3.3 N etw ork typology 64
3.4 Knowledge codification and innovation 76
6.1 U SPTO utility patents 1965-2002 *59
6.2 The SPR U innovation database: The intersectoral flow o f
innovations t^3
6.3 D efining technological innovation— C o m m u n ity In n ovation
Survey (CIS) 164
6.4 Resources devoted to innovation activities in 1996 166
8.1 Universities' perform ance share o f total national R & D , 19 8 1-9 9 216
8*2 R8cD perform ed by the higher education sector as a percentage
o f GDP 217
8.3 Share o f higher education R & D financed b y industry, 19 9 1-2 0 0 0 218
8 4 U S research university patents as a percentage o f all
dom estic-assignee U S patents, 1963-99 230
8.5 Technology field o f U S “ research university" patents, 19 6 0-9 9 231
17.1 D iffusion o f electric m otors in U S m anufacturing 467
17.2 D iffusion o f m ajor innovations in the U nited States 468
18.1 Long-run growth in the w orld econom y 488
18.2 A pproxim ate chronology o f technological revolutions 498
18.3 The Jones critique 507
19.1 G D P per capita growth 19 6 0 -20 0 1 527
19.2 Third-level enrollm ent in relation to age group, 2 0 -2 4 years old
(1965-95) 528
19.3 Ratio o f first university degrees in natural sciences and engineering
to 24-year-olds in the population, 1999 (all values in % ) 529
19 4 R8cD as percentage o f GDP, 1960s and 1990s 531
19.5 U S patenting per m illion inhabitants (log scale) 532
21.1 Share o f new products in sales and em ploym ent change 574
22.1 Relationship between science, technology, and innovation policy 615
22.2 The scientific and technological architecture o f Europe, 2001 622
 L i s t of T a b l e s

4.1 Exam ples o f technological convergence and vertical disintegration 92

4.2 M atching corporate technology and organizational practices
with m arket needs and dem ands 103
5л M intzberg's structural archetypes and their innovative potentials 120
6.1 G E R D /G D P ratios across countries 156
6.2 Classification o f industries based on R & D intensity 157
6.3 A ppendix 6л: Recent (2002 onwards) jou rnal publications using
C IS data 170
6.4 A ppendix 6*2: Publications using C IS data sponsored by the
European C o m m issio n — topic and institutions 171
8л The relevance o f university science to industrial technology 223
8.2 Im portance to industrial R & D o f sources o f inform ation on
public R & D (including university research) 224
9л Venture capital and early-stage investm ent as a percentage o f G D P 253
10л Eras in the h istory o f patents and IP 268
xo*2 C hronological overview o f m ajor events in U S post-w ar IP R
developm ent (through 2000) 276
10.3 M eans for com m ercializing new product technologies 281
10.4 Sensitivity o f the R & D investm ents o f large Japanese corporations
to length o f term (1992) 282
12л A taxon om y o f the globalization o f innovation 320
12.2 Rates o f grow th o f industrial R & D and patenting in the O E C D
countries 321
12.3 Selected indicators o f FD I and international production,
19 8 2-20 0 1 (U S $b illion at current prices and percentage values) 323
12.4 R & D expenditure o f foreign affiliates as a percentage o f
total R & D expenditures b y all firm s in selected host econom ies,
1998 o r latest year 324
12.5 Shares o f U S patenting o f largest nationally owned industrial
firm s due to research located abroad, 19 20 -9 0 324
12.6 Share o f U S patents o f the w o rld s largest firm s attributable to
research in foreign locations b y m ain area o f origin o f parent
firm s, 1969-95 325
XIV LIST OF TABLES

15.1 M arket shares, 1970-93 (ratio o f national exports to w orld

exports, % ) 412
16л Share (% ) o f gross value added in services in total G D P in E U
countries 434
17л D iffusion in Japanese households (% ) 469
19л Incom e groups, 1960-99 (G D P per capita, 10 3 $U S , 1990
constant PPPs) 526
20л The decom position o f the predicted grow th in national m arket
shares from an estim ated em pirical m odel o f cross-cou n try
com petitiveness, for 19 6 1-73 548
20.2 Actual grow th rates achieved b y countries, 19 6 0-89 , over and
above that predicted by (inter alia) their rates o f capital
accum ulation 549
20.3 Educational enrollm ents in technical subjects at tertiary level as
a percentage o f the total population in selected countries,
in 1995 or closest year available 554
20.4 The industries in which the largest nationally ow ned firm s have
persistently held com parative advantage in innovation,
1920-39 and 197&-95 556
20.5 The shares o f patenting o f the largest industrial firm s attributable
to research facilities located in the biggest single region o f
selected European countries, in 1969-95 558
20.6 The statistically significant determ inants o f com parative
growth am ong the w orld's largest industrial firm s, 19 72-8 2 560
21л A sum m ary o f approaches to innovation and em ploym ent 570
21.2 Effects o f innovation on the quantity o f em ploym ent 577
21.3 Effects o f innovation on the quality o f em ploym ent: skills,
wages, and organizations 585
 L i s t of B o x e s

1.1 SPRU, Freem an, and the spread o f innovation studies 3

1.2 The innovation theorist Joseph Schum peter 6
1.3 W hat innovation is not: the linear m odel 8
1 4 W hat is high-tech? Pavitt s taxonom y 16
1.5 Industrial dynam ics— an evolutionary interpretation 17
5.1 Burns and Stalker: m echanistic and organic structures 118
5.2 Japan: an exam ple o f organizational com m unity m odel o f learning 129
5.3 Silicon Valley: an exam ple o f professional team m odel o f learning 130
5 4 Oticon: the rise and decline o f the “ spaghetti o rg an izatio n 5 131
6.1 Bibliom etric data 153
7.1 Systems o f innovation— m ain term s used 182
9.1 Venture capital in com parative-h istorical perspective 251
10.1 International IP conventions 270
10.2 Trade-related aspects o f Intellectual Property Rights (T R IP S) 275
11.1 Baden-W iirttem berg s regionally networked innovation system 301
11.2 US biotechnology clusters: entrepreneurial regional innovation
systems 304
12.1 Asset exploiting, asset augm enting or both? 329
12.2 H ow innovation system s affect the internationalization
of R&D 332
12.3 Host coun try effects: technology gaps, technological
upgrading, and absorptive capacity 338
13.1 Technological diffusion in the first industrial revolution 353
13.2 Josiah W edgw ood and “ m odern” m anagem ent in pottery
fabrication 357
13.3 The foundation o f R8cD laboratories by Bayer and D u Pont 362
13 4 The Internet 370
15.1 Planetary science in the N orth Sea 410
15.2 W hat's so clever about a rubber tube? 413
17.1 The Q W E R T Y controversy— d iffusion w ith netw ork externalities 463
18.1 Technological change and uncertainty 494
18.2 Evolution and the blind w atchm aker 495
18.3 The “ Jones critique” and sem i-endogenous grow th m odels 506
19.1 O rganizational in n ovation in Jap an 520
XVI LIST OF BOXES

19.2 The financial crisis in Korea 522

19.3 H ow to access foreign technology? The O EM system 533
19 4 A tale o f two countries 537
21.1 Technology and unem ploym ent: a classical debate 371
21.2 Evidence on innovation and em ploym ent 573
22.x O EC D and the evolving discourse around science,
technology and innovations policy 603
22.2 The neoclassical econom ics o f innovation po licy 613
22.3 Innovation systems and innovation policy 617
2 2 4 N orm ative principles for design o f STI policies 618
 L i s t of C o n t r i b u t o r s

Virginia Acha Research Fellow, SPRU , U niversity o f Sussex, U K .

Bjorn Asheim Professor, Departm ent o f Social and Econom ic G eography and
Centre for Innovation, Research and Com petence in the Learning E conom y
(C IR C LE), U niversity o f Lund, Sweden, and Centre for Technology, Innovation
and Culture (T IK ), U niversity o f Oslo, Norway.

Susana B o rras Associate Professor, D epartm ent o f Social Sciences, Roskilde U n i­

versity, D enm ark.

Kristine B ru lan d Professor, D epartm ent o f H istory, U niversity o f Oslo, Norway.

John Cantw ell Professor, Rutgers University, U SA and U niversity o f Reading, U K.

Charles Edquist Professor, D ivision o f Innovation, D epartm ent o f Design, Lund

Institute o f Technology, Lund University, Sweden and Centre for Innovation, R e­
search and Com petence in the Learning E con om y (C IR C L E ), Lund University,
Sweden.

Jan Fagerberg Professor, Centre for Technology, Innovation and Culture (T IK ),

University o f Oslo, N orway.

Meric G ertler Professor, D epartm ent o f G eography and M u n k Centre for Inter­
national Studies, U niversity o f Toronto, Canada, and Centre for Technology, In n o v­
ation and Culture (T IK ), U niversity o f Oslo, N orway.

M anuel M . G o d in h o Associate Professor, ISE G , U niversidade Tecnica de Lisboa,

Portugal.

Ove G ran stran d Professor, Center for Intellectual Prop erty Studies (C IP ), D epart­
ment o f Industrial M anagem ent and Econom ics, School o f Technology M anage­
ment and Econom ics, C halm ers U niversity o f Technology, Sweden.

Stine G rod al D octoral Candidate in M anagem ent Science and Engineering, Stan­
ford University, U SA .

B ron w yn H all Professor, D epartm ent o f Econom ics, U niversity o f C aliforn ia at

Berkeley, U SA .

Alice L am Professor, School o f Business and M anagem ent, Brunei U niversity, U K .

ХѴІІІ LIST OF CONTRI BUTORS

w m ia m Lazonick U niversity Professor, U niversity o f M assachusetts Lowell, U SA

and D istinguished Research Professor, IN SE A D , France.

Bengt-A ke LundvaU Professor, D epartm ent o f Business Studies, A alb o rg Univer-

sity, Denm ark.

Franco M alerba Professor, C E S P R I and Istituto di E con om ia Politica, Bocconi

University, Italy.

Ian M iles Professor, PREST, Institute o f In novation Research, U n iversity o f M an ­

chester, U K.

D avid C M ow ery Professor, Haas School o f Business, U n iversity o f C aliforn ia at

Berkeley, USA.

Rajneesh N arula Professor, D epartm ent o f International E con om ics 8c M anage­

ment, Copenhagen Business School, D enm ark and Centre for Technology, In nov­
ation and Culture (T IK ), U niversity o f Oslo, N orw ay.

R ichard R. N elson Professor, C olu m b ia University, U SA .

M a ry O ’Sullivan Associate Professor, Strategy and M anagem ent, IN SE A D , France.

Keith Pavitt Professor, SPRU , U niversity o f Sussex, U K .

M ario Pianta Professor, Faculty o f Econom ics, U n iversity o f U rbin o, Italy.

W alter W. Powell Professor o f Education, Sociology, and O rganizational Behavior

at Stanford University, USA.

Bhaven N. Sam pat Assistant Professor, School o f Public Policy, G eorgia Institute o f
Technology, USA.

Keith Sm ith Professor, D epartm ent o f Industrial D ynam ics, C h alm ers U niversity
o f Technology, Sweden.

N ick von Tunzelm ann Professor, SPRU , U niversity o f Sussex, U K .

B art Verspagen Professor, Eindhoven Centre for In novation Studies (Ecis), Eind ­
hoven U niversity o f Technology, the N etherlands, and Centre for Technology,
Innovation and Culture (T IK ), U niversity o f Oslo, N orw ay.

A ntonello Zan fei Professor, Faculty o f Econom ics, U niversity o f U rbino, Italy.
 C H A P T E R 1

IN N O V A T I O N
A GUIDE TO THE
* ль* Ж. « Ж Ж . л Ж* Ж*. .Ж. JL Ж J

LITERATURE

JAN FAGERBERG

i.i In t r o d u c t io n 1

In n o v a t io n is not a new phenom enon. Arguably, it is as old as m ankind itself.

There seems to be som ething inherently “ hum an” about the tendency to think about
new and better ways o f d oing things and to try them out in practice. W ithout it, the
world in w hich we live w ould lo o k very, very different. Try for a m om ent to think o f a
world w ithout airplanes, autom obiles, telecom m unications, and refrigerators, just
to m ention a few o f the m ore im portant innovations from the not-too-distant past.
Or— from an even longer perspective— where w ould we be w ithout such fu nda­
mental innovations as agriculture, the wheel, the alphabet, o r printing?
In spite o f its obviou s im portance, innovation has not always received the
scholarly attention it deserves. For instance, students o f long-ru n econom ic change
used to focus o n factors such as capital accum ulation or the w orking o f m arkets,
rather than on innovation. T his is now changing. Research on the role o f innovation
in econom ic and social change has proliferated in recent years, particularly w ithin
the social sciences, and w ith a bent tow ards cross-disciplinarity. In fact, as illustrated
in Figure 1.1, in recent years the num ber o f social-science publications focusing on
innovation has increased m uch faster than the total num ber o f such publications.
2 IAN F A G E R B E R G

Fig. 1.1 Scholarly Articles with "Innovation" in the title, 1955-2004

(per 10,000 social science articles)
Note: The source is the IS! Web of Knowledge, Social Sciences Citation Index (SSCI).

As a result, our knowledge about innovation processes, their determ inants and social
and econom ic im pact has been greatly enhanced.
W hen innovation studies started to em erge as a separate field o f research in the
1960s, it did so m ostly outside the existing disciplines and the m ost prestigious
universities. An im portant event in this process w as the form ation in 1965 o f the
Science Policy Research Unit (SPR U ) at the U niversity o f Sussex (see B o x 1.1). The
nam e of the center illustrates the tendency for innovation studies to develop unde
other (at the time m ore acceptable?) terms, such as, for instance, “ science studies” or
“ science policy st udies.” But as we shall see in the follow ing, one o f the m ain lessons
from the research that came to be carried out is that science is only one am ong several
ingredients in successful innovation. As a consequence o f these findings, not only the
focus o f research in this area but also the notions used to characterize it changed.
D uring the late twentieth/early twenty-first century, a n um ber o f new research
centers and departm ents have been founded, focusing on the role o f innovation in
 innovation : a g u i d e to t h e l it e r a t u r e 3

Box 1.1 SPRU, Freeman, and the spread of innovation studies

SPRU— Science Policy Research Unit— at the University o f Sussex, UK was founded in
1965 with Christopher Freeman as its first director. From the beginning, it had a cross-
disciplinary research staff consisting o f researchers with backgrounds in subjects as
diverse as economics, sociology, psychology, and engineering. SPRU developed its own
cross-disciplinary Master and Ph.D. programs and carried out externally funded
research, much o f which came to focus on the role o f innovation in economic and
social change. It attracted a large number o f young scholars from other countries who
came to train and work here.
The research initiated at SPRU led to a large number o f projects, conferences, and
publications. Research Policy* which came to be the central academic journal in the
field, was established in 1972, with Freeman as the first editor (he was later succeeded by
Keith Pavitt, also from SPRU). Freemans influential book, The Economics o f Industrial
Innovation* was published two years later, in 1974, and has since been revised twice. In
1982, the book, Unemployment and Technical Innovation* written by Freeman, Clark,
and Soete, appeared, introducing a systems approach to the role o f innovation in long­
term economic and social change. Freeman later followed this up with an analysis o f
the national innovation system in Japan (Freeman 1987). He was also instrumental in
setting up the large, collaborative IFIAS project which in 1988 resulted in the very
influential book, Technical Change and Economic Theory* edited by Dosi, Freeman,
Nelson, Silverberg, and Soete (both Dosi and Soete were SPRU Ph.D. graduates).
In many ways, SPRU came to serve as a role model for the many centers/institutes
within Europe and Asia that were established, mostly from the m id- 1980s onwards,
combining cross-disciplinary graduate and Ph.D. teaching with extensive externally
funded research. Most o f these, as SPRU itself, were located in relatively newly formed
(so-called “ red-brick” ) universities, which arguably showed a greater receptivity to
new social needs, initiatives, and ideas than the more inert, well-established academic
“ leaders,” or at other types o f institutions such as business or engineering schools.
SPRU graduates were in many cases instrumental in spreading research and teaching
on innovation to their own countries, particularly in Europe.

econom ic and so cial change. M a n y o f these have a cro ss-d iscip lin ary o rien tatio n ,
illustrating the need fo r in n o va tio n to be studied fro m different perspectives. Several
jou rn als an d p ro fessio n al association s have also been fo u n d ed .
The leaning tow ards cross-d isciplin arity that characterizes m uch scholarly w o rk in
this area reflects the fact that no single discipline deals w ith all aspects o f innovation.
Hence, to get a com preh ensive overview , it is necessary to com bine insights from
several disciplines. Traditionally, fo r instance, econom ics has dealt primarily w ith the
allocation o f resources to in n ovation (in com petition with other ends) and its
econom ic effects, w hile the in n o vation process itself has been m o re or less treated as
a “ black b o x ” W h at h ap pen s w ithin this “box” has been left to scholars from other
disciplines. A lo t o f w h at h appens o b vio u sly has to d o w ith learning, a central topic in
cognitive science. Such, learn in g occu rs in organized settings (e g . groups* teams, firms.
4 IAN F A G E R B E R G

and networks), the working o f which is studied w ithin disciplines such as sociology,
organizational science, management, and business studies. M oreover, as economic
geographers point out, learning processes tend to be linked to specific contexts or
locations. The way innovation is organized and its localization also undergo im port­
ant changes through time, as underscored by the w o rk w ithin the field o f economic
history. There is also, as historians o f technology have pointed out, a specific techno­
logical dim ension to this; the way innovation is organized, as well as its econom ic and
social effects, depends critically on the specific nature o f the technology in question.
Two decades ago, it was still possible for a h ard -w orkin g student to get a fairly
good overview o f the scholarly w ork on innovation b y devoting a few years o f
intensive study to the subject. N ot any m ore. Today, the literature on innovation is
so large and diverse that even keeping up-to-date w ith one specific field o f research is
very challenging. The purpose o f this volum e is to provide the reader w ith a guide to
this rapidly expanding literature. We do this under the follow in g b road headings:
I Innovation in the M aking
II The System ic N ature o f Innovation
III H ow Innovation Differs
IV Innovation and Perform ance.
Part One focuses on the process through which innovations occu r and the actors
that take part: individuals, firm s, organizations, and netw orks. A s we will discuss in
m ore detail below, innovation is by its very nature a system ic phenom enon, since it
results from continuing interaction between different actors and organizations. Part
Two outlines the systems perspective on innovation studies and discusses the roles o f
institutions, organizations, and actors in this process at the national and regional
level. Part Three explores the diversity in the m anner in w hich such system s work
over time and across different sectors or industries. Finally, Part Fo u r exam ines the
broader social and econom ic consequences o f innovation and the associated policy
issues. The rem ainder o f this chapter sets the stage for the discussion that follows by
giving a broad overview o f some o f the central topics in innovation studies (includ­
ing conceptual issues).

1 .2 W h a t is In n o v a t io n ?

A n im portant distinction is norm ally m ade between invention and innovation.2

Invention is the first occurrence o f an idea for a new produ ct or process, while
innovation is the first attempt to carry it out into practice. Som etim es, invention
and innovation are closely linked, to the extent that it is hard to distinguish one from
 innovation : a g u i d e to t h e l i t e r a t u r e 5

another (b io tech n o lo g y for instance). In m an y cases, how ever, there is a consider­

able tim e lag betw een the two. In fact, a lag o f several decades or more is not
u n co m m o n (R ogers 1995). Such lags reflect the d ifferent requirem ents for working
out ideas and im p lem en tin g them . W h ile invention s m ay be carried out anywhere,
for exam ple in universities, in n o vation s occu r m o stly in firm s, though they may also
occu r in oth er types o f organizations, such as pu b lic h ospitals. To be able to turn an
invention into an in n o vatio n , a firm n o rm ally needs to co m b in e several different
types o f know ledge, capabilities, skills, and resources. F o r instance, the firm may
require p ro d u ctio n know ledge, skills and facilities, m arket know ledge, a well­
fu n ctio n in g d istrib u tio n system , sufficient finan cial resources, an d so on. It follows
that the role o f the in n o vato r,3 i.e. the perso n o r organ izatio n al unit responsible for
co m b in in g the factors necessary (w hat the in n o vatio n theorist Joseph Sch u m p eter
(see B o x 1.2) called the “ en trepren eu r” ), m ay be quite d ifferent fro m that o f the
inventor. In deed, h isto ry is replete w ith cases in w h ich the in ven to r o f m a jo r
tech nological advances fails to reap the profits from his b reakth rou gh s.
L o n g lags betw een in ven tion and in n o vatio n m ay have to do w ith the fact that, in
m an y cases, som e o r all o f the co n d itio n s fo r co m m ercializatio n m a y be lacking.
There m ay n o t be a su fficien t need (yet!) o r it m ay be im p o ssib le to p ro d u ce an d /or
m arket because som e vital in pu ts o r co m p lem en tary factors are n o t (yet!) available.
T h us, alth o u gh L eo n ard o da V in ci is rep orted to have h ad som e q u ite advanced
ideas for a flyin g m ach in e, these w ere im possib le to c a rry o u t in practice due to a lack
o f adequate m aterials, p ro d u ctio n skills, and— above all— a p o w er source. In fact,
the realization o f these ideas had to w ait fo r the invention and subsequent co m m er­
cialization (and im p ro vem en t) o f the internal co m b u stio n engine.4 H ence, as this
exam ple show s, m an y in ven tion s requ ire co m p lem en tary Inventions and in n o v ­
ations to succeed at the in n o vatio n stage.
A n o th er co m p licatin g factor is that in ven tion and in n o va tio n is a co n tin u o u s
process. F o r instance, the car, as w e k n o w it today, is rad ically im p ro ved co m p ared to
the first co m m ercial m odels, due to the in co rp o ra tio n o f a v e ry large n u m b er o f
different in ven tion s/in n ovatio n s. In fact, the first version s o f v irtu a lly all sign ifican t
in n ovation s, fro m the steam engine to the airplan e, w ere crude, un reliab le version s
o f the devices that even tu ally d iffu sed w idely. K lin e and R o sen b erg (1986), in an
influential paper, p o in t out:

it is a serious mistake to treat an innovation as if it were a well-defined, homogenous thing

that could be identified as entering the economy at a precise date— or becoming available at a
precise point in tim e.. . . The fact is that most important innovations go through drastic
changes in their lifetimes— changes that may, and often do, totally transform their economic
significance. The subsequent improvements in an invention after its first introduction m aybe
vastly more important, economically, than the initial availability o f the invention in its
original form. (Kline and Rosenberg 1986: 283)

T h u s, w h at w e th in k o f as a single in n o va tio n is the result o f a lengthy

o ften
process involving- m a n y interrelated in n o vatio n s. This is one o f the reasons why
6 JAN F A G E R B E R G

Box 1.2 The innovation theorist Joseph Schumpeter

Joseph Schumpeter ( 1883- 1950) was one o f the most original social scientists o f A&
twentieth century* He grew up in Vienna around the turn o f the century* where he
studied law and economics. For most o f his life he worked as an academic* but he also
tried his luck as politician* serving briefly as finance minister in the first post-World
War I (socialist) government* and as a banker (without much success). He became
professor at the University o f Bonn in 1925 and later at Harvard University in the USA
( 1932), where he stayed until his death. He published several books and papers in
German early on, among these the Theory o f Economic Development, published in 1911
and in a revised edition in English in 1934*Among his most well-known later works are
Business Cycles in two volumes (from 1939) > Capitalism , Socialism and Democracy
(1943), and the posthumously published History o f Economic Analysis (1954)*
Very early he developed an original approach, focusing on the role o f innovation in
economic and social change. It was not sufficient, Schumpeter argued, to study the
economy through static lenses, focusing on the distribution o f given resources across
different ends. Economic development, in his view, had to be seen as a process of
qualitative change, driven by innovation, taking place in historical time. As examples
o f innovation he mentioned new products, new methods o f production, new sources
of supply, the exploitation of new markets, and new ways to organize business. He
defined innovation as “ new combinations” o f existing resources. This combinatory
activity he labeled “ the entrepreneurial function” (to be fulfilled by “ entrepreneurs” ),
to which he attached much importance. One main reason for the important role played
by entrepreneurs for successful innovation was the prevalence o f inertia, or “ resistance
to new ways” as he phrased it, at all levels of society that entrepreneurs had to fight in
order to succeed in their aims. In his early work, which is sometimes called “ Schump­
eter Mark I ” Schumpeter focused mostly on individual entrepreneurs. But in later
works he also emphasized the importance o f innovation in large firms (so-called
“ Schumpeter Mark II” ), and pointed to historically oriented, qualitative research
(case studies) as the way forward for research in this area.
In his analysis of innovation diffusion, Schumpeter emphasized the tendency for
innovations to “cluster” in certain industries and time periods (and the derived effects
on growth) and the possible contribution o f such “clustering” to the formation of
business cycles and long waves” in the world economy (Schumpeter 1939). The latter
suggestion has been a constant source of controversy ever since. No less controversial,
and perhaps even better known, is his Inspired discussion o f the institutional changes
under capitalism (and its possible endogenous transformation into “ socialism” ) in the
book Capitalism, Socialism and Democracy (1943).
Sources. Swedberg 1991; Shionoya 1997; Fagerberg 2003.

m any students o f technology and innovation find it natural to app ly a systems

perspective rather than to focus exclusively on individual inventions/innovations.
Innovations may also be classified according to “ type,” Schum peter (see Box 1.2)
distinguished between five different types; new products, new m ethods o f produc­
tion, new sources o f supply, the exploitation o f new m arkets, and new ways to
 innovation : a g u i d e to t h e l i t e r a t u r e 7

organize business, H ow ever, in econom ics, m ost o f the focus has been on the tw o
first o f these. S ch m o o k ler (1966), fo r instance, in his classic work on “ Invention and
E con om ic G row th ,” argu ed that the distinction betw een “ product technology” and
“ p ro d u ctio n tech n o lo g y” w as “ critical” for o u r u n d erstan d in g o f this phenomenon
(ibid. 166). H e defin ed the fo rm er type as know ledge abou t h o w to create or improve
products, and the latter as know ledge ab ou t h ow to produce them . Similarly, the
term s “ p ro d u ct in n o va tio n ” and “ process in n o va tio n ” have been used to character-
ize the occurren ce o f new o r im p ro ved good s an d services, and im provem ents in the
w ays to p ro d u ce these g o o d and services, respectively,5 T h e argu m en t for focusing
particu larly on the d istin ctio n betw een p ro d u ct and process in n o vatio n often rests
on the assu m p tio n that their eco n o m ic and social im pact m a y differ. For instance,
while the in tro d u ctio n o f new p ro d u cts is c o m m o n ly assum ed to have a clear,
positive effect on grow th o f in com e and em plo ym en t, it has been argu ed that process
in n ovation , due to its co st-cu ttin g nature, m ay have a m o re am b ig u o u s effect
(Edquist et al. 2001; P ian ta in this vo lu m e). H ow ever, w h ile clearly d istin guish able
at the level o f the in d ivid u al firm or indu stry, such differences tend to becom e
blurred at the level o f the overall econom y, because the p ro d u ct o f o n e firm (or
industry) m ay end up as b ein g used to p ro d u ce good s o r services in another.6
The focus o n p ro d u ct and process in n o vation s, w h ile u sefu l fo r the analysis o f
som e issues, sh o u ld not lead us ign ore other im p o rtan t aspects o f in n o vatio n . F o r
instance, d u rin g the first h a lf o f the tw entieth century, m an y o f the in n o vatio n s that
m ade it possible fo r the U n ited States to “ forge ah ead ” o f other capitalist econ o m ies
were o f the o rgan izatio n al k in d , in vo lvin g entirely new w ays to organ ize p ro d u ctio n
and d istrib u tion (see B ru la n d an d M o w e ry in this vo lu m e, w hile L a m pro vid es an
overview o f o rgan izatio n al in n o vatio n ). E d q u ist et al. (2001) have suggested d iv id ­
ing the catego ry o f process in n o va tio n into “ tech n ological process in n o va tio n s” and
“ organizational p ro cess in n o vation s,” the fo rm er related to n ew types o f m ach in ery,
and the latter to n ew w ays to organ ize w o rk . H ow ever, organ izatio n al in n o vatio n s
are not lim ited to n ew w ays to organize the process o f p ro d u ctio n w ith in a given
firm . O rgan ization al in n o va tio n , in the sense used b y Sch u m p eter ,7 also includes
arrangem ents across firm s su ch as the reo rgan izatio n o f entire indu stries. M oreover,
as exem plified b y the case o f the U S A in the first h a lf o f the p revio u s century, m an y o f
the m ost im p o rta n t o rgan izatio n al in n o vatio n s have occu rred in d istrib u tio n , w ith
great consequences fo r a w h o le ran ge o f indu stries (C h an d ler 1990)*
A n oth er ap p ro ach , also based o n S c h u m p e te rs w o rk , has been to classify innov­
ations acco rd in g to h o w rad ical th ey are co m p ared to cu rren t technology (Freeman
and Soete 1997). F ro m this perspective, co n tin u o u s im p ro vem en ts o f the type
referred to above are o ften characterized as “ in crem en tal” or “ marginal” innov­
ations,8 as o p p o se d to “ rad ical” in n o vatio n s (such as the introduction o f a totally
new type o f m ach in ery) o r “ tech n o logical rev o lu tio n s” (con sistin g o f a cluster o f
in n o vation s that togeth er m a y have a v e ry far-reaching impact)* Schumpeter
focused in p a rtic u la r o n the latter tw o categories, which he believed to be o f greater
8 JAN F A G E R B E R G

im portance. It is a w idely held view, however, that the cum ulative im pact of
increm ental innovations is just as great ( if not greater), and that to ignore these
leads to a biased view o f long run econom ic and social change (Lun d vail et aL 1992).
M oreover, the realization o f the econom ic benefits fro m radical innovations in
m ost cases (including those o f the airplane and the autom obile, discussed earlier)
requires a series o f incremental im provem ents. A rguably, the b u lk o f economic
benefits com e from incremental innovations and im provem ents.
There is also the question o f how to take different contexts into account. I f A for
the first time introduces a particular innovation in one context, w hile B later
introduces the sam e innovation in another, w ou ld we characterize both as innov­
ators? This is a m atter o f convention. A w idely used practice, based on Schum peter’s
work, is to reserve the term innovator for A and characterize B as an im itator. But
one might argue that, follow ing Schum peter’s ow n d efin ition, it w ou ld be equally
consistent to call B an Innovator as well, since B is in trodu cin g the innovation for the
first tim e in a new context. This is, for instance, the position taken b y H obday (2000)
in a discussion o f innovation in the so-called “ new ly indu strializing countries” in
A sia.9 One might object, though, that there is a qualitative difference between (a)
com m ercializing som ething for the first tim e and ( b) copyin g it and introducing it in
a different context. The latter arguably includes a larger dose o f im itative behavior
(im itation), or what is som etim es called “ technology transfer.” T h is does not
exclude the possibility that im itation m ay lead to new in n o vation (s). In fact, as
pointed out by Kline and Rosenberg (1986, see B o x 1.3), m any econom ically sig­
nificant innovations occur while a product or process is diffusing (see also Hall in
this volum e). Introducing som ething in a new context often im plies considerable
adaptation (and, hence, increm ental innovation) and, as h isto ry has show n, organ­
izational changes (or innovations) that m ay significantly increase productivity and
competitiveness (see G odinho and Fagerberg in this v o lu m e ).10

Box 1-3 What innovation is not: the linear model

Sometimes it easier to characterize a complex phenomenon by clearly pointing out

what it is NOT. Stephen Kline and Nathan Rosenberg did exactly this when they, in an
influential paper from 1986, used the concept “ the linear model” to characterize a
widespread but in their view erroneous interpretation o f innovation.
Basically, the linear model” is based on the assumption that innovation is applied
science. It Is linear because there is a well-defined set o f stages that innovations are
assumed to go through. Research (science) comes first, then development, and finally
production and marketing. Since research comes first, it is easy to think o f this as the
critical element. Hence, this perspective, which is often associated with Vannevar
Bushs programmatic statements on the organization o f the US research systems
(Bush 1945), is well suited to defend the interests o f researchers and scientists and
the organizations in which they work.
 innovation : a g u i d e to t h e l i t e r a t u r e 9

The problems with this model, Kline and Rosenberg point out, are twofold. First,
it generalizes a chain of causation that only holds for a minority o f innovations.
Although some important innovations stem from scientific breakthroughs, this is
not true most o f the time. Firms normally innovate because they believe there is a
commercial need for it, and they commonly start by reviewing and combining existing
knowledge. It is only if this does not work, they argue, that firms consider investing in
research (science). In fact, in many settings, the experience o f users, not science, is
deemed to be the most important source o f innovation (von Hippel 1988; Lundvall
1988). Second, “ the linear model” ignores the many feedbacks and loops that occur
between the different “ stages” o f the process. Shortcomings and failures that occur at
various stages may lead to a reconsideration o f earlier steps, and this may eventually
lead to totally new innovations.

1 .3 In n o v a t io n in t h e M a k in g

Leaving d efin itio n s aside, the fu n d am en tal question fo r in n o vatio n research is o f

course to exp lain h o w in n o vatio n s occur. O ne o f the reasons in n o vatio n w as ign ored
in m ain stream so cial science fo r so lo n g w as that this w as seen as im possib le to do.
The best on e co u ld do, it w as c o m m o n ly assum ed, w as to lo o k at in n o va tio n as a
ran d o m p h en o m en o n (o r “ m an n a fro m heaven ” as som e scholars used to ph rase it).
Schum peter, in his early w ork s, w as one o f the first to object to this practice. H is ow n
account o f these processes em phasized three m ain aspects. T h e first w as the fu n d a ­
m ental u n certain ty inherent in all in n o vatio n projects; the secon d w as the need to
m ove q u ick ly b efo re so m e b o d y else d id (and reap the poten tial eco n o m ic rew ard).
In practice, Sch u m p eter argu ed, these tw o aspects m eant that the stand ard b eh a v ­
ioral rules, e.g., su rveyin g all in fo rm atio n , assessing it, an d fin d in g the “ o p tim a l”
choice, w o u ld not w o rk . O ther, q u icker w ays had to be fo u n d . T h is in h is v ie w
involved lead ersh ip an d v isio n , tw o qualities he associated w ith entrepren eursh ip.
The th ird aspect o f the in n o va tio n process w as the prevalence o f “ resistance to new
w ays” — o r in ertia— at all levels o f society, w h ich threatened to d estro y all n o vel
initiatives, an d fo rced en trepren eu rs to figh t h ard to succeed in th eir projects. O r as
he p u t it: “ In the breast o f one w h o w ishes to d o so m eth in g new, the forces o f h ab it
raise u p an d b ear w itn ess again st the em b ryo n ic p ro ject” (Sch u m p eter 1934: 86).
Such in ertia, in S ch u m p ete r’s view , w as to som e extent en d o gen ou s, since it reflected
the em b ed ded ch aracter o f existin g kn ow led ge an d habit, w h ich , th o u g h “ en ergy­
saving,” ten d ed to b ias d ecisio n -m a k in g again st n ew w ays o f d o in g things.
H ence, in S ch u m p ete r’s early w o rk (so m etim es called “ Sch u m p eter M a rk Г )
in n o va tio n is the o u tco m e o f co n tin u o u s struggle in h isto rical tim e betw een in d i­
v id u a l entrepreneurs , ad vo catin g n o vel so lu tio n s to p a rticu la r problem s* a n d s s c it f
10 JAN FAGERBERG

inertia, with the latter seen as (partly) endogenous. T h is may, to som e extent, have
been an adequate interpretation o f events in Europe aro u n d the turn o f the nine­
teenth century. But during the first decades o f the tw entieth century, it becam e clear
to observers that innovations increasingly involve team w ork and take place within
larger organizations (see Bruland and M o w ery (Ch. 13)? L am (C h. 5)> and Lazonick
(Ch. 2) in this volum e). In later work, Schum peter acknow ledged this and empha­
sized the need for systematic study o f "coop erative” entrepreneurship in big firms
(so-called "Schum peter M ark II” ). However, he did not analyze the phenom enon in
m uch detail (although he strongly advised others to ).11
System atic theoretical and em pirical w ork on in n ovation -projects in firm s (and
the m anagem ent o f such projects) was slow to evolve, but d u rin g the last decades a
quite substantial literature has emerged (see chapters b y Pavitt and Lam in this
volum e). In general, research in this area coincides w ith Schum peter's em phasis on
uncertainty (Nelson and W inter 1982; N onaka and Takeuchi 1995; Van de Yen et a l
1999). In particular, for potentially rew arding innovations, it is argued, one may
sim ply not know what are the m ost relevant sources o r the best options to pursue
(still less how great the chance is o f success).12 It has also been em phasized that
innovative firm s need to consider the potential problem s that "p ath dependency”
m ay create (Arthur 1994). For instance, if a firm selects a specific innovation path
very early, it m ay (if it is lucky) enjoy “ first m over” advantages. But it also risks being
“ locked in” to this specific path through various self-reinforcing effects. I f in the end
it turns out that there actually existed a superior path, w hich som e other firm
equipped with m ore patience (or luck) happened to find, the early m over may be
in big trouble because then, it is argued, it m ay sim p ly be too costly or too late to
switch paths. It has been suggested, therefore, that in the early phase o f an inno vation
project, before sufficient knowledge o f the alternatives is generated, the best strategy
m ay sim ply be to avoid being “ stuck” to a particular path, and rem ain open to
different (and com peting) ideas/solutions. At the level o f the firm , this requires a
pluralistic leadership” that allows for a variety o f com peting perspectives ( Van de
Ven et al. 1999), in contrast to the hom ogenous, u n itary leader style that, in the
m anagem ent literature, is som etim es considered as the m ost advan tageou s.13
Openness to new ideas and solutions? is considered essential for innovation
projects, especially in the early phases. The principal reason for this has to do with a
fundam ental characteristic o f innovation: that every new in n ovation consists o f a
new com bination o f existing ideas, capabilities, skills, resources, etc. It follows
logically from this that the greater the variety o f these factors w ithin a given system,
the greater the scope for them to be com bined in different ways, producing new
innovations which will be both m ore com plex and m ore sophisticated. This evolu­
tionary logic has been used to explain why, in ancient tim es, the inhabitants of the
large Eurasian landmass came to be m ore innovative, and technologically sophisti­
cated, than small, isolated populations elsewhere around the globe (D iam on d 1998).
Applied m echanically on a population o f firm s, this logic m ight perhaps be taken to
 innovation : a g u i d e to t h e l i t e r a t u r e П

im p ly that large firm s sh ou ld be expected to be m o re in n ovative than small firms*14

H ow ever, m o d ern firm s are not d o se d system s co m parab le to isolated populations
o f ancient tim es. F irm s have learnt, b y necessity, to m o n ito r closely each o th e rs
steps, and search w id ely for new ideas, inputs, and sources o f in sp iratio n . T h e m o re
firm s on average are able to learn fro m interacting w ith external sources, the greater
the pressure on others to fo llo w suit. T h is greatly enhances the innovativeness o f
both in d iv id u a l firm s and the econ om ic system s to w h ich they b elon g (region s o r
countries, fo r instance). A rguably, this is o f p articu lar im p o rtan ce for sm aller firm s,
w hich have to com pensate fo r sm all internal resources b y being go od at interacting
w ith the outsid e w o rld . H ow ever, the grow in g co m p lexity o f the know ledge bases
necessary fo r in n o va tio n m eans that even large firm s in creasin gly d ep en d o n exter­
nal sources in their in n o vative activity (G ran stran d , Patel, and Pavitt, 1997; and in
this vo lu m e: Pavitt; Pow ell and G ro d al; N aru la and Z an fei).
H ence, cu ltivatin g the cap acity fo r ab sorb in g (outside) know ledge, so-called
'"absorptive cap a city ” (C o h en and Levin th al 1990), is a m u st fo r in n o vative firm s,
large or sm all. It is, how ever, so m eth in g that firm s o ften find v e ry challenging; the
"n o t invented h ere” syn d ro m e is a w ell-k n o w n feature in firm s o f all sizes. T h is
argu ab ly reflects the cu m u lative and em bedded character o f firm -sp ecific k n o w ­
ledge. In m ost cases, firm s d evelop their kno wledge o f h o w to do things in crem en ­
tally. Such know ledge, then, consists o f "ro u tin e s” that are rep ro d u ced th ro u gh
practice ("o rg an iz a tio n a l m e m o ry ” ; N elson and W in ter 1982). O ver tim e, the o r­
ganizational stru ctu re o f the firm and its know ledge base ty p ica lly co -evo lve into a
set-up that is b en eficial fo r the d a y -to -d a y o p eratio n s o f the firm . It has been argu ed ,
however, that su ch a set-up, w hile facilitatin g the d aily internal co m m u n icatio n /
interaction o f the firm , m a y in fact con strain the fir m s cap acity fo r ab so rb in g n ew
know ledge created elsew here, especially i f the new external kn ow ledge sign ifican tly
challenges the existin g set-u p /kn ow led ge o f the firm (so-called "co m p eten ce d es­
troying tech n ical ch an ge” ; T u sh m an an d A n d erso n 1986). In fact, such p ro b lem s
m ay o ccu r even fo r in n o vatio n s that are created internally. X ero x, fo r instance,
developed both the P C an d the m ou se, b u t failed to exp lo it co m m ercially these
in n o vation s, p rim a rily because th ey d id n o t seem to be o f m u ch valu e to the fir m s
existing p h o to -c o p ie r business (R o gers 1995).
T h u s o rgan izin g fo r in n o va tio n is a delicate task. R esearch in this area has, a m o n g
other things, p o in ted to the need fo r in n o vative firm s to allo w g ro u p s o f p eo p le
w ith in the o rgan izatio n su fficien t freed o m in exp erim en tin g w ith n ew so lu tio n s
(Van de V en 1999), an d establish in g patterns o f in teractio n w ith in the firm that a llo w
it to m o b ilize its entire kn o w led ge base w h en co n fro n tin g n ew challenges (N o n a k a
and Takeuchi 1995; L am , C h . 5 in this v o lu m e ). Su ch o rg an izin g d oes n o t sto p at th e
gate o f the firm , b u t extends to relations w ith extern al partn ers. T ie s to p artn ers w ith
w h o m c o m m u n ic a tio n is frequent are o ften called "stro n g ties,” w h ile th o se th at a re
m o re o cca sio n a l are d en o ted as "w e a k ties” (G ran o vetter 13731 see P ow ell an d
G ro d al, C h . 3 in th is v o lu m e ). P artn ers lin ked together w ith stro n g ties» e ith e r
12 JAN F A G E R B E R G

directly, or indirectly via a com m on partner, m ay self-organize into (relatively

stable) networks. Such networks m ay be very useful for m an agin g and maintaining
openness. But just as firm s can display sym ptom s o f path-dependency, the same can
happen to established networks, as the participants converge to a com m on percep­
tion o f reality (so-called “ group-think ). Innovative firm s therefore often find it
useful to also cultivate so-called weak ties in order to m ain tain a capacity for
changing its orientation (should it prove necessary).

1.4 T h e S y s t e m ic N a t u r e of I n n o v a t io n

As is evident from the preceding discussion, a central fin d in g in the literature is that,
in m ost cases, innovation activities in firm s depend h eavily on external sources. One
recent study sums it up well: “ Popular folklore notw ithstandin g, the innovation
journey is a collective achievem ent that requires key roles fro m n u m erou s entrepre­
neurs in both the public and private sectors” (Van de Yen et al. 1999:149)* In that
particular study, the term “ social system for innovation d evelopm en t” was used to
characterize this “ collective achievem ent ” H ow ever, this is ju st one am ong several
exam ples from the last decades o f h ow system concepts are applied to the analysis of
the relationship between innovation activities in firm s and the w ider fram ework in
which these activities are em bedded (see Edquist, C h. 7 in this volu m e).
One m ain approach has been to delineate system s on the basis o f technological,
industrial, or sectoral characteristics (Freem an et al. 1982; H ughes 1983; Carlsson and
Stankiewicz 1991; M alerba, Ch. 14 in this volum e) but, to a varyin g degree, to include
other relevant factors such as, for instance, institutions (laws, regulations, rules,
habits, etc.), the political process, the public research infrastructure (universities,
research institutes, support from public sources, etc.), financial institutions, skills
(labor force), and so on. To explore the technological dynam ics o f innovation, its
various phases, and how this influences and is influenced b y the w ider social,
institutional, and econom ic fram eworks has been the m ain focus o f this type of
analysis. Another im portant approach in the innovation-system s literature has
focused on the spatial level, and used national or regional borders to distinguish
between different systems. For exam ple, Lund vail (1992) and N elson et al. (1993)
have used the term national system o f innovation” to characterize the systemic
interdependencies within a given cou ntry (see Edquist in this volum e), while
Braczyk et al. (1997) sim ilarly have offered the notion o f “ regional innovation
systems (see Asheim and Gertler, Ch. 11 in this volu m e). Since the spatial systems
are delineated on the basis o f political and adm inistrative borders, such factors
 innovation : a g u i d e to t h e l i t e r a t u r e 13

n atu rally tend to play an im p o rtan t role in analyses based on this approach, which
has proven to be influ ential am o n g po licy m akers in this area, especially in Europ e
(see Lu n dvall and B o rras, C h. 22 in this volu m e). (Part II o f this v o lu m e analyzes
som e o f the constituent elem ents o f such system s in m o re d e t a il15)
W hat are the im p licatio n s o f ap p lyin g a system perspective to the study o f
in n ovation ? System s are— as netw orks— a set o f activities (or actors) that are
interlinked, and this leads n atu rally to a focus on the w o rk in g o f the linkages o f
the system .16 Is the poten tial fo r co m m u n icatio n and interaction th rou gh existing
linkages su fficien tly exploited? A re there po ten tial linkages w ith in the system that
m ight p ro fitab ly be established? Such questions ap p ly o f course to netw orks as well
as system s. H ow ever, in the n o rm al usage o f the term , a system w ill typ ically have
m o re “ stru ctu re” than a netw ork, and be o f a m ore en d u rin g character. T h e structu re
o f a system w ill facilitate certain patterns o f in teraction and ou tcom es (an d constrain
others), and in this sense there is a parallel to the role o f “ in ertia” in firm s. A d yn am ic
system also has feedbacks, w h ich m ay serve to reinforce— o r w eaken— the existin g
structure/ fu n ctio n in g o f the system , lead in g to “ lo ck in “ (a stable co n fig u ratio n ), o r
a change in o rien tatio n , o r— eventually— the d issolu tion o f the system . H ence,
system s m ay— ju st as firm s— be locked into a specific path o f d evelopm en t that
su p po rts certain types o f activities and constrains others. T h is m a y be seen as an
advantage, as It pu shes the p articip atin g firm s an d other actors in the system in a
direction that is d eem ed to be beneficial. B u t it m ay also be a d isadvantage, i f the
co n figu ratio n o f the system leads firm s to ign ore p o ten tially fru itfu l avenues o f
exp loratio n . T h e ch aracter o f such processes w ill be affected b y the extent to w h ich
the system exchanges im pu lses w ith its en viron m en t. T h e m o re open a system is fo r
im pulses fro m o u tsid e, the less the chance o f being “ locked o u t” fro m p ro m isin g
new path s o f d evelo p m en t that em erge ou tsid e the system . It is, therefore, im p o rtan t
for “ system m an ag ers” — such as p o licy m akers— to keep an eye o n the openness o f
the system , to avo id the p o ssib ility o f in n o vatio n activities b eco m in g u n d u ly c o n ­
strained b y self-rein fo rcin g path -d ep en den cy.
A n o th er im p o rta n t feature o f system s that h as com e into focu s is the stron g
com plem en tarities that co m m o n ly exist betw een the co m p o n en ts o f a system . If, in a
d yn am ic system , on e critical, co m p le m e n tary co m p o n en t is lackin g, o r foils to
progress o r develop, this m a y b lo ck o r slo w d o w n the grow th o f the entire system .
T h is is, as p o in ted o u t earlier, one o f the m ain reasons w h y there is o ften a v e ry
co n siderab le tim e lag betw een in ven tion an d in n o vatio n . E c o n o m ic h isto rian s h ave
co m m o n ly used concep ts such as “ reverse salients” an d “ b ottlen ecks” to ch aracter­
ize such p h en o m en a (H u gh es 1983; R o sen b erg 1982). H ow ever, such co n strain ts
need n o t be o f a p u re ly tech n ical ch aracter (such as, fo r instance, the failu re to invent
a decent b attery, w h ic h has severely con strain ed the d iffu sio n o f electric cars fo r
m o re th an ce n tu ry ), b u t m a y h ave to d o w ith lack o f p ro p e r in firn tractu re, fin an ce,
skills, etc. Some o f the m o st im p o rta n t in n o va tio n s o f this century, su ch as d e e tric it j
an d au to m o b ile s (M o w e ry a n d R o sen b erg 19 9 8 ), w ere d e j» id e n t o n v e ry ш т ш т
14 JAN FAGERBERG

infrastructural investments (w iring and roads/distribution-system s for fuel, re*

spectively). M oreover, to fulfil the potential o f the new in n o vation , such investments
often need to be accom panied by radical changes in the o rgan ization o f production
and distribution (and, m ore generally, attitudes: see Perez 19 8 3,19 8 5; Freeman and
Lou^a 2001). There are im portant lessons here fo r firm s an d p o licy m akers. Firms
m ay need to take into account the w ider social and econ o m ic im plications o f an
innovation project. The m ore radical an innovation is, the greater the possibility that
it m ay require extensive infrastructural investm ents an d /or organizational and
social change to succeed. I f so, the firm needs to th in k th rou gh the w ay in which it
m ay join up with other agents o f change in the private o r p u b lic sector. Policy
m akers, for their part, need to consider what different levels o f governm ent can do to
prevent “ bottlenecks” to occur at the system level in areas such as skills, the research
infrastructure, and the broader econom ic infrastructure.

1.5 How In n o v a t io n D if f e r s

One o f the striking facts about innovation is its variab ility over tim e and space* It
seems, as Schum peter (see Box 1.2) pointed out, to “ cluster,” not only in certain
sectors but also in certain areas and tim e periods. O ver tim e the centers o f innov­
ation have shifted from one sector, region, and co u n try to another. Fo r instance, for a
long period the worldw ide center o f innovation was in the U K , and the productivity
and incom e o f its population increased relative to its n eighb orin g countries, so that
by the m id-nineteenth century its productivity (and incom e) level was 50 per cent
higher than elsewhere; at about the beginning o f the twentieth cen tu ry the center of
innovation, at least for the m odern chem ical and electrical technologies o f the day,
shifted to Germ any; and now, for a long tim e, the w orld w ide center o f innovation
has been in the U SA, which during m ost o f the twentieth centu ry enjoyed the highest
productivity and living standards in the world. As explained b y Bruland and
M ow ery in this volum e, the rise o f the U S to w orld technological leadership was
associated with the growth o f new industries, based on the exploitation of
econom ies o f scale and scope (Chandler 1962, 1990) and m ass production and
distribution.
H ow is this dynam ic to be explained? Schum peter, extending an earlier line of
argum ent dating back to Karl M a rx ,17 held technological com petition (competition
through innovation) to be the driving force o f econom ic developm ent. I f one firm in
a given industry or sector successfully introduces an im p o rtan t innovation, the
argum ent goes, it will be am ply rewarded b y a higher rate o f profit. This functions
 innovation : a g u i d e to t h e l i t e r a t u r e 15

as a signal to other firm s (the im itators), w hich, if entry conditions allow, will
“ sw arm ” the in d u stry o r sector w ith the h ope o f sh aring the benefits (with the result
that the initial in n o vato r's first m over advantages m ay be quickly eroded). This
“ sw arm in g ” o f im itators im plies that the grow th o f the sector or industry in which
the in n o vatio n occu rs w ill be quite high for a while. S o o n er o r later, however, the
effects on grow th (created b y an in n o vatio n ) w ill be depleted and growth w ill slow
dow n.
To this essentially M a rx ia n sto ry Sch u m p eter added an im p o rtan t modification.
Im itators, he argued, are m u ch m o re likely to succeed in their aim s if they improve
on the origin al in n o vation , i.e., becom e in n o vato rs them selves. T h is is all the more
natural, he co n tin u ed , because one (im p o rtan t) in n o va tio n tends to facilitate
(induce) other in n o vatio n s in the sam e or related fields. In this way, in n o v a tio n -
d iffu sio n becom es a creative process— in w h ich one im p o rtan t in n o vatio n sets the
stage for a w h o le series o f subsequent in n o vatio n s— and n o t the passive, adaptive
process often assu m ed in m u ch d iffu sio n research (see H all in this v o lu m e). T h e
system ic interdepend en cies betw een the initial and indu ced in n o vatio n s also im ply
that In n ovation s (and grow th) “ tend to concentrate in certain sectors and their
su rro u n d in g s” o r “ clu sters” (Sch u m p eter 1939: 10 0 - 1) . Schum peter, as is well
kn o w n , lo o ked at this d yn am ic as a possible exp lan ato ry facto r b eh in d business
cycles o f vario u s lengths (Freem an and Lou^a 2001).
This sim ple schem e has been rem arkably successful in inspiring applications in
different areas. Fo r instance, there is a large am ou n t o f research that has adapted the
M an c-Sch u m p eter m odel o f technological com petition to the study o f industrial
grow th, international trade, and com petitiveness,18 although som etim es, it m ust be
said, w ith ou t acknow ledging the source for these ideas. A n early and very influential
contribution w as the so-called “ product-life-cycle th eo ry” suggested b y V ernon
(1966), in w hich indu strial grow th follow in g an im po rtan t product in n o vation w as
seen as com posed o f stages, characterized b y changing conditions o f and location o f
p ro d u ctio n .19 Basically w hat w as assum ed w as that the ab ility to d o produ ct in n o v­
ation m attered m ost at the early stage, in w hich there w ere m an y different and
com peting versions o f the produ ct o n the m arket. H ow ever, w ith tim e, the p ro d u ct
was assum ed to standardize, and this was assum ed to be accom panied by a greater
em phasis o n process innovation, scale econom ics, and cost-com petition. It was
argued that these changes in co m p etitive co n d itio n s m igh t initiate tran sfer o f the
tech n o lo gy fro m the in n o va to r c o u n try (h igh in com e) to co u n tries w ith large
m arkets an d /o r lo w costs. Su ch transfers m igh t also be associated w ith in tern atio n al
capital flo w s in the fo rm o f so-called fo reign d irect investm ents (F D Is), an d the
th eo ry has therefore also b eco m e k n o w n as a fra m e w o rk fo r exp lain in g such flo w s
(see N aru la an d Z a n fe i in this vo lu m e).
The “ product-life-cycle theory,” attractive as it was in its simplicity, was not
always corroborated by subsequent research. While it got some o f the general,
conjectures (b o rro w ed fro m Schumpeter) right, the rigorous scheme it aided.
16 JAN F A GE RB E R G

with well-defined stages, standardization, and changing com petitive requirements,

was shown to fit only a m inority o f industries (W alker 1979>C o h en 1995). Although
good data are hard to com e by, what emerges from em pirical research is a much
m ore com plex picture,^® with considerable differences across indu strial sectors in
the w ay this dynam ic is shaped. As exem plified b y the taxo n o m y suggested by Pavitt
(see Box 1.4), exploration o f such differences (“ indu strial d yn am ics” ) has evolved
into one o f the m ain areas o f research w ithin innovation studies (see in this volume:

Box 1.4 What is high-tech? Pavitt's taxonomy

The degree o f technological sophistication, or innovativeness, o f an industry or sector

is something that attracts a lot o f interest, and there have been several attempts to
develop ways o f classifying industries or sectors according to such criteria. The most
widely used in common parlance is probably the distinction between “ high-tech/5
“ medium-tech/5 and “ low-tech/5 although it is not always clear exactly what is meant
by this. Often it is equated with high, medium, and low R&D intensity in production
(or value added), either directly (in the industry itself) or including R&D embodied in
machinery and other inputs. Based on this, industries such as aerospace, computers,
semiconductors, telecommunications, pharmaceuticals, and instruments are com­
monly classified as “ high-tech/5 while “ medium-tech55 typically include electrical and
non-electrical machinery, transport equipment, and parts o f the chemical industries.
The remaining, “ low-tech/5 low R&D category, then, comprises industries such as
textiles, clothing, leather products, furniture, paper products, food, and so on (Fager-
berg 1997; see Smith in this volume for an extended discussion).
However, while organized R&D activity is an important source o f innovation in
contemporary capitalism, it is not the only one. A focus on R&D alone might lead one
to ignore or overlook innovation activities based on other sources, such as skilled
personnel (engineers, for instance), learning by doing, using, interacting, and so forth.
This led Pavitt ( 1984) to develop a taxonomy or classification scheme which took these
other factors into account. Based a very extensive data-set on innovation in the UK (see
Smith in this volume), he identified two (“ high-tech55) sectors in the economy, both
serving the rest of the economy with technology, but very different in terms o f how
innovations were created. One, which he labeled “ science-based/5 was characterized by
a lot o f organized R&D and strong links to science, while another— so-called “ special­
ized suppliers (of machinery, instruments, and so on)— was based on capabilities in
engineering, and frequent interaction with users. He also identified a scale-intensive
sector (transport equipment, for instance), also relatively innovative, but with fewer
repercussions for other sectors. Finally, he found a number o f industries that, although
not necessarily non-innovative in every respect, received most o f their technology from
other sectors,
An important result of Pavitt s analysis was the finding that the factors leading to
successful innovation differ greatly across industries/ sectors. This obviously called into
question technology or innovation polices that only focused on one mechanism, such
as, for instance, subsidies to R&D.
 innovation : a g u i d e to t h e l i t e r a t u r e 17

Ch. 14 by M alerb a; C h . 15 by V onT unzelm ann and Acha; Ch, 16 by M iles). Inspired, to
a large extent, b y the sem inal w o rk by N elson and W inter (see Box 1.5), research in
this area has exp lored the m an n er in w hich in du stries and sectors differ in terms o f
their internal d yn am ics (or “ tech nological regim es” : see M alerb a and O rsenigo
1997), fo cu sin g, in particular, on the differences across sectors in knowledge bases,
actors, n etw orks, and in stitu tion s (so called “ sectoral system s” : see M alerb a, Ch. 14
in this volu m e). A n im p o rtan t result fro m this research is that, since the factors that
influence in n o vatio n differ across indu stries, p o licy m akers have to take such
differences into accou n t w h en designing policies. T h e sam e p o licy (and policy
instrum ents) w ill not w o rk equ ally well everyw here.

Box 1.5 Industrial dynamics—an evolutionary interpretation

The book An Evolutionary Theory o f Economic Change (1982) by Richard Nelson and
Sidney Winter is one o f the most important contributions to the study o f innovation
and long run economic and social change. Nelson and Winter share the Schumpeterian
focus on “capitalism as an engine o f change” However, building on earlier work by
Herbert Simon and others (so-called “procedural” or “ bounded” rationality), Nelson
and Winter introduce a more elaborate theoretical perspective on how firms behave. In
Nelson and Winter’s models, firms’ actions are guided by routines, which are repro­
duced through practice, as parts o f the firms’ “ organizational memory.” Routines
typically differ across firms. For instance, some firms may be more inclined towards
innovation, while others may prefer the less demanding (but also less rewarding)
imitative route. I f a routine leads to an unsatisfactory outcome, a firm may use its
resources to search for a new one, which— if it satisfies the criteria set by the firm— will
eventually be adopted (so-called “ satisficing” behavior).
Hence, instead o f following the common practice in much economic theorizing o f
extrapolating the characteristics o f a “ representative agent” to an entire population
(so-called “ typological thinking” ), Nelson and Winter take into account the social, and
economic consequences o f interaction within populations o f heterogeneous actors (so-
called “ population thinking” ). They also emphasize the role o f chance (the stochastic
element) in determining the outcome o f the interaction. In the book, these outcomes
are explored through simulations, which allow the authors to study the consequences
off varying the value o f key parameters (to reflect different assumptions on techno-
logical progress, firm behavior, 'etc.). ■ They, 'distinguish ■ between a n ' тштттШш:
regime” Ы which. Ae-teAnologi^d/.iromfier Is assumed t o - o f f -

pmgressis ш огі endogenous and depends on what the.firms

“ cumulative” .regime). They also vary the ease/di.fficuity o f innovation and jtu itattaa
..,,' Nelsonahd Wmttfiwmk Ь^^.Ь^папітроШт. source of ip^iration ^subsequent
.. regmes^ and“indu^rMdynamk^r
'ioine' im prtant -
Sources Nelson and Winter і9^8;і| | Й е ^
ІШ
18 JAN F A G E R B E R G

1.6 In n o v a t io n a n d E c o n o m ic
Per fo r m an ce

The M arx-Schum peter model was not intended as a m odel o f industrial dynamics; its
prim ary purpose was to explain long run econom ic change, what Schum peter called
“ development ” The core o f the argument was (1) that technological com petition is the
m ajor form o f com petition under capitalism (and firm s not responding to these
demands fail), and (2) that innovations, e.g. 'n e w com binations"5 o f existing know­
ledge and resources, open up possibilities for new business opportunities and future
innovations, and in this way set the stage for continuing change. This perspective, while
convincing, had little influence on the econom ics discipline at the tim e o f its publica­
tion, perhaps because it did not lend itself easily to form al, m athem atical modeling of
the type that had become popular in that field. M ore recently, however, economists
(Rom er 1990), drawing on new tools for m athem atical m odeling o f econom ic phe­
nomena, have attempted to introduce som e o f the above ideas into form al growth
models (so-called4new growth theory55 or “ endogenous grow th theory55).21
In developing this perspective, Schum peter (1939) w as, as noted, particularly
concerned with the tendency o f innovations to “ cluster55 in certain contexts, and
the resulting structural changes in production, organization, dem and, etc. Although
these ideas were not well received b y the econom ic co m m u n ity at the tim e, the big
slum p in econom ic activity w orldw ide during the 1970s led to renew ed attention,
and several contributions emerged view ing long run econom ic and social change
from this perspective. Both M ensch (1979) and Perez (19 8 3,19 8 5 ), to take just two
exam ples, argued that m ajor technological changes, such as, fo r instance, the ICT
revolution today, or electricity a century ago, require extensive organizational and
institutional change to run their course. Such change, how ever, is difficult because of
the continuing influence o f existing organizational and institutional patterns. They
saw this inertia as a m ajor grow th-im peding factor in periods o f rapid technological
change, possibly explaining some o f the variation o f grow th over tim e (e.g. booms
and slumps) in capitalist econom ies. W hile the latter p ro p o sitio n rem ains contro­
versial, the relationship between technological, organizational, and institutional
change continues to be an im portant research issue (Freem an and Lou^a 2001),
with im portant im plications both for the analysis of the d iffu sion o f new technolo­
gies (see Hall in this volum e) and the policy discourse (see Lu ndvall and Borras in
this volum e).
A lthough neither M arx nor Schum peter applied their dynam ic perspective to the
analysis o f cross-national differences in growth perform ance, fro m the early 1960s
onwards several contributions emerged that explore the potential o f this perspective
for explaining differences in cross-country grow th. In what cam e to be a very
influential contribution, Posner (1961) explained the difference in econom ic growth
 innovation : a g u i d e to t h e l i t e r a t u r e 19

between tw o cou ntries, at different levels o f econ o m ic and technological develop­

m ent, as resulting fro m tw o sources: in n o vatio n , w hich enhanced the difference, and
im itation, w h ich tended to reduce it. T h is set the stage for a lo n g series o f co n trib u ­
tions, often labeled “ tech n o lo gy gap” o r “ n o rth -s o u th ” m odels (or app roach es),
focusin g on exp lain in g such differences in econ om ic grow th across countries at
different levels o f d evelopm en t (see Fagerberg 1994, 1996 for details). A s for the
lessons, o n e o f the theoretical co n trib u tors in this area su m m ed it up w ell when he
concluded that: “ Like A lice and the Red Q ueen, the developed region has to keep
ru n n in g to stay in the sam e place” (K ru gm an 1979: 262).
A w eakness o f m u ch o f this w o rk w as that it w as based on a v ery stylized
representation o f the glo bal d istrib u tion o f in n o vation , in w h ich in n o vatio n w as
assum ed to be concentrated in the developed w o rld , m ain ly in the U S A . In fact, as
argued b y Fagerberg and G o d in h o in this vo lu m e, the successful catch -u p in
tech n o logy and in com e is n o rm ally not based o n ly o n im itatio n , b u t also involves
in n o vatio n to a sign ifican t extent. A rgu ab ly, this is also w h at one sh ou ld expect fro m
the Sch u m p eterian perspective, in w h ich in n o vatio n is assu m ed to be a p ervasive
p h en om en on . Fagerb erg (1987, 1988) identified three factors affecting differential
grow th rates across countries: in n o vatio n , im itatio n , and o th er efforts related to
the co m m ercial exp lo itatio n o f technology. The analysis suggested that su p erio r
in n o vative activity w as the p rim e factor b eh in d the huge difference in p erfo rm an ce
between A sian an d Latin A m erican N IC cou n tries in the 1970s and early 1980s.
Fagerberg an d V erspagen (2002) likew ise fo u n d that the co n tin u in g rap id grow th
o f the A sia n N IC s relative to other co u n try g ro u p in g s in the decade that follow ed
w as p rim a rily caused b y the rap id grow th in the in n o vative p erfo rm an ce o f
this region. M o reo ver, it has been sh o w n (Fagerberg 1987; Fagerb erg and V erspagen
2002) that, w h ile im ita tio n has b eco m e m o re d em an d in g o ver tim e (and hence
m ore d ifficu lt an d /o r co stly to u n d ertak e), in n o vatio n has grad u ally b ecom e a
m ore p o w erfu l facto r in exp lain in g differences across co u n tries in econ o m ic
grow th.

1.7 W h a t do w e K n o w a b o u t In n o v a t io n ?
A n d w h a t d o w e N e e d t o

L e a r n m o r e a b o u t ?

Arguably, we have a good understanding of the role played by innovation in long run
economic and social change, and many of its consequences:
20 JAN F A G E RB E R G

* The function o f innovation is to introduce novelty (variety) into the economic

sphere. Should the stream o f novelty (innovation) d ry up, the econ o m y will settle
into a “ stationary state" with little or no grow th (M etcalfe 1998). Hence, innov­
ation is crucial for long-term econom ic growth.
* Innovation tends to cluster in certain industries/sectors, w h ich consequently grow
m ore rapidly, im plying structural changes in p ro d u ction and dem and and, even­
tually, organizational and institutional change. The cap acity to undertake the
latter is im portant for the ability to create and to benefit fro m innovation.
* Innovation is a powerful explanatory factor behind differences in performance
between firm s, regions, and countries. Firm s that succeed in in n ovation prosper,
at the expense o f their less able com petitors. In novative countries and regions have
higher productivity and incom e than the less in n ovative ones. Countries or
regions that wish to catch up with the innovation leaders face the challenge of
increasing their ow n innovation activity (and “ absorptive cap acity” ) towards
leader levels (see G odinho and Fagerberg in this volu m e).
Because o f these desirable consequences, p olicy m akers an d business leaders
alike are concerned with ways in w hich to foster innovation. N evertheless, in spite
o f the large am ount o f research in this area during the past fifty years, we know
m uch less about w hy and how innovation occurs than w hat it leads to. Although
it is by now well established that innovation is an organizational phenomenon,
m ost theorizing about innovation has tradition ally looked at it fro m an individual­
istic perspective, as exemplified b y Schum peter's “ psych olo gical” th eo ry o f entre­
preneurial behavior (Fagerberg 2003). Sim ilarly, m ost w o rk on cognition and
knowledge focuses on individuals, not organizations. A n im p o rtan t exception
was, o f course, Nelson and W inter (1982), w hose focus on “ organizational m em ory”
and its links to practice paved the w ay for m uch subsequent w o rk in this area.22
But our understanding o f how knowledge— and in n o vation — operates at the
organizational level rem ains fragm entary and further conceptual and applied re­
search is needed.
A central finding in the innovation literature is that a firm does not innovate in
isolation, but depends on extensive interaction w ith its environm ent. Various
concepts have been introduced to enhance our understanding o f this phenomenon,
most of them including the terms “ system ” or (som ew hat less am bitious) “ netw ork”
Som e o f these, such as the concept o f a “ national system o f innovation,” have become
popular am ong policy makers, who have been constrained in their ability to act by
lack o f a sufficiently developed fram ew ork for the design and evaluation o f policy.
Still, it is a long w ay from pointing to the system ic character o f in n ovation processes
(at different levels o f analysis), to having an approach that is sufficiently developed
to allow for systematic analysis and assessm ent o f po licy issues. Arguably, to be
really helpful in that regard, these system approaches are in need o f substantial
elaboration and refinement (see the chapter by Edquist in this volum e).
 in n o v a t io n : a guide to the literature 21

O ne obstacle to im p ro vin g o u r u n d erstan d in g is that in n o vation has been studied

b y different co m m u n ities o f researchers w ith different b ackgro u n d s, and the failure
o f these co m m u n ities to co m m u n icate m ore effectively w ith one anoth er has
im peded progress in this field. O ne consequence o f these co m m u n icatio n d ifficulties
has been a certain degree o f “ fuzziness” w ith respect to basic concepts, w hich can
o n ly be im p ro ved by b rin g in g these d ifferent co m m u n ities together in a constructive
d ialogue, and the present vo lu m e sh ou ld be seen as a co n trib u tio n tow ard s this aim ,
D ifferent, and to som e extent com petin g, perspectives sh ou ld not alw ays be seen as a
problem : m a n y social p h en o m en a are to o co m p lex to be analyzed p ro p e rly fro m a
single d iscip lin ary perspective. A rgu ably, in n o vatio n is a p rim e exam ple o f this.

N otes

1. I wish to thank my fellow editors and contributors for helpful comments and suggestions.
Thanks also to Ovar Andreas Johansson for assistance in the research, Sandro Mendon^a
for his many creative inputs (which I unfortunately have not have been able to follow to
the extent that he deserves), and Louise Earl for good advice. The responsibility for
remaining errors and omissions is mine.
2. A consistent use o f the terms invention and innovation might be to reserve these for the
first time occurrence o f the idea/concept and commercialization, respectively. In practice
it may not always be so simple. For instance, people may very well conceive the same idea
independently o f one another. Historically, there are many examples o f this; writing,
for instance, was dearly invented several times (and in different cultural settings)
throughout history (Diamond 1998). Arguably, this phenomenon may have been
reduced in importance over time, as communication around the globe has progressed.
3. In the sociological literature on diffusion (i.e, spread o f innovations), it is common to
characterize any adopter o f a new technology, product, or service an innovator. This then
leads to a distinction between different types o f innovators, depending on how quick they
are in adopting the innovation, and a discussion o f which factors might possibly explain
such differences (Rogers 1995). While this use o f the terminology may be a useful one in
the chosen context, it dearly differs from the one adopted elsewhere. It might be preferable
to use terms such as “ imitator” or “ adopter” for such cases.
4. Similarly for automobiles: while the idea o f a power-driven vehicle had been around for a
long time, and several early attempts to commercialize cars driven by steam, electricity,
and other sources had been made, it was the incorporation o f an internal combustion
engine driven by low-cost, easily available petrol that made the product a real hit in the
market (Mowery and Rosenberg 1998).
5. A somewhat similar distinction has been suggested by Henderson and Clark (1990). They
distinguish between the components (or modules) o f a product or service and the way these
components are combined, e.g. the product “ d e s ip ” or “ architecture” A change only in
the former is dubbed “ modular innovation” change only in the latter “architectural
innovation” They a r p e that these two types o f innovation rely on different types o f
knowledge (and, hence, create different challenges for the firm).
22 IAN F A G E R B E R G

6. In fact, many economists go so far as to argue that the savings in costs, following a process
innovation in a single firm or industry, by necessity will generate additional income and
demand in the economy at large, which will “ compensate” for any initial negative effects
of a process innovation on overall employment. For a rebuttal, see Edquist 2001 and
Pianta, Ch. 21 in this volume.
7. Schumpeter 1934: 66.
8. In the sociological literature on innovation the term "reinvention” is often used to
characterize improvements that occur to a product or service, while it is spreading in a
population of adopters (Rogers 1995)*
9. In the Community Innovation Survey (CIS) firms are asked to qualify novelty with
respect to the context (new to the firm, industry or the world at large). See Smith in this
volume for more information about these surveys.
10. Kim and Nelson (2000a) suggest the term "active imitation” for producers who, by
imitating already existing products, modify and improve them.
11. For Instance, in one o f his last papers, he pointed out: "To let the murder out and start my
final thesis, what is really required is a large collection o f industrial and locational
monographs all drawn up according to the same plan and giving proper attention on
the one hand to the incessant historical change in production and consumption func­
tions and on the other hand to the quality and behaviour o f leading personnel”
(Schumpeter 1949/1989:328).
12. Even in cases where the project ultimately is successful in aims, entrepreneurs face the
challenge of convincing the leadership o f the firm to launch it commercially (which may
be much more costly than developing it). This may fail if the leadership o f the firm has
doubts about its commercial viability. It may be very difficult for management to foresee
the economic potential of a project, even if it is "technically” successful Remember, for
instance, IBM director Thomas Watsons dictum in 1948 that “ there is a world market for
about five computers” (Tidd et al. 1997: 60)!
13. "A unified homogenous leadership structure is effective for routine trial-and-error
learning by making convergent, incremental improvements in relatively stable and
unambiguous situations. However, this kind o f learning is a conservative process that
maintains and converges organizational routines and relationships towards the existing
strategic vision. . . although such learning is viewed as wisdom in stable environments, it
produces inflexibility and competence traps in changing worlds” (Van de Yen et al.
1999:117).
14. It would also imply that large countries should be expected to be more innovative than
smaller ones, consistent with, for instance, the prediction o f so-called "new growth”
theory (Romer 1990)- See Verspagen in this volume.
15. See, in particular, Ch, 10 by Granstrand (intellectual property rights), Ch. 8 by Mowery
and Sampat (universities and public research infrastructure), and Ch. 9 by O'Sullivan
(finance).
16. This is essentially what was suggested by Porter (1990).
17. See Fagerberg 2002, 2003 for a discussion of this “ Marx-Schumpeter” model.
18. See Fagerberg (1996), Wakelin (1997), and Cantwell, Ch. 20 in this volume for overviews
o f some of this literature.
19. For a more recent analysis in this spirit, with a lot o f empirical case-studies, see Utterback
(1994).
 in n o v a t io n : a g u id e to th e lit e r a t u r e 23

20** Available econometric evidence suggests that innovation, measured in various ways (see
Smith in this volume), matters in many industries, not only those which could be
classified as being in the early stage o f the product-cycle (Soete 1987; Fagerberg 1995),
21, For an overview, see Aghion and Howitt (1998). See also the discussion in Fagerberg
(2002, 2003), and Ch. 18 by Verspagen in this volume.
22. For a discussion o f the role o f different types o f knowledge in economics, including the
organizational dimension, see Cowan et a l (2000) and Ancori et a l (2000).

R efer en c es

A g h io n , R, and H o w itt , R (1998), Endogenous Growth Theory> Cambridge, Mass.: M IT

Press.
A n co ri , B*, B u r e t h , A., and C o h en d et , P. (2000), “ The Economics o f Knowledge: The
Debate about Codification and Tacit Knowledge/' Industrial Dynamics and Corporate
Change 9: 255-87.
A n d e r se n , E. S. (1994b Evolutionary Economics, Post-Schumpeterian ContributionsyLondon:
Pinter.
A r t h u r , W. B. (1994), Increasing Returns and Path Dependency in the Economy Ann Arbor:
University o f Michigan Press.
B r a c z y k , H. J. et al. (1998), Regional Innovation Systems, London: UCL Press.
B u sh , V. (1945), Science: The Endless Frontier. Washington: US Government Printing Office*
C a r lsso n , B., and S t a n k ie w ic z , R. (1991), “ On the Nature, Function and Composition o f
Technological Systems," Journal of Evolutionary Economics 1: 93-118.
C handler , A. D. (1962), Strategy and Structure: Chapters in the History o f the American
Industrial Enterprise, Cambridge, Mass.: M IT Press.
------ (1990) Scale and Scope: The Dynamics o f Industrial Capitalism, Cambridge, Mass.:
Harvard University Press.
C o h en , W. (1995), “ Empirical Studies o f Innovative Activity," in P. Stoneman (ecL),
Handbook o f the Economics o f Innovation and Technological Change, Oxford: Blackwell,
182-264.
*------ and L e v in t h a l , D. (1990), “Absorptive Capacity: A New Perspective on Learning and
Innovation," Administrative Science Quarterly 3 5 :123-33.
C ow an , R., D a v id , P. A., and F oray , D. (2000), “ The Explicit Economics o f Knowledge
Codification and Tacitness," Industrial Dynamics and Corporate Change 9: 211-53.
D iamond , J. (1998), Guns, Germs and Steel: A Short History o f Everybody for the Last 13000
Years, London: Vintage.
D osi, G. (1988), “ Sources, Procedures and Microeconomic Effects o f Innovation," Journal o f
Economic Literature 26 :1120 -71.
------ F r eem a n , C , N elson , R., S il v e r b e r g , G., and S oete , L. G. (eds.) (1988), Technical
Change and Economic Theory London: Pinter.
E dquist , C , H om m en , L., and M c K e l v e y , M. (2001), Innovation and Employment: Process
versus Product Innovation, Cheltenham: Elgar.

* Asterisked items are suggestions for further reading.

24 JAN FAGERBERG

F a g e r b e r g , J. (1987), “A Technology Gap Approach to Why Growth Rates Differ,” Research

Policy 16:87-99, repr. as ch. 1 in Fagerberg (2002).
____ (1988), “ Why Growth Rates Differ,” in Dosi et al. 1988: 432-57.
____ (1994), “ Technology and International Differences in Growth Rates,” Journal ofEco-
nomic Literature 32(3): П47- 75*
____ (1995), “ Is There a Large-Country Advantage in High-Tech?,” NUPI Working Paper No.
526, Norwegian Institute o f International Affairs, Oslo, repr. as ch. 14 in Fagerberg (2002),
____(1996), “ Technology and Competitiveness,” Oxford Review of Economic Policy 12; 39-51,
repr. as ch. 16 in Fagerberg (2002).
____ (1997), “ Competitiveness, Scale and R&D,” in J. Fagerberg et al., Technology and
International Trade, Cheltenham: Edward Elgar, 38-55. repr. as ch. 15 in Fagerberg (2002).
____ (2000), “ Vision and Fact: A Critical Essay on the Growth Literature ” in J. Madrick (ed.),
Unconventional Wisdom: Alternative Perspectives on the New Economyt New York: The
Century Foundation, 299-320, repr. as ch. 6 in Fagerberg (2002).
* ---- (2002), Technology Growth and Competitiveness: Selected Essays, Cheltenham: Edward
Elgar,
------(2003), “ Schumpeter and the Revival o f Evolutionary Economics: An appraisal of the
Literature,” Journal of Evolutionary Economics 13:125-59.
------and V erspagen , B. (2002), “ Technology-Gaps, Innovation-Diffusion and Transfor­
mation: An Evolutionary Interpretation,” Research Policy 31:1291-304.
F reem an , C. (1987), Technology Policy and Economic Performance: Lessons from Japan,
London: Pinter.
------C lark , ]., and Soete , L, G. (1982), Unemployment and Technical Innovation: A Study of
Long Waves and Economic Development, London: Pinter,
* ----and S oete , L. (1997), The Economics of Industrial Innovation, 3rd edn, London:
Pinter.
------and LougA, F. (2001), As Time Goes By: From the Industrial Revolutions to the Infor­
mation Revolution, Oxford: Oxford University Press.
G ranovetter , M. (1973), “ The Strength of Weak Ties,” American Journal of Sociology 78:
1360-80.
G ranstrand , O., P atel , P., and Pa v it t , K. (1997), “ Multi-technology Corporations: Why
They Have Distributed* rather than 'Distinctive Core* Competencies,” California Man­
agement Review 39: 8-25.
H enderson , R. M., and C lark , R. B. (1990). “Architectural Innovation: The Reconfigura­
tion of Existing Product Technologies and the Failure o f Established Firms,” Adminis­
trative Science Quarterly 29: 26-42.
H obday , M. (2000), East versus Southeast Asian Innovation Systems: Comparing OEM-
and TNC-led Growth in Electronics ” in Kim and Nelson 2000 b: 129—69.
H u g h e s , T. P. (1983), Networb of Power>Electrification in Western Society 1880-1930, Balti­
more: The Johns H opkins University Press.
K im , L., and N e l so n , R. R. (2000a) “ Introduction,” in Kim and N elson 2000b: 13-68.
(2000b), Technology Learning and Innovation: Experiences of Newly Industrializing
Economies, Cam bridge: Cam bridge University Press.
* Kline , S. J„ and Rosenberg, N. (1986), “An Overview of Innovation,” in R. Landau and N.
Rosenberg (eds.), The Positive Sum Strategy: Harnessing Technology for Economic Growth,
Washington, DC: National Academ y Press, 275—304
 in n o v a t io n : a. g u i d e t o t h e literature 25

K rugm an , P. (i979)> UA Model o f Innovation, Technology Transfer and the World Distribu­
tion o f Incom e” Journal of Political Economy 87: 253-66.
Lu n d vall , B. A. (1988), "‘Innovation as an Interactive Process: From User-Producer Inter­
action to the National System o f Innovation” in Dosi et al. 1988: 349—69.
------ (ed.) (1992), National Systems of Innovation: Towards a Theory of Innovation a n d
Interactive Learning, London: Pinter.
M a ler b a , R, and O rsenigo , L. (1997), “ Technological Regimes and Sectoral Patterns o f
Innovative Activities” Industrial and Corporate Change6: 83-117.
------ N elson , R. R., O rsenigo , L., and W in ter , S. G. (1999), “ ‘History-friendly' Models o f
Industry Evolution: The Computer Industry,” Industrial Dynamics and Corporate Change
8 :1-36.
M ensch , G. (1979), Stalemate in Technology, Cambridge, Mass.: Ballinger Publishing C om ­
pany.
M etc a lfe , J. S. (1998), Evolutionary Economics and Creative Destruction, London: Routledge.
*M ow ery , D., and Ro senberg , N. (1998), Paths of Innovation, Technological Change in 20th-
Century America, Cambridge: Cambridge University Press.
N elson , R. R. (ed.) (1993), National Systems of Innovation: A Comparative Study; Oxford:
Oxford University Press.
------ and W in ter , S. G. (1982), An Evolutionary Theory o f Economic Change, Cambridge,
Mass.: Harvard University Press.
*N o naka , I., and T ak eu ch i , H. (1995), The Knowledge Creating Company, Oxford: Oxford
University Press.
*P a v it t , K. (1984), “ Patterns o f Technical Change: Towards a Taxonomy and a Theory,”
Research Policy 13: 343-74.
Perez , C. (1983), “ Structural Change and the Assimilation o f New Technologies in the
Economic and Social System,” Futures 15: 357-75.
— (1985), “ Micro-electronics, Long Waves and World Structural Change,” World Develop­
ment 13: 441-63.
Porter , M. E. (1990), “ The Competitive Advantage ofNations,” H arvard Business Review 68:
73- 93.
Posner , M. V. (1961), “ International Trade and Technical Change,” Oxford Economic Papers
13: 323-41.
^Rogers , E. (1995), Diffusion o f Innovations, 4th edn., New York: The Free Press.
Romer , P. M. (1990), “ Endogenous Technological Change,” Journal o f Political Economy 98:
S71-S102.
Rosenberg , N. (1976), Perspectives on Technology, New York: Cambridge University Press.
------ (1982), Inside the Black Box: Technology and Economics, New York: Cambridge Univer­
sity Press.
S chmookler , J. (1966), Invention and Economic Growth, Cambridge, Mass.: Harvard Uni­
versity Press.
Sch u m peter , J. (1934), The Theory o f Economic Development, Cambridge, Mass.: Harvard
University Press.
------ (1939)» Business Cycles: A Theoretical, Historical, and Statistical Analysis o f the Capitalist
Process, 2 vols., New York: McGraw-Hill.
* ------ (1943), Capitalism, Socialism and Democracy, New York: Harper.
-— — (1949), “ Economic Theory and Entrepreneurial History* Change m i the Ernmprmmm
63-84, repr. in J. Schumpeter (1989), Essays on Entrepreneurs, Innovations» Business
26 JAN FAGERBERG

and the Evolution of Capitalism, ed. Richard V. Clemence, New Brunswick, NJ: Transaction
Publishers, 253-61.
Schumpeter , R. (1954), History o f Economic Analysis, New York: Allen & Unwin.
Shionoya, Y. (1997), Schumpeter and the Idea o f Social Science, Cambridge: Cambridge
University Press.
S oete , L. (1987), “ The Impact o f Technological Innovation on International Trade Patterns;
The Evidence Reconsidered” Research Policy 16 :101-30 .
Sw edberg , R, (1991), Joseph Schumpeter: His Life and Work, Cambridge: Polity Press.
T idd, J., B ess an t , J., and Раѵітт , K. (1997), Managing Innovation: Integrating Technological
Market and Organizational Change, Chichester: John Wiley 8c Sons.
T ushman , M. L., and A nderson , P. (1986). “ Technological Discontinuities and Organiza­
tional Environments ” Administrative Science Quarterly 31(3): 439--65.
U tterba ck , J. M. (1994), Mastering the Dynamics of Innovation, Boston: Harvard Business
School Press.
V an de V en , A., P olley , D. E., G arud , R., and V en k a ta r a m a n , S. (1999), The Innovation
Journey; New York: Oxford University Press.
V ernon , R. (1966), “ International Investment and International Trade in the Product Cycle,”
Quarterly Journal of Economics So: 190-207.
V on H ip p e l , E. (1988), The Sources of Innovation, New York: Oxford University Press.
W a k elin , K. (1997)> Trade and Innovation: Theory and Evidence, Cheltenham: Edward Elgar.
W a lk er , W. B. (1979), Industrial Innovation and International Trading Performance, Green­
wich: JAI Press.
I NNOVATI ON IN
THE M A K I N G
 In t r o d u c t io n t o P a r t I

M o st innovations occur in firm s o r other types o f organizations.

The contributions in this section survey o u r current know ledge on
the organizational structure and context o f the process o f in n o v­
ation. Chapter 2, by Lazonick, provides a historical perspective on
the developm ent o f innovative firm s, fro m the sm all and m ed iu m ­
sized firm s o f the First Industrial R evolution th ro u gh the m u lti­
divisional diversified industrial firm s o f the U S and Jap an in the
twentieth century to the current debate on the “ N ew E co n o m y”
and network-based business m odels. Powell and G ro d al deal m ore
extensively with the role o f networks in in n ovation in the subse­
quent chapter. Chapter 4, b y Pavitt, discusses in n o vatio n processes
within firm s, and uses an extensive survey o f the relevant literature
to provide an analytical perspective on the factors affecting the
perform ance and m anagem ent o f innovation w ith in the large firm .
A com plem entary chapter by Lam (C hapter 5) focuses on firm s'
experiences with organizational innovation. Finally, C h apter 6 by
Sm ith deals with an indispensable prerequisite fo r the stu d y o f
innovation, the m easurem ent o f innovation-related activities, p ar­
ticularly in firms.
 C H A P T E R 2

THE INNOVATIVE
FIRM

WILLIAM LAZONICK

2.1 In t r o d u c t io n

W h a t m akes a firm Innovative? How have the characteristics o f innovative firms

changed over time? To address these questions, one requires a conceptual ffamework
for analyzing how a firm, tran sfo rm s p ro d u ctive resources into goods and services
that customers want at prices they can afford. To make this productive transform­
ation , a firm must engage in three generic activities: strategizm g, financings and
organizing. The types o f strategy, finance, and o rg an izatio n that support A e innov­
ation process change over time and can vary m ark ed ly across industrial activities and
institutional environments at any point in time. The innovative film, must, Aero-
fore, be analyzed in comparative-historical perspective. This chapter presents and
illustrates a framework for analyzing A e “social conditions of mnovatiweenterprise*
in A e comparative—historical experiences of A e advanced economies.
Section 2.2 builds upon prominent theories o f A e innovative firm to derive the
“social ronditioiis of innovative enterprise” framework. Section 2*3 focuses on the
regional agglomerations of capabilities,, now known as щМ в 1ш!1іііп industrial
d is A e t^ that» by the late nineteenA century, had enabled. Britain Ш emerge m
A t m m M % first md»trial nation. Section%4 provides a pmpecrive on A e enter*
®ftfc ilS mmmgmd-ootpocation that д о р с М the-US есм м яу
A.,;intei»tMiiilal'.iiid'iii^Ai^fc®^^^#'.A№ring the first half o f A e 'tw eiAetriettucy*1
30 WILLIAM LAZONICK

Over the past few decades, the greatest challenges to the U S m anagerial corporation
have com e from Japan. Section 2.5 identifies the social co n d itio n s o f innovative
enterprise that have characterized the Japanese m odel, w hile Section 2.6 outlines the
distinctive characteristics o f the U S N ew E con om y firm that has gained competitive
advantage in a num ber o f critical product m arkets in the in fo rm atio n and commu­
nication technology (IC T) industries. Section 2.7 draw s som e general conclusions
from this essay's com parative—historical perspective con cern in g strategy, finance,
and organization in the innovative firm , and the m eth o d o lo g y for studying these
phenom ena.

2.2 S o c i a l C o n d i t i o n s o f
In n o v a t iv e E n t e r p r is e

Firm s strategize when they choose the product m arkets in w h ich they want to
com pete and the technologies w ith which they h ope to be com petitive. Firm s finance
when they m ake investments to transform technologies and access m arkets that can
only be expected to generate revenues som etim e in the future. Firm s organize when
they com bine resources in the attem pt to tran sform them into saleable products,
To strategize, finance, and organize is not necessarily to innovate. B y definition,
innovation requires learning about how to tran sform technologies and access
markets in ways that generate higher quality, low er cost products. Learning is a
social activity that renders the innovation process uncertain, cum ulative, and
collective (O 'Sullivan 2000b). The innovation process is uncertain because, by
definition, what needs to be learned about tran sform in g technologies and accessing
markets can only becom e know n through the process itself. By investing in learning,
an innovative strategy confronts the uncertain character o f the innovation process.
The innovation process is cum ulative when learning cannot be done all at once; what
is learned today provides a foundation for what can be learned tom orrow . Invest­
ments in cum ulative learning, therefore, require sustained, com m itted finance. The
innovation process is collective when learning cannot be done alone; learning
requires the collaboration o f different people with different capabilities. Investments
in collective learning, therefore, require the integration o f the w ork o f these people
into an organization.
W hat is the theory o f the firm that can com prehend h o w strategizing, financing,
and organizing can support the Innovation process? O ver the past century, the
theoretical efforts o f econom ists have focused m ain ly on the optim izing firm
rather than the innovating firm . The optim izing firm takes as given technological
 THE I NN OV AT I VE FIRM 3]

capabilities and m arket prices (for inputs as well as outputs), and seeks to m axim ize
profits on the basis o f these technological and m arket constraints. In sharp con trast,
in the attem pt to generate higher quality, low er cost products than had previously
been available, and thus differentiate itself from com petitors in its industry, the
innovating firm seeks to tran sform the technological and m arket conditions that the
optim izing firm takes as “ given” constraints. Hence, rather than constrained o p ti­
m ization, the innovating firm engages in what I call “ historical tra n sfo rm a tio n ” a
m ode o f resource allocation that requires a theoretical perspective on the processes
o f industrial and organizational change (Lazonick 2002a).
The distinction between the innovating and optim izing firm is im plicit in the
w ork o f A lfred M arshall, w hose Principles o f Economics, published in eight editions
between 1890 and 1920, placed the th eo ry o f the firm at the center o f econom ic
analysis. A lth ou gh M arshall's follow ers used his argum ents to construct the theory
o f the optim izing firm that rem ains entrenched in econom ics textbooks, M arshall
(1961:3x5) h im self displayed considerable insight into the dynam ics o f the in n o vat­
ing firm , as revealed in the follow in g passage:

An able man, assisted by some strokes o f good fortune, gets a firm footing in the trade, he
works hard and lives sparely, his own capital grows fast, and the credit that enables him to
borrow more capital grows still faster; he collects around him subordinates of more than
ordinary zeal and ability; as his business increases they rise with him, they trust him and he
trusts them, each o f them devotes himself with energy to just that work for which he is
specially fitted, so that no high ability is wasted on easy work, and no difficult work is
entrusted to unskillful hands. Corresponding to this steadily increasing economy o f skill, the
growth o f his firm brings with it similar economies o f specialized machines and plants o f all
kinds; every improved process is quickly adopted and made the basis o f further improve­
ments; success brings credit and credit brings success; success and credit help to retain old
customers and to bring new ones; the increase o f his trade gives him great advantages in
buying; his goods advertise one another and thus diminish his difficulty in finding a vent for
them. The increase o f the scale o f his business increases rapidly the advantages which he has
over his competitors, and lowers the price at which he can afford to sell.

W hat then constrains the grow th o f such a firm ? In Industry and Trade>published
in 1919, A lfred M arsh all acknow ledged that over the previou s decades the large-scale
enterprise had becom e d o m in an t in advanced nations such as the U nited States and
Germ any. H e invoked, how ever, the aphorism , “ shirtsleeves to shirtsleeves In three
generations” (M arsh all 1961: 621) to explain the lim it to the grow th o f the firm that
w ould prevent a sm all n u m b er o f large firm s fro m d om inatin g an industry. A n
ow ner-entrepreneur o f exceptional ab ility w o u ld fou nd and build a successful firm .
In the second generation, control w o u ld pass to descendants w h o could not be
expected to have the capabilities o r drive o f the founder, and as a result the firm
would grow m o re slowly o r even stagnate. The third generation w o u ld lose touch
w ith the innovative legacy o f the first generation, and the firm w ou ld w ither aw ay in
the face o f n ew en trepreneurial com petition.
32 WILLIAM LAZONICK

W riting in the first decades of the twentieth century, Jo sep h Schum peter (1934)
also focused on the innovative entrepreneur w ho, b y creating “ n ew combinations”
o f productive resources, could disrupt the “ circu lar flo w o f econom ic life as
conditioned by given circum stances ” In effect, Schu m peter was arguing that,
through entrepreneurship, which he called the fu n dam en tal phenom enon of
econom ic developm ent,” innovating firm s could challenge o p tim izin g firms, and
thereby drive the developm ent o f the econom y. In i9it> w hen he first published
The Theory of Economic Development (in G erm an ), Schum peter, like Marshall,
viewed the innovative firm as the result o f the entrepreneurial w o rk o f an extra­
ord in ary individual. Over the subsequent decades, how ever, as Schum peter ob­
served the actual developm ent o f the leading econom ies, he cam e to see the large
corporation as the innovating firm , engaged in what he called a process o f creative
destruction” ; the creation o f new m odes o f p rodu ctive tran sfo rm ation destroyed
existing m odes that had themselves been the result o f in n ovative enterprise in
the past.
In Capitalism , Socialism, and Democracy, first published in 1942, Schumpeter
( 19 5 0 :118 ,13 2 ) argued that “ technological 'progress’ tends, through systemization
and rationalization o f research and m anagem ent, to becom e m ore effective and
sure-footed” as it is undertaken as “ the business o f team s o f trained specialists who
turn out what is required and make it w ork in predictable w ays ” In a series o f major
works, Alfred Chandler (1962, 1977, 1990) docum ented the rise o f the managerial
corporation in the United States from the last decades o f the nineteenth century, the
evolution o f its m ultidivisional structure from the 1920s, and the emergence of
m anagerial enterprise in Britain and Germ any. In The Theory o f the Growth of the
Firm , first published in 1959, Edith Penrose (1995) conceptualized the modem
corporate enterprise as an organization that adm inisters a collection o f human
and physical resources. People contribute labor services to the firm , not merely as
individuals, but as m em bers o f teams who engage in learning about how to make
best use o f the firm ’s productive resources— including their ow n.
At any point in time, this learning endow s the firm w ith experience that gives it
productive opportunities unavailable to other firm s, even in the sam e industry, that
have not accum ulated the same experience. The accu m u lation o f innovative experi­
ence enables the firm to overcom e the “ m anagerial lim it” that in the theory o f the
optim izing firm causes the onset o f increasing costs and constrains the growth of
the firm (Penrose 19 95- chs. 5,7, and 8). The innovating firm can transfer and reshape
its existing productive resources to take advantage o f new m arket opportunities.
Each m ove into a new product m arket enables the firm to utilize unused productive
services accum ulated through the process o f organizational learning. These unused
productive services can provide a foundation for the grow th o f the firm , through
both in-house com plem entary investments in new product developm ent and the
acquisition o f other firms that have already developed co m p lem en tary productive
services.
 THE INNOVATIVE FIRM S3

*E lbaum , B., and L azo nick , W. (eds.) (1986), The Decline of the British Economy>Oxford:
Oxford University Press.
Fa r n ie > D. (1990), uThe Textile Machinery-Making Industry and the World Market,
1870—1960,” Business History 32(4): 150-70.
G a l b r a it h , J. K. (1967), The New Industrial State, Boston: Houghton Mifflin.
G im e in , M., D a s h , E., M unoz , L., and S u n g , J. (2002), “ You Bought They Sold/' Fortune
146(4): 64-72.
G om pers , P., L erner , J., and Sch a r fstein , D. (2003), “ Entrepreneurial Spawning: Public
Corporations and the Genesis o f New Ventures, 1986-1999,” N BER Working Paper 9816,
July.
G ordon , A. (1985), The Evolution of Labor Relations in Japan: Heavy Industry; 1853-1955,
Cambridge, Mass.: Harvard University Press.
* H a n n a h , L. (1983), The Rise of the Corporate Economy: The British Experience, 2nd edn.,
London: Methuen.
H o u n sh ell , D. (1984), From the American System to Mass Production, 1800-1932, Baltimore:
Johns Hopkins University Press.
H u b e r m a n , M. (1996), Escape from the Market: Negotiating Work in Lancashire, Cambridge:
Cambridge University Press.
J a ik u m a r , R. (1989), “ Japanese Flexible Manufacturing Systems: Impact on the United
States,” Japan and the World Economy 1(2): 113-43.
K ogut , B., and Z a n d e r , U. (1996), “ What Firms Do? Coordination, Identity, and Learning,”
Organization Science 7: 502-18.
K o rtu m , S., and L e r n e r , J. (2000), “Assessing the Contribution o f Venture Capital to
Innovation,” Rand Journal of Economics 31(4): 674-92.
La zo n ic k , W. (1986), “ Strategy, Structure, and Management Development in the United
States and Britain,” in K. Kobayashi and Н. Morikawa (eds.), Development of Managerial
Enterprise, Tokyo: University o f Tokyo Press, 101-46.
------ (1990), Competitive Advantage on the Shop Floor, Cambridge, Mass.: Harvard University
Press.
------ (1999), “ The Japanese Economy and Corporate Reform: What Path to Sustainable
Prosperity?” Industrial and Corporate Change8(4): 607-33.
------ (2001), “ Organizational Learning and International Competition: The Skill-Base
Hypothesis,” in W. Lazonick and M. O'Sullivan (eds.), Corporate Governance and Sustain­
able Prosperity, Basingstoke: Palgrave, 37- 17*
------ (2002a), “ Innovative Enterprise and Historical Transformation,” Enterprise & Society
3(i): 35- 54-
------ (2002b), “ Innovative Enterprise, The Theory of,” in M. Warner, (ed.), International
Encyclopedia of Business and Management, 2nd edn., Stamford, Conn.: Thomson Learning,
3055-76.
-----(2002c), “The US Industrial Corporation and The Theory of the Growth ofthe Firm,” in
Christos Pitelis (ed.), The Growth of the Firm: The Legacy o f Edith Penrose, Oxford: Oxford
University Press, 249-77.
— —(2003), “The Social Foundations of Innovative Enterprise,” Franco Momigliano Lec­
ture, Institute per la Cultura e la Storia dTmpresa, Terni, Italy, July 10.
— — (2004), “ Corporate Restructuring ” in S. Ackroyd, R* Batt, R Thompson, and R Tolbert
(eds.), The Oxford Handbook o f Work and Organization, Oxford: Oxford University Pies
(forthcoming).
34 WILLIAM LAZONICK

along the cum ulative paths that are the fou n d ation s o f its distinctive competitive
success. The perspective em phasizes the role o f h u m an agency in determining
whether and how the enterprise accum ulates in n ovative capability, and thus adds
an explicitly social dim ension to w ork on dynam ic capabilities. Specifically,
strategic control determines h ow strategic decision m akers choose to build on
“ asset position s"; financial com m itm ent determ ines w hether the enterprise will
have the resources available to it to persist along an “ evo lu tio n ary path" to the
point where its accum ulation o f innovative capability can generate financial returns;
and organizational integration determ ines the structure o f incentives that charac­
terize “ organizational processes" that can tran sform in d ivid u al actions and individ­
ual capabilities (including those o f strategic m anagers) into collective learning.
O f central im portance to the accum ulation and tran sfo rm atio n o f capabilities in
knowledge-intensive industries is the skill base in w hich the firm invests in pursuing
its innovative strategy. W ithin the firm , the d ivisio n o f lab or consists o f different
functional specialties and hierarchical responsibilities. A t any point in time a
firm's functional and hierarchical division o f lab or defines its skill base. In the
effort to generate collective and cum ulative learning, those w h o exercise strategic
control can choose how to structure the skill base, in clu d in g h o w em ployees move
around and up the functional and hierarchical d ivisio n o f lab or over the course of
their careers. At the same tim e, however, the organization o f the skill base will be
constrained by both the particular learning requirem ents o f the industrial activities
in which the firm has chosen to com pete and the alternative em ploym ent oppor­
tunities o f the personnel for w hom the firm m ust com pete.
In cross-national com parative perspective, the skill base that enterprises employ
to transform technologies and access m arkets can v a ry m arkedly even in the same
industrial activity during the same historical era, w ith different innovative out­
comes. Precisely because innovative enterprise depends on social conditions, the
developm ent and utilization o f skill bases that occu r in one institutional environ­
ment m ay not, at a point in time at least, be possible in another institutional
environm ent. M oreover, even within the sam e in d u stry and sam e nation, dynamic
capabilities that yielded innovative outcom es in one historical era m ay becom e static
capabilities that inhibit innovative responses in a subsequent historical era.
I h e innovative firm requires that those w ho exercise strategic control be able
to recognize the com petitive strengths and weaknesses o f their firm 's existing
skill base and, hence, the changes in that skill base that w ill be necessary for an
innovative response to com petitive challenges. These strategic decision makers must
also be able to mobilize com m itted finance to sustain investm ent in the skill base
until it can generate higher quality, lower cost produ cts than were previously
available. As the follow ing com parative-h istorical syntheses illustrate, given stra­
tegic control and financial com m itm ent, the essence o f the innovati ve firm is the
organizational integration of a skill base that can engage in collective and cumulative
learning.
 T H E I N N O V A T I V E FIRM.' 33

From the 1980s m an y business sch ool academ ics, w o rk in g in the strategy area,
cited Penrose's 1959 b o o k as an intellectual fo u n d atio n for a “ reso u rce-b ased ” view
o f the firm . R esou rce-b ased th eo ry focused on the characteristics o f valuab le re­
sources that one firm possessed and that co m p etitor firm s fo u n d it d ifficult to
im itate. R esou rce-b ased theory, how ever, pro vid ed no perspective on w h y and
h o w som e firm s rather than others accu m u lated valuable and in im itab le resources,
o r indeed w h at m ad e these resources valu able and in im itab le (see Lazonick 2002a).
In depen d en tly o f the resou rce-based perspective, how ever, R ich ard N elson and
Sid n ey W in ter (1982) fashion ed a th eo ry o f the persistence o f the large in du strial
co rp o ratio n based o n o rgan izatio n al capabilities, characterized b y tacit know ledge
and em bed ded in o rgan izatio n al routines, thus ad d in g a cu m u lative d im en sio n to
the th e o ry o f the firm . D raw in g on a h igh ly eclectic set o f sources fro m a n u m b er o f
disciplines, B ru ce K o gu t and U d o Z a n d er (1996: 502) argu ed that “ [f firm s are
o rganizations that represent social know ledge o f co o rd in atio n and learning,” thus
em ph asizing the collective d im en sio n in the th eo ry o f the firm .
In “ W h y D o F irm s D iffer, and H o w D oes It M atter?” N elson (19 9 1:7 2 ) argu ed that
“ it is o rgan izatio n al differences, especially differences in abilities to generate and
gain fro m in n o vatio n , rather than differences in co m m an d over p articu lar tech n o lo ­
gies, that are the sou rce o f durable, n ot easily im itable, differences a m o n g firm s.
Particular tech n ologies are m u ch easier to understand, and im itate, than b ro ad er
firm d yn am ic capabilities.” D a v id Teece, G a ry Pisan o, an d A m y Shuen (1997: 516)
defined “ d yn a m ic capabilities as the fir m s ab ility to integrate, b u ild , and reconfigure
internal and extern al com petences to address rap id ly ch an gin g environ m en ts.” T h e y
also argu ed that the firm 's strategy entails ch oo sin g am o n g and co m m ittin g to lo n g ­
term path s o r trajectories o f com petence d evelopm en t (Teece at al. 1997: 524).
W hereas the firm 's asset p o sitio n s determ ine its com petitive advantage at an y
p o in t in tim e an d its e vo lu tio n a ry path constrains the types o f in d u strial activities
in w h ich a firm can be co m petitive, its o rgan izatio n al processes tran sfo rm the
capabilities o f the firm o ver tim e.
W h ile Teece et al. (19 9 7 :5 19 ) stressed the im p o rtan ce o f learn in g processes that are
“ in trin sically so cial an d collective,” their d yn am ic capabilities perspective lacks
social content. T h e fra m e w o rk does n o t ask w h at typ es o f peo p le are able an d w ilin g
to m ake the strategic investm ents that can result in in n o vatio n , h o w these strategic
d ecision m akers m o b ilize the necessary fin an cial resources, and h o w they create
incentives fo r th o se people w ith in the firm 's h ierarch ical and fu n ctio n a l d iv isio n o f
labor to co o p erate in the im p lem en tatio n o f the in n o vative strategy. These qu estio n s
ab ou t the roles o f strategizing, fin an cin g, an d o rgan izin g in the in n o va tin g firm are
have called the “ so cial co n d itio n s o f
at the center o f w h at M a r y O 'S u liv a n and I
in n o vative e n terp rise” perspective (L azo n ick and O 'S u lliv a n 2000; O 'S u lliv a n
2000b; L a z o n ick 2002b).
T h is p ersp ective asks h o w an d u n d er w h at co n d itio n s the exercise of strategic
co n tro l ensures- that the en terprise seeks to g ro w u sin g the e o le e tiv e p ro cesses m i
36 W I L L I A M L A Z O N I C K __________ _____________________________________ _

labor induced new specialized m anufacturing firm s, often fo u n d ed b y craft workers

themselves, to set up in these districts. The grow th o f a district indu ced other firms to
invest in regionally specific com m unication and d istrib u tion facilities for the supply
o f materials, the transfer o f w ork-in-progress across vertically specialized firms, and
the m arketing o f output.
Regional concentration encouraged vertical specialization, w hich in turn eased
firm entry into a particular speciality, thus resulting in h igh levels o f horizontal
com petition. Firm s could be owned and m anaged b y the sam e people, there was no
need to invest in the types o f m anagerial organization that b y the late nineteenth
century were becom ing central to the grow th o f firm s in the U nited States, Germany,
and Japan. In the industrial districts, econom ies o f scale w ere, as M arshall argued,
external, rather than internal, to the firm .
As producers and users o f m achinery, craft w orkers constituted the prim e source
o f innovation in a particular region. O ver tim e they devised increm ental techno­
logical and organizational im provem ents that, th rou gh the local trade press (includ­
ing workers' newspapers) as well as the m ovem ent o f w orkers (especially trained
apprentices) to new em ployers, diffused across firm s in the district. Som e specialized
engineering firm s distinguished themselves through in -h o u se learning. But even the
strongest o f these firm s— for exam ple, the textile m ach in ery firm o f Platt Brothers
based in O ldham — did no in-house R8cD, and from the last h a lf o f the nineteenth
century generated no significant technological innovations. T h eir strength resided
in their em ploym ent o f craft labor that could flexibly produ ce custom ized machines
for m any different types o f users (Farnie 1990).
The im portance o f localized craft labor to the in n ovative capabilities o f local firms
meant that it was the industrial district, and often a particu lar tow n w ithin a district,
not the individual firm , that constituted the learning entity. At the firm level, craft
workers made countless “ strategic" decisions to im p ro ve produ cts and processes.
For both individual firm s and the district as a w hole, the fixed costs o f developing
this source o f innovation were, in historical and com parative perspective, low. At the
same time, craft-oriented em ploym ent system s encouraged a high level o f utilization
o f the plant and equipm ent in place. U nion bargains protected the tenure and
rem uneration o f senior workers who, paid by the piece, w ere w illin g to work long,
hard, and steady. The inducem ent for ju n io r w orkers, typ ically paid tim e wages, was
that they could eventually join the aristocracy o f labor. There is evidence that, within
an industrial district, those localities in which negotiated piece-rate bargains shared
productivity gains between em ployees and em ployers on a stable and equitable basis
saw the fastest growth in productivity and m arket share (Lazon ick 1990: chs. 3-5;
H uberm an 1996).
Based on craft organization, British industrial districts were high ly innovative (see
also Bruland and M owery, this volum e). The fact that it was the industrial district as
a whole, rather than the individual enterprise w ithin it, that w as the innovating
entity gave rise to the notion that differences am ong firm s in an industrial activity
 THE INNOVATIVE FIRM 35

2.3 T h e B r it is h In d u s t r ia l D is t r ic t

In last h a lf o f the nineteenth century, B ritain becam e k n o w n as the “ w o rk sh o p o f the

w o rld /’ B rita in ’s p o sitio n in the w orld eco n o m y ow ed m uch to its m ercantile power,
developed th ro u g h global co m m erce and related w ars w ith oth er lead in g nations
over the p revio u s centuries. M ercan tilism gave B ritish in d u stry access to w orld
pro d u ct m arkets and sources o f raw m aterials, b u t it w as the tran sfo rm ation o f
p ro d u ctio n fro m the late eighteenth cen tu ry that enabled B ritain to em erge as the
w o rld ’s lead in g (and indeed first) indu strial nation.
In the late nineteenth century, B rita in ’s p ro d u ctive p o w er resided in in du strial
districts that, fo r b u ild in g m ach in es and using them to m an u factu re p ro d u cts as
varied as cloth an d ships, possessed an im m ense accu m u latio n o f capabilities.
B eyon d even in g courses at local “ m ech an ics’ institutes,” fo rm al vocatio n al or p ro ­
fessional ed u catio n played no role in the d evelopm en t o f B rita in ’s skilled lab o r force.
N o r did B ritish in d u stry m ake use o f corporate, university, o r govern m en t research
labs to d evelop n ew technology. R egio n ally based o n -th e-jo b apprenticesh ip ar­
rangem ents, th ro u gh w h ich craft w orkers passed on their skills to the next gen er­
ation, constituted in effect the “ n ation al in n o vatio n system ” o f the w o rld ’s first
indu strial econom y.
W hat accoun ts fo r the im p o rtan ce o f the craft w o rk er fo r B ritain ’s in d u strial
leadership? W h ile the m ech an ization o f the fa c to ry w as a central feature o f the
B ritish in d u strial revo lu tio n — and in its tim e a w o n d er o f the w o rld — the stan d ard ­
ization o f m aterials an d the au to m atio n o f m ach in ery that B ritish in d u stry achieved
d u rin g its in d u strial revo lu tio n w ere, in h isto rical retrospect, incipient. Skilled craft
w orkers m ain tain ed critical roles in keeping im perfect m ach in ery in m o tio n and
en surin g h igh levels o f th ro u g h p u t o f w o rk -in -p ro g ress m ad e fro m im perfect
m aterials. W ith in the firm , experien ced w orkers typ ically w ere respon sible fo r
train in g yo u n ger w o rk ers in the craft, su p ervisin g their w o rk , an d co o rd in atin g
the flo w o f w o rk th ro u g h the p ro d u ctio n process. In som e indu stries, the central
em p lo ym en t relatio n to o k the fo rm o f an in tern al su b con tract system ; fo r exam ple,
in the co tto n sp in n in g indu stry, em ployers p aid piece-rates to sen io r w o rk ers,
kn o w n as “ self-actin g m inders,” w h o in turn trained, su p ervised , an d paid tim e
w ages to ju n io r w o rk ers k n o w n as “ piecers” and “doffers” In the m etalw o rk in g
industries, specialized w o rk ers such as “ tu rn ers” and “ fitters” w ere gen erally clas­
sified as “ engineers,” an ap p ellation that in the B ritish context sign ified membership
in the “ lab o r a risto cra cy ” o f skilled p ro d u c tio n w o rk ers (L azon ick 1990; chs. 1- 6 ) .
T h e localized , o n -th e-jo b ch aracter o f skill fo rm a tio n w as the major factor
u n d erlyin g the grow th o f in d u strial districts that m ad e use o f p articu lar specialized
craft skills. A s A lfred Marshall (1961: 271) fa m o u sly put it, in the British industrial
districts “ m ysteries o f the trade b eco m e n o m ysteries; but are as it were in the t i f f In
p e rio d s o f stro n g p ro d u ct-m ark e t d em an d , the read y availability o f specialized е ш і
38 WILLIAM LAZONICK

2.4 T h e US M a n a g e r ia l C o r p o r a t io n

Marshall located the limits to the growth o f the firm in the problem o f succeeding the
original owner-entrepreneur. In The Theory ofEconomic Development, Schumpeter
(1934: 156) concurred using the sam e aphorism as M arsh all, literally clothed in
different garb and specifically identified as a U S p h en om en on : “A n American
adage expresses it: three generations from overalls to overalls ” Critical to this
perspective were two assum ptions: first, that the entrepreneur w as the essence of
the innovative firm , and second, that the integration o f ow n ersh ip and control was a
necessary condition for entrepreneurship. N otw ith stan d in g his ow n important
study o f com parative trends in industrial organization pu blish ed in Industry and
Trade, M arshall (1919) declined to recognize, as ultim ately Schu m peter did, that the
problem o f innovative succession could be resolved b y the separation o f ownership
and co n tro l
Taking place during the same decades in w hich M arsh all w rote his influential
books, the separation o f share ow nership from strategic control w as the essence of
what Chandler (am ong others) w ould call “ the m anagerial revolu tion ” in American
business. D uring this period G erm any and Jap an also experienced managerial
revolutions (Chandler 1990; Chandler et al. 1997; M o rik aw a and Kobayashi 1986;
M orikaw a 1997). M any British firm s, especially in the science-based chemical and
electrical industries also m ade investm ents in m anagerial organ ization , but in such a
constrained m anner that it can hardly be said that a m anagerial revolution occurred
in Britain during the first h alf o f the twentieth century (H an n ah 1983; Lazonick 1986;
Chandler 1990; Owen 2000).
In the United States, the m anagerial revolution began in the 1890s in industries
such as steel, oil refining, m eatpacking, tobacco, agricultural equipm ent, telecom­
m unications, and electric power that ow ner-entrepreneurs had built up over the
previous decades. Wall Street (and especially the firm o f J. P. M o rgan ) organized the
m erger o f the leading com panies, and in the process did w hat w o u ld later become
known as initial public offerings” (IPO s) in order to allow the owner-entrepreneurs
to cash in on their ownership stakes. M an y o f them then retired from active
m anagem ent o f the company. Taking their places in strategic decision-making
positions were salaried m anagers, m ost o f w h o m had them selves been recruited
years or even decades earlier to help build the innovative firm s that they now
controlled. Hence, M arshall’s “ entrepreneurial” lim it to the grow th o f the firm was
overcom e. By the turn o f the century, the separation o f ow nership and control in
m any o f the m ost successful industrial corporations served as a pow erful induce­
ment for bright young, and typically W hite, A n g lo -Saxo n , Protestant, men to
consider careers as corporate executives (Lazonick 1986; O ’ Su llivan 2000a: ch. 3).
Also from the beginning o f the twentieth century, а fo u r-year undergraduate
college degree became im portant for entry into m anagerial careers, and in 1908
 THE INNOVATIVE FIRM 3J

were u n im p o rtan t to econ om ic p erform an ce, and indeed that they could all be
characterized b y d ep ictin g a “ representative firm ” that optim ized subject to given
tech nological and m arket constraints. W ithin the M arsh allian perspective, even
in n o vation at the district level did not require strategic direction , since the industrial
arts w ere “ in the a ir ” In deed, M arsh all (1919: 6 0 0 -1) described the o rgan ization o f
the Lan cash ire cotton textile industry, w ith its h igh degrees o f h orizon tal co m p e ti­
tion and vertical specialization, as “ perhaps the present instance o f concentrated
o rgan isatio n m a in ly a u to m a tic ” Yet ju st as M arsh all w as w ritin g these w ords, the
cotton textile indu stry, w h ich had accoun ted fo r o n e-q u arter o f British exp orts on
the eve o f W orld W ar I, entered into a lo n g -ru n decline fro m w hich it never
recovered, and the other m a jo r B ritish in d u strial districts suffered a sim ilar fate
(E lb au m an d Lazo n ick 1986).
From the late 1970s, how ever, the n o tio n o f the “ M arsh allian in d u strial d istrict” as
a d river o f in n o vative enterprise saw an academ ic resurgence, based on the rap id
grow th d u rin g the 1960s and 1970s o f m an y h igh ly specialized and localized districts
in w hat b ecam e k n o w n as “ the T h ird Italy” (B ru sco 1982; Sabel 1982; Becattini 1990).
O n the basis o f this experience, a n u m b er o f U S academ ics, h ead ed b y C h arles Sabel,
M ich ael P iore, an d Jo n ath an Z eitlin , posited a n ew m o d el o f “ flexible sp ecialization ”
as an alternative to m ass p ro d u ctio n on the U S co rp orate m o d el (P iore and
Sabel 1984; Sabel an d Z eitlin 1985). T h e in d u strial activities o f the districts o f the
T h ird Italy fo cu sed on, am o n g oth er things, textiles, footw ear, and light m ach in ery,
just as the B ritish districts had done. Large n u m bers o f vertically specialized p ro p ri­
etary firm s in w h ich craft lab o r w as a p rim e source o f co m p etitive advantage
p o p u lated each in d u strial activity, and m an y entrepreneurs h ad p re vio u sly been
craft w orkers.
There w ere, how ever, tw o im p o rtan t differences betw een the B ritish in d u strial
districts that M arsh all h ad observed in the late nineteenth cen tu ry and those that
experien ced rap id grow th in the T h ird Italy m ore recently. T h e first difference w as
the extent to w h ich in Italy collective in stitu tion s su p p o rted the in n o vative activities
o f sm all firm s. Seb astian o B ru sco (1992) has em phasized the im p o rtan ce o f the “ red ”
local go vern m en ts in E m ilia -R o m a g n a in p ro m o tin g p o licies to su p p o rt the activ ­
ities o f sm all enterprises, an d in p articu lar in facilitatin g coop eratives that p ro vid e d
these firm s w ith “ real services” related to business ad m in istratio n , m arketin g, and
train in g. W h ile co n su m er co op eratives sp ru n g u p in the B ritish in d u strial d istricts
o f the late nineteenth century, p ro d u cer co op eratives w ere rare. T h e secon d d iffe r­
ence, w h ich b ecam e m o re evident in the 1990s, w as the extent to w h ich , in some
districts an d in so m e in du stries, “ lead in g” firm s co u ld em erge, d raw in g o n the
resources o f the in d u strial d istricts w h ile, th ro u g h th eir o w n in tern al growth*
tra n sfo rm in g the in n o vative ca p a b ility o f the d istricts (see, for exam p le, lelussi
1999). In co n trast, w h en in the first h a lf o f the tw entieth ce n tu ry competitive
challenges co n fro n ted the B ritish in d u strial districts, d o m in a n t firms failed to
em erge t o lead a restru ctu rin g process.
40 W IL L IA M LA Z O N JCK _________ ______ ____________________ _

intervention o f a skilled worker to m ake the parts fit together. A s D avid Homshell
(1984) has shown, it took a century o f investm ent in p ro d u ctive capabilities by many
com panies in m any sectors o f U S indu stry before, d u rin g the b o o m o f the 1920s,
mass production, so defined, becam e a reality. T h e p ro d u ctivity o f the mass-
production enterprise, nevertheless, still relied u p on the stable employment of
“ sem i-skilled” production workers w ho tended h igh -th rou gh p u t, and very expen­
sive, m achinery (Lazonick 1990: chs. 7-8 ).
D uring the Great Depression o f the 1930s, such stable em ploym en t disappeared,
leading sem i-skilled workers at the m ajor m ass p ro d u cers to turn to industrial
unionism (B rod y 1980: ch. 3). The m ajor achievem ent o f m ass-p ro d u ction unionism
in the United States was long-term em ploym ent secu rity fo r so-called “hourly*
workers, with seniority as the governing principle fo r internal p ro m o tio n to higher
pay grades and continued em ploym ent during co m p an y layoffs. In return, these
unionized employees accepted unilateral m anagerial con trol over the organization
o f w ork and technological change. D uring the post-W o rid W ar II decades, produc­
tion workers enjoyed em ploym ent security and rising w ages but they were not in
general integrated with m anagerial personnel into the co m p an y's organizational
learning processes.
The result was that going into the second h a lf o f the tw entieth century US
industrial corporations had pow erful m anagerial organ ization s for developing
new technology. These corporations also had devised arrangem ents with their
unionized labor forces to ensure the high level o f utilization o f these technologies.
In em ploying thousands and in som e cases tens o f thou sand s o f production workers
who were not integrated into the com pany's organizational learning processes,
however, this US m odel o f the innovative firm had a fu ndam ental weakness that,
in the 1970s and 1980s, w ould be exposed in international com petition . The Japanese
in particular w ould dem onstrate the innovative capab ility that could be created by
not only building highly integrated m anagerial organizations, as the Am ericans had
done, but also, as a com plem ent, developing the skills o f sh o p -flo o r workers and
integrating their efforts into the firm's collective learning processes.
Even the m ost insightful o f the theories o f the U S m anagerial corporation could
not, without elaboration, account for the Japanese challenge (Lazon ick 2002c). Both
Penrose (1995) and Chandler (1962 and 1977) focused exclusively on the managerial
organization, as did the influential perspective o f Jo h n Kenneth Galbraith (1967)
with its notion o f the technostructure" as the essence o f the m o d ern firm . Penrose
did not see that, once confronted by the Japanese challenge, the U S managerial
corporation would have to develop the capabilities o f the sh o p -flo o r w orker to make
use o f unused m anagerial resources. Chandler focused on speed or throughput as a
basis for achieving econom ies o f scale and scope, but ignored the role o f the shop-
floor worker in the process o f transform ing high fixed costs into low unit costs, and
hence did not perceive an im portant lim itation o f the U S m anagerial model
(Lazonick 1990).
 THE INNOVATIVE FIRM 39

Harvard University launched the first graduate school in business administration* In

1900 about 2 per cent of 18 -2 4 year olds were enrolled in institutions o f higher
education; in 1930 over 7 per cent; and in 1950 over 14 per cent. By the 1920s the top
managers o f m any large industrial corporations had college degrees. As employers of
university graduates as well as beneficiaries o f university research, big business took
an active role in shaping the form and content o f higher education to meet its needs
for “ knowledge assets” (Noble 1977; Lazonick 1986).
As they expanded, US industrial corporations tended to diversify into newlines o f
business. Capabilities developed for generating goods for one product market could
be used as a basis for gaining entry to new product markets. Moreover, as companies
were successful, they could use internally generated revenues to finance these new
investments. Profitable US corporations generally paid ample dividends to share­
holders, but they still generated enough revenues to invest for the future, including
growing expenditures on R&D (Mowery and Rosenberg 1989: ch. 4).
Besides transform ing technology, a critical role o f the managerial organization
was to gain access to product markets. Without high levels of sales, the high fixed
costs of developing technology and investing in production facilities would have
simply resulted in high levels of losses. The building of national transportation and
communications infrastructures— themselves largely put in place by managerial
enterprises— created the possibility for manufacturing enterprises to sell on mass
markets. To take advantage of this opportunity, however, the industrial corporations
had to make complementary investments in distribution capabilities, including sales
personnel, sales offices, advertising, and in some cases even customized transporta­
tion facilities. As Chandler (1990) has shown, from the late nineteenth century,
a “three-pronged” investment in production, distribution, and management was a
necessary condition for the growth of the industrial enterprise.
If the so cial condition for the growth o f the US industrial corporation was an
integrated managerial organization, a distinguishing feature of the same corpor­
ation was a sharp organizational segmentation between salaried managers and what
became known as “ hourly” workers. This segmentation had its roots in the first half
of the nineteenth century when in d u strial managers faced a skilled labor force that
w as h ig h ly m o b ile not o n ly fro m one firm to anoth er but also fro m on e o ccu p atio n
and o n e lo cality to another. In contrast, in B ritain the local p o o ls o f specialized craft
labor generated b y apprenticeship system s meant that employers had access to
am ple su p p lies o f skilled labor, even in b o o m s. A s a result, there w as m u ch less
pressure in B rita in th an in the United States fo r m an agers to invest in the d evelo p­
m ent o f sk ill-d isp lacin g tech n ologies. In the U n ited States, b u t n ot B rita in , firm s
integrated tech n ical specialists in to th eir m an agerial organ izatio n s fo r p recisely that
p u rp ose. H en ce the em ergen ce b y the m id -n in eteen th cen tu ry o f the distinctive
“American system o f m an u factu res” (Hounshdl 19^4* As. 1-2).
' ■ ■ 'The k e y to th is system w as the m ass p m d u A o n o f standaidized» р ш ш ш »
ш р п е е г е і p a rts th at c o u ld be u sed in terch an geab ly in a p m d u c t w ith o u t tile
42 WILLIAM LAZONICK __________ ________________________________ _

1997), Taking control o f strategic decision m aking w ere "th ird -ra n k executives,”
prim arily engineers plucked from the ranks o f m iddle m anagem ent to take leader­
ship positions o f com panies whose challenge was to find n o n -m ilita ry markets for
their com panies' accum ulated capabilities.
W ith the reopening o f the stock m arket in 1949? these yo u n g and ambitious
executives feared that the new public shareholders m igh t jo in forces to demand
their traditional rights as owners. To defend them selves against these outside
interests, the com m unity o f corporate executives engaged in the practice o f cross­
shareholding. Com m ercial banks and industrial com panies to o k equities off the
market by holding each other's shares. T h ou gh not contractual, cross-shareholding
was sustained by the willingness o f the entire Japanese business co m m u n ity to accept
that one com pany would not sell its shareholdings o f anoth er com pany.2 By 1975,
according to its broadest, and m ost relevant, defin ition as stock in the hands o f such
stable shareholders, cross-shareholding represented 60 per cent o f outstanding
stocks listed on the Tokyo Stock Exchange. It peaked at 67.4 p er cent in 1988, but
by 2000 had declined to 57.1 per cent, m ainly because the beleaguered banking sector
had been forced to reduce their shareholdings.
D uring the "era o f high-speed grow th" from the early 1950s to the early 1970s,
m ost o f the financial com m itm ent o f Japanese com panies cam e from bank loans,
with the com panies' d eb t-eq u ity ratios often at 6 : 1 or 7 : 1 . Each m ajor industrial
com pany had a "m ain ban k" whose job it was to convince other banks to join it in
m aking loans to the com pany and to take the lead in restru ctu ring its client company
should it fall into financial distress. Som e econom ists (e.g. A o k i and Patrick 1994)
have accorded the m ain banks a m ajor role in m o n ito rin g the behavior o f Japans
corporate managers. In funding the grow th o f Japanese com panies, however, the
Japanese banks were relatively passive agents o f govern m en t developm ent policy,
with "overloans" being m ade by the Bank o f Jap an to its m em b er banks for provid­
ing highly leveraged finance to grow ing industrial com panies. Japanese banks, that
is, played a critical role in providing financial com m itm ent, but no significant role in
the exercise o f strategic control.
Integrated organizations o f m anagers and w orkers, not financial interests, moni­
tored the behavior o f the top executives o f Japanese co rp oratio n s (Lazonick 1999)’
The m ain m ode o f achieving this organizational integration was the lifetime em­
ploym ent system, which extended from top executives to m ale (but not female)
shop-floor workers. The origins o f the lifetim e em ploym ent system can be found in
the widespread em ploym ent in industry ot university graduates as salaried technical
and adm inistrative personnel during the early twentieth cen tu ry (Yonekawa 1984)*
Som e com panies extended the prom ise ot lifetim e em ploym ent to shop-floor
workers as well when dire econom ic conditions and dem ocratization initiatives of
the late 1940s had given rise to a m ilitant labor m ovem ent. The goal o f the new
industrial unions was to im plem ent "p ro d u ctio n co n tro l": the takeover o f idle
factories so that workers could put them into operation and earn a living (Gordon
 THE INNOVATIVE FIRM 41

2.5 T h e J a p a n e se C h a l l e n g e

Within the new structure of cooperative industrial relations that emerged out the
conflicts of the depression years, US industrial corporations were able to take
advantage of the post-World War II boom to re-establish themselves as the world’s
pre-eminent producers of consumer durables such as automobiles and electrical
appliances and related capital goods such as steel and machine tools* With the help of
U S go vern m en t research su p p o rt and contracts, U S com panies also becam e the
leaders in the computer and semiconductor industries.
In the 1970s and 1980s, however, Japanese companies challenged the US industrial
corporations in the very mass-production industries— steel, memory chips,
machine tools, electrical machinery, consumer electronics, and automobiles— in
which even as late as the 1960s US corporations seemed to have attained an insur­
mountable competitive advantage. During the 1950s and 1960s many Japanese
companies had developed innovative manufacturing capabilities, often on the
basis o f tech nologies b o rro w ed fro m abroad to p ro d u ce m ain ly for the h om e
market. As Japanese exports to the United States increased rapidly in the last half
of the 1970s, many observers attributed the challenge to the lower wages and longer
working hours that prevailed in Japan. By the early 1980s, however, with real wages in
Japan continuing to rise, it became clear that Japanese advantage was based on
superior capabilities for generating higher quality, lower cost products*
The three social institutions that, in combination, formed the foundation for
Japan’s remarkable success were cross-shareholding, the main b an k system, and
lifetime employment. Cross-shareholding provided the managers of Japanese
industrial corporations with the strategic control to allocate resources to invest­
ments that could generate higher quality, lower cost products. The main bank system
provided these companies with levels of fin an cial commitment that permitted them
to sustain the innovation processes until they could generate returns, first on home
and then on foreign product markets. Given this financial support for strategic
industries, lifetime employment enabled the companies involved to put in place a
new model of hierarchical and functional integration that enabled them to
mobilize broader and deeper skills bases for collective and cumulative learning
(Lazonick 2001). Let us look briefly at how these institutions became embedded in
the functioning of the Japanese industrial enterprise in the post-World War II
decades.
In 1948 the Supreme Commander for the Allied Powers (SCAP)—the occupation
authority in Japan—began the dissolution of the zaibatsu, the giant holding com­
panies that had dominated the Japanese economy from the Meiji era of the late
nineteenth century to World War II. The dissolution process not only dispossessed
the families that owned the zaibatsu but also removed from office the top im rnge-
meut layers of the zaibatsu holding companies and major affiliated firms (Могікаші
44 WILLIAM LAZONICK

developm ent cycles. M uch o f the discussion o f fu n ctio n al integration focused un its
role in “ new product developm ent" in international co m p arative perspective, with,
as Clark and Fujim oto (1991) showed for the au tom o b ile industry, the U S manager­
ial corporation perform ing quite poorly.
Given that the innovative pow er o f the U S in d u strial co rp o ratio n resided in its
integrated m anagerial organization, w hy should it have suffered from functional
segm entation in com petition w ith the Japanese? O ne reason w as that, given the
hierarchical segm entation o f sh op -floor activities fro m organizational learning
processes in US com panies, U S engineers were not forced to com m unicate across
their disciplines to solve “ real-w orld" m anu factu ring problem s. A n oth er had to do
with the increasing interfirm m obility o f U S engineers fro m the 1960s— mobility
that, as we shall see, was related to the rise o f the "N e w E c o n o m y " high-tech firm.
The prospects for interfirm m o b ility gave scientists an d engineers an interest in
developing their reputations am ong their peers w ith in their particular area of
specialization, even if it detracted from integrating their specialist knowledge across
functional areas within the particular firm for w hich they w ere w orking. By contrast,
in the Japanese firm both the hierarchical integration o f m anagers and workers and
low levels o f interfirm m obility o f engineering person nel fostered functional
integration.
The evolution o f the sem iconductor in d u stry pro vid es a v iv id exam ple o f the
com petitive power, but also the lim its, o f Japanese o rgan ization al integration. From
the late 1970s the Japanese m ounted a form idable com petitive challenge to US
producers in dynam ic random access m em o ry (D R A M ) chips, forcing most US
com panies, including Intel, to w ithdraw from the m arket after 1985. Already a
powerhouse in sem iconductors before the Japanese challenge, Intel reemerged
even stronger in the 1990s as the leader in m icroprocessors, a product in which it
was the pioneer in the early 1970s and for w hich d u rin g the 1980s it secured the
franchise for the IB M PC and the subsequent IB M clones (B u rgelm an 1994).
O rganizational integration was critical to the Japanese challenge in DRAM s. As
Daniel O kim oto and Yoshio Nishi (1994) have show n, the m ost critical interactions
in product and process developm ent in Japanese sem icon d u ctor com panies were
between personnel in divisional R8cD labs and facto ry engineering labs, with
engineering capability being concentrated in the fa cto ry labs. T h ey argue that in
Japan hands-on m anufacturing exp erien ce. . . is alm ost a requirem ent for upward
career and post-career m obility [whereas] [i]n the U nited States, by contrast,
m anufacturing engineers carry the stigm a o f being second-class citizens" (Okimoto
and Nishi 1994:195).
Value added in m icroprocessors is in the design that determ ines the use of the
product, an activity for which US skill bases in sem icon d u ctors were m ore suited.
Value added in m em ory chips is in process engineering that reduces defects and
increases chip yields, an activity for which Japanese skill bases in semiconductors
were m ore suited. By the 1980s Japanese com panies such as Fujitsu, Hitachi, and
 THE INNOVATIVE FIRM 43

1985). Lead in g com pan ies such as Toyota, Toshiba, and Hitachi fired militant
w orkers and created enterprise u nions o f w h ite-collar (technical and adm inistra­
tive) and b lu e-collar em ployees. Forem en and su p erviso rs were members o f the
enterprise un io n s, as w ere all u niversity-edu cated personnel, fo r at least the first ten
years o f em p lo ym en t b efore they m ade the official tran sition into “ m anagem ent.”
The m ost im p o rtan t achievem ent o f enterprise u n io n ism w as the in stitu tio n al­
ization o f lifetim e em p lo ym en t, a system that, w hile not con tractu ally guaranteed,
gave w h ite-co llar and b lu e-collar w ork ers em p lo ym en t security, at first to the
retirem ent age o f 55, then fro m the 1980s to the age o f 60, and cu rren tly (in
tran sition ) to the age o f 65 (Sako and Sato 1997). T h is em p lo ym en t secu rity both
w on the co m m itm en t o f the w o rk er to the co m p an y and gave the co m p an y the
incentive to d evelop the p ro d u ctive capabilities o f the w orker. The system d id not
differ in p rin cip le fro m the organ izatio n al integration o f technical and a d m in is­
trative em ployees that w as at the heart o f the U S m an agerial revolu tion , except in one
extrem ely im p o rta n t respect. In the U nited States there w as a sharp segm entation
betw een salaried m an agers and sh o p -flo o r w orkers, w hereas the Japan ese c o m ­
panies o f the p o st-W orld W ar II decades integrated sh o p -flo o r w ork ers into a
co m p an y-w id e process o f organ ization al learning.
T h ro u g h their engagem ent in processes o f cost red u ction , Japanese sh o p -flo o r
w orkers w ere co n tin u o u sly involved in a m o re general process o f im p ro vem en t of
products and processes that, b y the 1970s, enabled Japanese com pan ies to em erge as
w orld leaders in fa c to ry a u to m atio n (Ja ik u m a r 1989). B y the early 1990s the stock o f
robots in Jap an ese factories w as o ver seven tim es that o f the U n ited States. Also o f
great im p o rtan ce w as the ab ility o f Japan ese m anu factu rers to elim inate w aste in
p ro d u ctio n ; b y the late 1970s, fo r exam ple, Japan 's com petitive advantage in televi­
sion sets w as n ot in lab o r costs o r even scale econ om ies but in a savin gs o f m aterials
costs (O w en 2000: 278; Fagerberg and G o d in h o in this vo lu m e). T h is p ro d u ctive
tran sfo rm atio n b ecam e p articu la rly im p o rtan t in in tern atio n al co m p etitio n in the
1980s as Jap an ese w ages ap p ro ach ed the levels o f those in N o rth A m erica and
W estern E u ro p e and, especially fro m 1985, as the valu e o f the Jap an ese yen d ra m a t­
ically strengthened. D u rin g the 1980s an d 1990s, in flu en ced b y n o t only Japan s
exp ort p e rfo rm an ce b u t also the im p act o f Jap an ese direct investm ent in N o rth
A m erica an d W estern E u ro p e, m a n y W estern com pan ies sou gh t, w ith varying
degrees o f success, to im p lem en t Japan ese h igh -q u ality, lo w -co st m ass-p ro d u ctio n
m ethods.
D u rin g the 1980s m o st W estern analyses o f the sources o f Japanese competitive
advantage fo cu sed o n the h ierarch ical in tegration o f the sh o p -flo o r w o rk er into the
o rgan izatio n al learn in g process. B y the early 1990s, how ever, as Japanese companies
captured h ig h er valu e -a d d e d segm ents o f the p ro d u cts m arkets in which they
co m peted , the em p h asis shifted to the role o f “ cro ss-fu n ctio n al management,”
“ c o m p a n y -w id e q u a lity co n tro l ” o r “ co n cu rren t en gin eerin g” in generating not
o n ly lo w er cost b u t also h igh er q u a lity p ro d u cts within highly accelerated product
4б W I L L I A M L A Z O N I C K _________ ________ _____ __________________________

chip, with Fairchild Semiconductor and Texas Instruments in the forefront of

creating the technology that would become the standard of the semiconductor
industry. Through the early 1960s the US government provided virtually all of the
demand for semiconductors. From the second half of the 1960s, however, a growing
array of commercial opportunities for electronic chips induced the creation of
semiconductor startups. A new breed of venture capitalist, many with prior man*
agerial or technical experience in the semiconductor industry, backed so many
semiconductor startups clustered in the region around Stanford University that by
the early 1970s the district was dubbed Silicon Valley. Innovation in semiconduct­
ors, and especially the development of the microprocessor—in effect a computer on
a chip—created the basis for the emergence of the microcomputer industry from the
late 1970s, which in turn resulted in the enormous growth of an installed base of
powerful “hosts” in homes and offices that made possible the Internet revolution of
the 1990s.
As AnnaLee Saxenian (1994) has show n, intense, and often inform al, learning
networks that transcended the boundaries o f firm s con trib u ted to the success of
Silicon Valley. Like the M arshallian industrial districts o f a cen tu ry earlier, there is no
doubt that, in Silicon Valley, “ the m ysteries o f the t r a d e . . . w ere in the a ir ” But in its
strategy, finance, and organization, the N ew E co n o m y business m odel that emerged
in Silicon Valley differed significantly from the M arsh allian indu strial district. Of
particular im portance was the extent to w hich in Silicon Valley organizational
learning occurred within the firm , enabling som e p articu larly innovative firms
that grew to em ploy tens o f thousands o f em ployees to d rive the development of
the region. In its early stages this organizational learning tended to be backed by
venture capital, a m ode o f finance that through its success in Silicon Valley from the
1960s evolved into an industry in its ow n right. A lso o f great importance in
supporting the developm ent o f technology and the edu cation o f personnel available
to firm s in this high-tech industrial district w ere state fu n d in g and universities,
institutions that for a century had been central to the U S m anagerial m o d el
The founders o f new IC T firms were typically engineers w ho had gained special­
ized experience in existing IC T firms, although in som e cases they were university
faculty mem bers intent on com m ercializing their academ ic know ledge. While some
o f these entrepreneurs came from existing O ld E co n o m y com panies, where it was
often difficult for their new ideas to get internal backing, N ew E co n o m y companies
themselves becam e increasingly im portant as a source o f n ew entrepreneurs who left
their current em ployers to start a new firm (G o m p ers et al. 2003). Typically the
founding entrepreneurs o f a N ew E con om y startup sought com m itted finance from
venture capitalists with w hom they shared not o n ly ow nership o f the com pany but
also strategic c o n tro l Besides sitting on the board o f directors o f the new company,
the venture capitalists w ould generally recruit profession al m anagers, w ho would be
given com pany stock along with stock options, to lead the tran sfo rm ation o f the firm
from a new venture to a going concern. This stock-based com pensation gave these
N EC w ere able to achieve yield s in the p ro d u ction o f D R A M s that were 40 per cent
higher than the best U S com panies.
In the 1990s the Japanese e co n o m y as a w hole has stagnated, to the point where
many Western observers now blame its unique institutional framework, still largely
intact, for its lack of innovation. Yet, in industries such as electronics and automo­
biles, Japanese companies such as Sony and Toyota, among many others, remain
leading innovators in those types of products in which, as during the previous
decades, their integrated skill bases gave them international competitive advantage.
The main microeconomic problems in the Japanese economy are to be found in the
financial system and, relatedly, institutions for creation of new innovative firms.
During the boom of the 1980s the leading Japanese manufacturing companies
were able to reduce their reliance on bank debt, just as the banks were awash with
cash to lend. The banks then channeled fu nds into speculative investments in land
and stocks, thus fuelling the “ bubble economy” of the late 1980s. When the bubble
burst in 1990, the banks w ere saddled w ith m o u n tain s o f bad debt. A lth o u gh most of
this bad debt has now been written off, the banks remain in fragile condition because
m ost o f their loan s are b ein g m ad e to sm aller com pan ies that do n o t have anything
close to the grow th poten tial that w as realized b y m an y Jap an ese co m p an ies in the
previous eras ofhigh-speed growth and export expansion (Lazonick 1999). “Growth
potential,” how ever, is not exogen ou s to the “ social co n d itio n s o f in n o vative enter­
prise,” as illustrated b y the em ergence o f m o re p o w erfu l m odes o f strategy, finance
and organization in the rise of the “New Economy” model of the innovative firm in
the United States.

2.6 T h e N ew E c o n o m y M o d e l

During the 1970s and 1980s while Japanese enterprises were challenging established
U S m anagerial corporations in m any industries in which they had been dom inant,
there was a resurgence o f the U S inform ation and com m unications technology
(IC T ) industries, providing the foundation for what by the last h alf o f the 1990s
became known as the “ N ew Econom y.” Historically, underlying the emergence o f the
N ew Econom y were m assive post-W orld W ar II investments by the U S governm ent,
in collaboration with research universities and industrial corporations, in develop­
ing com puter and com m unications technologies.
B y the end o f the 1950s, this com bined business-governm ent investment effort
had resulted in not only the first generation o f com puters, w ith IBM as the leading
firm , b u t also the capability o f im bedding integrated electronic circuits on a silicon
48 WILLIAM LAZONICK

com pany taking on any long-term debt. N evertheless, w ith the burstin g o f the New
Econom y bubble from m id-2000, C isco spent b illion s o f dollars repurchasing its
ow n stock to support its sagging stock price (C arpen ter et al. 2003). Even during the
boom , when stock prices were rising, the extent to w h ich N ew E co n o m y companies
issued stock to make acquisitions and com pensate em ployees m eant that some of
them spent billions o f dollars on stock repurchases; d u rin g 19 9 7-20 0 0 , for example,
InteEs stock repurchases totalled $18.8 billion and M ic ro so ft’s $13.4 billion. By way of
com parison, over these years Intel s total expenditures on R8cD were $14.2 billion,
while M icrosoft’s were $11.2 billion.
As in the cases o f Intel, M icrosoft, and Cisco, b y the end o f the twentieth century a
num ber o f N ew Econom y com panies had grow n to be fo rm id ab le grow ing concerns
(Lazonick 2004). In 2002 the top 500 U S-based com panies b y sales included twenty
IC T firms founded no earlier than 1965 that had been neith er sp u n -o ff from nor
merged with an Old Econom y firm . These tw enty com pan ies h ad revenues ranging
from $35.4 billion for Dell C om puter to $3.0 b illio n fo r C o m p u ter Associates
International, with an average o f $10 .4 b illion. T h eir h eadcounts ranged from
78,700 for Intel to 8,100 for Q ualcom m , w ith an average o f 30,084, up from an
average for the same twenty com panies o f 6,347 in *993* N in e o f these twenty
com panies (and seven o f the top ten) were based in Silicon Valley, another two in
Southern California, and the other nine in eight states aro u n d the country. Compaq
Com puter, the forty-sixth largest U S com pany in 2001 w ith $33.6 billion in sales and
70,950 employees, would have been high up on this list in 2002 had it not been
acquired by Hewlett-Packard.
M any o f these large New Econom y com panies have becom e im portan t contribu­
tors to the patenting activity o f U S-based corp oration s. Sam u el K ortum and Josh
Lerner (2000) have shown that in the first h a lf o f the 1980s a sharp decline in
patenting by U S corporations was counterbalanced b y a m assive increase in early-
stage venture-capital disbursements. But from the last h a lf o f the 1980s patenting
picked up again, in part because it becam e im p o rtan t to the com petitive strategy of
high-growth N ew Econom y firms. In 2001 Intel was eighteenth in the num ber o f US
patents issued to all com panies, and seventh am o n g U S-b ased com panies. Ahead of
Intel were not only Old Econom y com panies such as IB M , Lucent Technologies,
General Electric, and H ewlett-Packard but also tw o m u ch sm aller, but still sizeable,
N ew Econom y sem iconductor com panies, M icro n Technology, founded in 1978
in Idaho, in fourth place, and Advanced M icro D evices (A M D ), founded in
Silicon Valley in 1969, in fourteenth place. In 2002 A M D w as the 535th largest US
com pany by sales and had 12,146 em ployees, w hile M icro n w as 554th and employed
18,700.
Innovative New Econom y com panies have tended to grow large b y upgrading and
expanding their product offerings within their m ain lines o f business, and thus far at
least have not engaged in the indiscrim inate diversification into unrelated technolo­
gies and markets that characterized, and ultim ately und erm in ed the perform ance of,
 THE INNOVATIVE FIRM 47

m anagers a p o w erfu l financial incentive to develop the in n ovative capabilities o f the

co m p an y to the p o in t where it cou ld do an IP O or private sale to an established
com pany. But, both b efore and after m akin g this tran sition , their tenure w ith, and
value to, the co m p an y depended on their m anagerial capabilities, not their fractional
ow nership stakes.
Key to m ak in g this tran sition fro m new venture to go in g concern w as the
o rgan ization al integration o f an exp an d in g b o d y o f technical and m anagerial
“ talent.75 Stock o p tio n s becam e an im p o rtan t m ode o f com pen sation , usu ally as a
partial substitute fo r cash salaries, for attractin g these h igh ly m o b ile people to the
startup and retain in g their services. T h e u n d erlyin g stock w o u ld becom e valuable if
and w hen th ey to o k the fo rm o f p u b licly traded shares. S h o rten in g the expected
period betw een the lau n ch o f a co m p an y and its IP O w as the practice o f m ost
ven ture-b acked h igh -tech startups o f go in g pu blic on the N A SD A Q exchange
(foun d ed in 19 7 1), w ith its m u ch less stringent listing requirem ents than the O ld
E co n o m y N ew Y o rk Sto ck Exchange. I f and w hen the firm did an IP O o r w as
acquired b y an o th er p u b licly listed com pany, the venture capitalists co u ld sell
their sh areh old in gs on the stock m arket, thereby exitin g fro m their investm ents in
the firm , w hile entrepreneu rs cou ld also tran sfo rm so m e o r all o f their ow n ersh ip
stakes into cash. W ith the co m p an y's stock being p u b licly traded, em ployees w h o
exercised their stock o p tio n s co u ld easily tu rn their shares into cash.
D u rin g the 1980s an d 1990s the liberal use o f stock as a co m p en satio n currency,
not o n ly fo r top executives as had been the case in O ld E c o n o m y com pan ies since the
1950s, b u t also fo r a b ro ad base o f n on -execu tive person nel becam e a d istin ctive
feature of New Economy firm s. For example, Cisco Systems, which grew from about
200 employees at the time of its IPO in 1990 to 38,000 em ployees in 2001, awarded
stock options to all of its employees, so that by 2001 stock options outstanding
accounted for over 14 per cent of the company's total stock ou tstan d in g. Since C isco
did hardly any of its own manufacturing—another distinctive characteristic of many
New Economy “ systems integrators"—the people in the skill base to whom these
o p tio n s w ere aw arded w ere alm ost all h ig h ly educated em ployees w h o w ere p o te n ­
tiallyhighly mobile on the labor market.
Besides using their own stock as a compensation currency, during the 1990s so m e
New Economy companies grew large by using their stock, instead o f cash, to acq u ire
other, smaller and typically younger, New Economy firms in order to gain access to
new technologies and markets. Cisco mastered this growth-through-acquisition
strategy; from 1993 through 2002 Cisco made seventy-eight acquisitions (forty-
one o f which were during 19 9 9 -2 0 0 0 , the peak years o f the New Economy boom)»
with stock providing the currency for over 98 per cent o f the total v alu e o f these
acquisitions.
At the same time Cisco conserved cash by paying no dividends,, a mode o f
financial commitment that also distinguished Mew .Economy f e w Old Іеоаощ у
companies. As a result, Cisco's astonishing growth in the 199-os occurred without the
50 WILLIAM L A Z O N I C K

2.7 U n d e r st a n d in g t h e In n o v a t iv e
F ir m : Im p l ic a t io n s f o r T h eo r y

This chapter has illustrated that the social characteristics o f the innovative firm have
varied m arkedly over time and across institutional environ m en ts. To study the
innovative firm in abstraction from the particu lar social co n d itio n s that enable it
to generate higher quality, lower cost products is to forgo an understanding o f why it
became innovative in the first place and h ow its in n o vative capabilities may be
rendered obsolete. A com parative-h istorical analysis enables us to learn from the
past and provide w orking hypotheses for on goin g research.
First, the com parative-historical experience o f the in n o vative firm suggests that,
contrary to a com m on belief that has persisted since the tim e o f M arshall, the form of
firm ownership is not the critical issue for u nd erstand in g the type o f strategic control
that supports innovative enterprise. Critical are the abilities and incentives o f those
managers who exercise strategic control. W hether they are m a jo rity owners of the
firm, state employees, or employees o f pu blicly listed com pan ies, we need to know
where and how these strategic m anagers gained the experience to allocate resources
to the innovation process, and the conditions under w h ich their personal rewards
depend on the firm's innovative success.
Second, the most fundam ental, if b y no m eans the only, source o f financial
com m itm ent for the innovative firm is to be fo u n d in those fu nds that are generated
by the firm itself. W hen bank finance is used to leverage financial commitment, it
requires close relations between financial institutions and in n ovative firms, as for
example in the Japanese model. In certain tim es and places, the stock market can
provide some well-positioned firm s with financial co m m itm en t. But as a financial
institution, the fundam ental role o f the stock m arket is to p ro vid e liquidity, not
com m itm ent. It enables ow ner-entrepreneurs and venture capitalists to cash out of
their investments, and it enables households to d iversify their savings portfolios so
that they can (hopefully) tap into the yields o f the stock m arket w ithout having to
devote time and effort to understanding the innovative capabilities o f the companies
that have listed their securities on it.
Third, while strategic control and financial co m m itm en t are essential to the
innovative firm, it is organizational integration that determ ines the innovative
capability that the firm actually possesses. The types o f organizational integration
that result in innovation vary across industries and institu tion al environments as
well as over time. The hierarchical and fu nctional d ivisio n s o f lab or that, when
integrated into learning organizations, generated in n o vatio n in the past cannot
necessarily be expected to do so in the future when faced w ith changes in technology,
markets, and com petition changes which to som e extent successful innovation in
itself brings about,
 THE IN N O V A TIV E FIRM 49

m an y lead in g O ld E co n o m y com panies in the 1960s and 1970s* At the same time,
N ew E co n o m y com pan ies have becom e less vertically integrated than Old Economy
com pan ies because equ ip m en t m anu factu rers such as C isco, Dell, and Sun M ic ro ­
system s have focused their investm ent strategies on activities that require o rgan iza­
tional learn in g in their core com petencies, w hile o u tso u rcin g activities that, as is the
case w ith sem ico n d u cto r fab rication , are too expen sive and co m plex to be don e in-
house, or, alternatively, as is the case w ith prin ted circu it board assem bly, have
becom e routine. Som e o f the largest IC T com panies in the U nited States are
upstream electronics co m p o n en ts suppliers, m ost o f w hich are N ew Economy
firm s. A m o n g the top 10 0 0 U S com panies b y sales in 2002 w ere eleven sem ico n ­
d uctor com pan ies, w ith a total em p lo ym en t o f 212,354, ran gin g from Intel w ith its
78,700 em ployees to N v id ia (a specialist p ro d u cer o f grap h ics processors fo u n d ed
in 1993) w ith 1,513 em ployees. T h e w o r ld s five largest contract m an u factu rers—
Flextronics, Solectron , S an m in a -S C I, C elestica, and Jab il C ircu it— to w h o m e q u ip ­
m ent m an u factu rers o u tsou rce the m ass p ro d u ctio n o f printed circu it b oard s
and oth er co m p o n en ts, em ployed a total o f 2 6 0 ,0 0 0 -2 70 ,0 0 0 people at the b egin ­
n ing o f 2003.
T h e severe d o w n tu rn in the IC T in du stries in 20 0 1 and 2002 raised question s
ab ou t the su stain ab ility o f the N ew E c o n o m y m o d e l A m a jo r w eakness o f the N ew
E co n o m y m o d el lay in the huge person al gains, often am o u n tin g to tens o f m illion s
and even h u n d red s o f m illio n s o f dollars, that top executives co u ld reap fro m stock-
based rew ards in a volatile stock m arket. W h en stock prices w ere risin g, executives
h ad stron g p erson al incentives to allocate resources (o r give the app earan ce o f d o in g
so) in w ays that en cou raged the speculative m arket. M a n y o f these allocative d eci­
sions u n d erm in ed the in n o vative capabilities o f the firm s o ver w h ich these execu ­
tives exercised co n tro l (C arp en ter et al. 2003). W h en stock prices b egan falling, the
sam e executives h ad stro n g person al incentives to cash in q u ick ly b y selling stock, so
that th ey m ad e im m en se fo rtu n es (in m o st instances w ith o u t b reakin g the law ) even
as their co m p an ies lost m o n ey and, in m an y cases, strag g led to su rvive (G im em et a l
2002).
A m a jo r p ro b lem fo r som e o f these co m p an ies w as the w a y in w h ich the use o f
stock as a co m b in atio n an d co m p en sation cu rren cy in the N ew E c o n o m y b o o m
affected the role o f the sto ck m arket as a source o f cash (O ’Su llivan 2003). Seven ty
years earlier, in the sto ck m arket b o o m o f the late 1920s, U S co rp o ratio n s had sold
stock at sp eculative prices to p a y d o w n debt o r b olster th eir treasuries, th u s making
th em less fin an cially vu ln erab le w h en the b o o m tu rn ed to bust. In the boom o f the
late 1990s co rp o ra tio n s d id n o t take advan tage o f the sp ecu lative market by selling
stock; i f an yth in g, these co m p an ies pu rch ased sto ck to su p p o rt th eir already inflated
sto ck prices* W h ile em ployees, an d p a rticu la rly h igh -level executives, benefited from
these sto ck p rice .increases, their co m p an ies w ere weakened financially as became
p a in fu lly evid en t fo r m a n y IC T co m p an ies fro m m id -a o o o when the sto ck maxkgt
tu rn ed d o w n .
52 W ILLIAM LAZONICK

N o t e s ________________ _______________ ____________ _ _ _ _ _ _ _

i, For a com parison of managerial corporations of European nations such as Britain, France,
Germany, and Italy with the US model, see Lazonick 2003.
2** When in financial distress, a com pany m ight raise cash b y selling som e o f its cross­
shareholdings to other com panies at the going m arket price but w ith an understanding
that the shares would be repurchased, also at the going m arket price, if and when its
financial condition improved.

R eferen ces

A o ki , M ., and Pa t ric k , H. (eds.) (1994)» T h e Ja p a n e s e M a in B a n k S ystem : Its Relevance for

D evelo p in g a n d T ran sform in g Econom ies? O xford: O xford U n iversity Press.
B ecattin i , G. (1990), “ The M arshallian Industrial District as a Socio-E con om ic Notion,” in
F, Pyke, G. Becattini, and W. Sengenberger (eds.), In d u s tr ia l D istricts a n d Inter-Firm
C ooperation in I t a ly International Institute for Labour Studies, 37-51.
B e l u s s i , E (1999), “ Path-Dependency versus Industrial D ynam ics: A n Analysis o f Two
Heterogeneous Districts,” H u m a n System s M a n a g e m e n t 18 :16 1- 7 4 .
B rody , D. (1980), Workers in In d u stria l A m e ric a : E ssays on the T w en tieth C en tu ry Struggle,
Oxford: Oxford University Press.
B rusco , S. (1982), “ The Emilian Model: Productive D ecentralisation and Social Integration”
C a m b rid g e Jo u rn a l o f Econom ics 6 :16 7 -8 4 .
------ (1992), “ Small Firms and the Provision o f Real Services,” in F. Pyke and W. Sengenberger
(eds.), In d u stria l D istricts a n d L o cal E co n o m ic R e g e n e ra tio n , International Institute for
Labour Studies, 177-96,
B urg elm a n , R, (1994), “ Fading M em ories: A Process T h eo ry o f Strategic Business Exit in
Dynam ic Environments,” A d m in is tra tiv e Scien ce Q u a rte rly 39(1): 24—56.
^Ca r pen ter , M ., Lazonick , W., and O 'S u lliv a n , M . (2003), “ T he Stock Market and
Innovative Capability in the New Econom y: The O ptical N etw orking Industry,” Industrial
a n d C orporate C h an ge 12(5): 963-1034.
C handler , A. (1962), Strategy a n d Stru ctu re: C h a p ters in the H is to ry o f the A m erica n Indus­
tria l E n terprise , Cambridge, Mass.: MIT Press,
* -(19 7 7 ), T he V isible H a n d : T he M a n a g e ria l R e v o lu tio n in A m e ric a n Business , Cam­
bridge, Mass,: H arvard University Press.
(1990b Scale a n d Scope: T he D y n a m ic s o f In d u s tria l C a p ita lis m , Cam bridge, Mass,:
Harvard University Press.
A m atori , E, and H ikino , T, (eds,) (1997), B ig B u sin ess a n d the W ealth o f Nations?
Cambridge: Cambridge University Press.
C lark , K., and F uji м о т о , T. ( i 9 9 i)> P ro d u ct D e v e lo p m e n t P e rfo rm a n c e : S trategy , Organiza­
tion a n d M a n a g em en t in the W orld A u to Industries? C am bridge, M ass.: H arvard Business
School Press.

Asterisked items are suggestions tor further reading,

 THE IN N O V A T IV E FIRM P

When a society's previously innovative firms no longer generate innovative

outcomes, there will be pressures on these firms to reallocate resources from invest­
m ents in existin g skill bases to investm ents in new types o f organizations. Such
o rgan ization al restru ctu rin g does not always o ccu r sm o o th ly o r successfully, as the
h istorical experien ces o f British indu strial districts, U S m anagerial co rp o ratio n s,
and Japanese enterprise grou p s have show n. Precisely because the in n o vative firm is
a social o rgan izatio n , the reallocation o f its resources is a social process in w hich
different gro u p s o f people can have very different interests. A n u n d erstan d in g o f the
ch an gin g organ izatio n o f the in n ovative firm is im p o rtan t for u n d erstan d in g not
o n ly h o w a so ciety innovates b u t also h o w a society copes w ith processes o f social
disruption in which the gains of some may be the losses of others.
In a th eo ry o f in n o vative enterprise, strategy, finance, and organ izatio n are
interlinked as a d yn am ic process w ith learn in g as an outcom e. To fully com preh end
the in n o vative firm, there is a need to understand the actual learn in g processes: the
relation betw een tacit know ledge and codified know ledge, betw een in d ivid u al
capabilities and collective capabilities, and betw een w hat is learned at a p o in t in
tim e and h o w that learn in g cum ulates o ver tim e. T h e prevailin g social co n d itio n s o f
innovative enterprise provide the context for those learning processes, sh apin g the
types o f learn in g that are attem pted, the extent to w h ich these processes are su s­
tained, an d the w ays in w h ich people interact both co gn itively and beh aviorally. T h e
influence o f the social context is m anifested b y the fu n ctio n al and hierarch ical
integration o f skill bases that, as this essay has illustrated, can v a ry d ram atically
across in du stries an d in stitu tion al en viron m en ts as w ell as o ver tim e.
A th e o ry o f in n o vative enterprise m u st be based on an u n d erstan d in g o f c o m p a ra ­
tiv e -h isto ric a l experien ce that is b road and deep en ou gh to evoke confidence that
the assu m p tio n s and relations fo rm in g the substance o f the th eo ry capture the
essence o f the reality to w h ich the th e o ry p u rp o rts to b e relevant. T h e d evelo p m en t
o f relevant th e o ry requ ires an iterative ap p ro ach in w h ich theoretical postulates are
d erived fro m the stu d y o f the h isto rical record and the resultant th eo ry is used to
analyze h isto ry as an o n g o in g an d u n fo ld in g process (L azon ick 20 0 2a). T h e in tel­
lectual challenge is to integrate th eo ry an d history.
A s Edith Penrose (19 8 9 :11) perceptively pu t it in an article w ritten late in her careen

“ Theory” is, by definition, a simplification o f“ reality” but simplification is necessary in order

to comprehend it at all, to make sense o f “ history” If each event, each institution, each feet,
were really unique in all aspects, how could we understand, or claim to understand, anything
at all about the past, or indeed the present for that matter? If, on the other hand, there are
common characteristics, and if such characteristics are significant in the determination ofthe
course o f events, then it is necessary to analyse both the characteristics and their significance
and “theoretically” to isolate them for that purpose.
I f w e need theory to make sense o f history, so we also need history to make sense o f
theory. A s Penrose concluded: "universal truths without reference to time and space
are unlikely to characterise economic affairs ”
54 WILLIAM L A Z O N I C K

Lazonick , W„ and O’Sullivan , M. (2000), “ Perspectives on Corporate Governance, Innov­

ation, and Economic Performance,” Report prepared for the project on Corporate Gov­
ernance, Innovation, and Economic Performance under the Targeted Socio-Economic
Research Programme of the European Commission, June (www.insead.edu/cgep).
Le slie , S., and Kargon , R. (1996), “ Selling Silicon Valley: Frederick Terman’s Model for
Regional Advantage,” Business History Review, 70(4). 435—7^-
M arshall , A. (1919), Industry and Trade, London: Macmillan.
____ (196!), principles of Economics, 9th (variorum) edn., London: Macmillan.
M orikawa , H. (1997), “ Japan: Increasing Organizational Capabilities o f Large Industrial
Enterprises, 1880S-1980S,” in A. Chandler, F. Amatori, and T. Hikino (eds.), B ig Business
and the Wealth of Nations, Cambridge: Cambridge University Press: 307-35-
------ and Kobayashi, K. (eds.) (1986), Development o f Managerial Enterprise, Tokyo: Uni-
versity o f Tokyo Press.
*M owery , D „ and R osenberg , N. (1989), T echnology a n d the P u r s u it o f E co n o m ic Growth ,
Cambridge: Cam bridge University Press,
N elson , R. (1991), “ W hy Do Firm s Differ, and H ow D oes It Matter?/" S trategic M anagem ent
Jo u rn a l и , Special Issue: 61-74.
------ and W inter , S. (1982), A n E v o lu tio n a ry T h e o ry o f E c o n o m ic C han ge. >Cam bridge, Mass.:
Harvard University Press.
N oble , D. (1977), A m erica by D esign : Science , Technology ; a n d th e R is e o f C o rp o ra te Capital­
ism, New York: Knopf.
O kimoto , D., and N ish i , Y. (1994), “ R & D O rganization in Japanese and Am erican Semi­
conductor Firms/" in M. Aoki and R. D ore (eds.), T h e Ja p a n e s e F ir m : T h e Sources of
C o m p etitive Strength , Oxford: O xford University Press: 178-208,
^O’S ullivan , M. (2000a), Contests fo r C o rp o ra te C o n tro l: C o rp o ra te G o v ern a n c e and Eco­
Oxford: Oxford University Press,
n om ic P erfo rm an ce in the U n ited States a n d G e r m a n y
------ (2000b), “The Innovative Enterprise and Corporate Governance/" C a m b rid g e Journal of
Econom ics 24(4): 393-416.
------ (2003), “ The Stock Market as a Source o f Cash in the U S Corporation/" IMSEAD
working paper.
*O w en , G. (2000), From E m p ire to E u ro p e: T h e D e c lin e a n d R e v iv a l o f B ritish In du stry since
the Second W orld W ar, New York: HarperColIins.
P enrose , E. (1989), History, the Social Sciences and Econ om ic 'T heory/ with Special
Reference to M ultinational Enterprise/" in A. Teichova, M . Levy-Leboyer, and H, Nuss-
baum (eds.), H istorical Studies in In te rn a tio n a l C o rp o ra te B u sin ess , Cam bridge: Cambridge
University Press, 7-13.
(*9 9 5 )> T he T h eory o f the G ro w th o f the F ir m , 3rd edn., O xford: O xford University Press
[first published 1959].
P iore , M „ and S abel , C. (1984b T h e S eco n d In d u s tr ia l D iv id e : P o ssib ilities fo r Prosperity ;
Basic Books.
S abel , C. G982), Work a n d Politics, Cam bridge: Cam bridge U niversity Press.
and Z eitlin , J. (1985), “ Historical Alternatives to M ass P roduction: Politics, Markets
^ and Technology in N ineteenth-Century Industrialization,” Past and Present 108:133-76-
Sako , M ., and Sato , H. (eds.) (1997), Japanese Labour and Management in Transition:
Diversity, Flexibility, and Participation, London: Routledge.
 THE I N N O V A T I V E FIRM 55

Sa x e n ia n , A. (1994), Regional Advantage: Culture and Competition in Silicon Valley and Route
128 . C am b rid ge, M ass.: H arvard U n iversity Press.
Sc h u m p e t e r , J. (1934), The Theory of Economic Developmenty C am bridge, M ass.: H arvard
U niversity Press.
------ (1950), Capitalismy Socialism, and Democracy; 3rd edn., N ew York: Harper.
T e e c e , D., P isa n o , G ., and S h u e n , A . (1997), “ D yn am ic Capabilities and Strategic M an age­
ment,” Strategic Management Journal 18(7): 509-33.
Y o n ekaw a , S. (1984), “ U n iversity G rad uates in Japanese Enterprises before the Second
W orld War,” Business History 26(3): 19 3-218 .
 CHAPTER 3

N E T W O R K S OF
INNOVATORS

W A L T E R W. P O W E L L
STINE GRODAL

3.1 In t r o d u c t io n

I n February o f 2001, two rival consortia published rou gh d raft (rou gh ly 90 per cent
complete) sequences o f the hum an genom e in Nature and Science. The puDiic
H um an Genom e Project consisted o f five key institutions and eleven collaborators,1
supported by the U S N ational Institutes o f H ealth, D epartm en t o f Energy, and the
W ellcome Trust in the United Kingdom . The rival "p riv a te ” consortia, led by the
biotech firm Celera, included both com m ercial firm s and academ ic researchers from
the U niversity o f C alifornia, Penn State, Case W estern, J ohns H op kin s, Cal Tech, Yale,
Rockefeller, as well as scientists in Spain, Israel, and A ustralia. These projects have
been acclaimed for their rem arkable scientific achievem ent; they were also the
product o f considerable organizational innovation. In contrast to the Manhattan
Project or Project Apollo, both o f w hich were h ierarch ically organized, national
projects, the H um an Genom e Project (H G P ) and the C elera team were pluralist,
m ultiorganizational, m ultinational confederations. These tw o groups were intensely
rivalrous, but collaborated intensively w ithin their ow n grou p s (Lam bright 2002).
H G P involved m anagem ent by two governm ent agencies and a private British

We are grateful to David Mowery and Jan Fagerberg for their careful readings o f earlier drafts.
 N E T W O R K S OF I N N O V A T O R S 57

foundation that coordinated activities in government labs, universities, and

nonprofit institutes in the US and England, As the lead firm, Celera’s organization
was more focused, but its research team included scientists and state-of-the-art
equipment at private firms, public and private universities, and nonprofit institutes
in four countries.
Both projects were organized as large-scale networks, and their rivalry spurred
each side to engage in a high-stakes learning race. While the cost, scale, and
distributed nature of these projects may have been unusual, the form of organiza­
tion—collaboration across multiple organizational boundaries and institutional
forms—is no longer rare. Indeed, many analysts have noted that the model of
networks of innovators has become commonplace over the past two decades (Powell
1990; Rosenbloom and Spencer 1996; Roberts and Liu 2001; Chesbrough 2003).
C o llab o ratio n a m o n g ostensible rivals was once regarded as a p ro visio n al o r
transitional step taken to enter n ew m arkets, spread risks, or to share early stage
R & D costs (M o w e ry 1988). Such forays w ere often follow ed b y m ergers as the
tran sitory activities b ecam e in co rp o rated inside the b ou n d aries o f the firm , Recent
studies suggest, h ow ever, that v ario u s fo rm s o f interorganization al partnerships are
now core co m p o n en ts o f co rp o rate strategy. Even w here these linkages endure fo r
relatively len gthy p erio d s o f tim e, they do not entail vertical integration (G om es-
Casseres 1996; H ag e d o o rn 1996; N o te b o o m 1999; A h u ja 20 0 0 a). C o n te m p o ra ry
studies o f in d u strial p e rfo rm an ce are replete w ith reports o f a significant upsurge
in variou s types o f in tero rg an izatio n al collab oration . W hile these collaborations can
take a n u m b er o f fo rm s (in clu d in g research consortia, jo in t ventures, strategic
alliances, and su b co n tractin g) and span a w ide range o f key fu nctions, a N ation al
Research C o u n c il an alysis o f trends in in du strial research and d evelopm ent (R&D)
suggests that the in n o va tio n process has u n d ergon e the m ost significant tra n sfo rm ­
ation over the past decade (M errill an d C o o p er 1999). In a su rvey o f the perio d
1960-98, H ag e d o o rn (2002) fin d s a sharp grow th in R & D collab oration s, b egin n in g
in the late 1970s an d co n tin u in g th ro u g h the m id-1990s.
A N atio n al R esearch C o u n c il assessm ent o f eleven U S-b ased industries, p u r­
posefully d iverse in ch aracter and tech n o lo gy b u t all resurgent in the 1990s, observes
in every sector an in creased reliance o n external sources o f R & D , n o tab ly u n iver­
sities, co n so rtia an d g o vern m en t labs, and greater co llab o ration w ith d om estic and
foreign co m p etito rs, as w ell as cu stom ers in the d evelopm ent o f n ew p ro d u cts
and processes (M o w e ry 19 9 9 :7 ). O th er surveys also p o in t to the enhanced centrality
o f in tero rgan izatio n al co llab o ratio n , especially in R & D . F o r exam ple, N ation al
Science F o u n d a tio n d ata sh o w a m arked increase in the n u m b er o f in tern atio n al
alliances betw een U S an d W estern E u ro p ean countries betw een 1980 an d щ §41
b u t b y the m id -19 9 0 s, the fo rm a tio n rates fo r in tran ation al alliances lin k in g U S
firm s w ith th eir d o m estic co m p etito rs o u tp ace in tern atio n al linkages (National
Science B o a rd 1998). T h e fo rm e r co llab o ration s w ere m o tivated largely b y concerns
w ith m arket access, w h ile the latter focu s m o re o n the d evelo p m en t o f new
technologies.
58 WALTER W, POWELL AND STINE GRODAL

Similarly, there is now ample research illustrating the growing links between US
firms and universities (Powell and Owen-Smith 1998), and greater involvement by
firms and government labs in research joint ventures (Link 1 9 9 6 , 1999). In the realm
of science, Hicks and Katz (1996) find that research papers are much more likely to
be co-authored and involve authors with multiple institutional affiliations that
span universities, government, and industry. Distributed networks of practice
are the organizing basis for many technical communities, suggesting both that
sources of knowledge are now more widely dispersed and that governance mechan­
isms are emerging to orchestrate distributed knowledge. The open source software
movement is but one highly visible example of this trend (O’Mahony 20 02; Weber
20 0 3), which illustrates how advances in information technology have greatly
facilitated virtual networks, In short, as Mowery (19 9 9 9 ) observes, “the diversity
-

of institutional actors and relationships in the industrial innovation process has

increased considerably.” Complex networks of firms, universities, and government
labs are critical features of many industries, especially so in fields with rapid
technological progress, such as computers, semiconductors, pharmaceuticals, and
biotechnology.
Our goal in this chapter is to assess the state of scholarly research on the role of
networks in the innovation process. We begin with a review of the factors that have
triggered the increased salience of networks. We discuss different types of networks,
distinguishing between networks that are based more on contractual or market
considerations, and those that are based on less formal, and more primordial
relationships, such as common membership in a technological community or a
regional economy. We then turn to a discussion of the analytical leverage provided
by the tools of network analysis. This stream of research, which spans sociology,
social psychology, organizational behavior, and business strategy, highlights key
distinctions between highly clustered, dense networks, steeped in overlapping ties
and high in trust, and weak-tie networks, that provide access to novel, non-redun­
dant information. We next review a number of empirical studies of the contribution
of networks to the innovative output of firms. We take up the issue of knowledge
transfer, examining how the codification of knowledge can shape what is transmit­
ted through networks. We briefly discuss the governance of networks, and then
conclude with an assessment of what types of organizations and settings derive the
greatest impact on innovation from participation in networks.
Research on the relationship between networks and innovation is a relatively
recent area of inquiry. While there is a good deal of work underway, direct analyses
measuring the impact of interfirm networks on performance are limited. Much of
the extant research focuses on the effects of networks on patenting, access to infor­
mation, and the generation of novel ideas. Moreover, the studies often examine
high-tech industries, where investment in R&D is pronounced. Attention to
the consequences of network ties for the financial performance of firms is relatively
rare.
 N E T W O R K S OF I N N O V A T O R S 59

3.2 W h y h a v e N e t w o r k s G r o w n
in Im p o r t a n c e ?

The advantages o f a h eterogen eou s grou p o f contacts are well established in both
social th eo ry an d n etw ork analysis. A stron g trad ition o f th eo ry and research,
running fro m Sim m el (1954) to M erto n (1957) to G ran ovetter (1973) to B u rt
(1992), m akes ab u n d an tly clear that there are in fo rm atio n al, status, and resource
advantages to h avin g b ro ad and d iverse social circles. B elow w e review an array o f
recent em p irical studies that d em on strate h o w interorganization al relationships
lead to v ario u s benefits w ith respect to in fo rm atio n d iffu sio n , resource sharing,
access to specialized assets, and in terorgan ization al learning. In science and tech ­
n ology-based fields, the advantages that accrue fro m diverse sources o f in fo rm atio n
and resources are considerable. N o t surprisingly, then, as the co m m ercialization o f
know ledge has assu m ed greater im p o rtan ce in econ o m ic grow th, co llab o ration
across o rgan izatio n al b o u n d aries has b ecom e m o re com m on place. In terorgan iza­
tional netw orks are a m eans b y w h ich organ ization s can p o o l or exchange resources,
and jo in tly d evelop n ew ideas and skills. In fields w here scientific o r tech nological
progress is d evelo p in g rapidly, and the sources o f know ledge are w id ely distributed,
no single firm has all the necessary skills to stay on top o f all areas o f progress and
bring sign ifican t in n o va tio n s to m arket (Pow ell and B ran tley 1992; Powell, K oput,
and S m ith -D o e rr 1996; H ag e d o o rn and D uysters 2002). In such settings, netw orks
can becom e the lo cu s o f in n o vatio n , as the creation o f know ledge is crucial to
im p ro vin g co m p etitive p o sitio n .
C o llab o rative n etw ork s have lo n g been central to the p ro d u ctio n process in craft-
based indu stries (Eccles 19 8 1), in in d u strial districts (B ru sco 1982; P iore and Sabel
1984), and in fields such as aerospace w h ere assem bly depended u p o n key inputs
from d iverse p articip an ts. T h e grow th o f kn o w led ge-in ten sive indu stries has
heightened the im p o rta n ce o f netw orks in R & D as w ell as p ro d u ct d evelopm ent
and d istrib u tion . A p ersisten t fin d in g fro m a diverse set o f em p irical studies is that
internal R & D in ten sity an d tech n o logical so p h isticatio n are p o sitively correlated
w ith both the n u m b e r an d in ten sity o f strategic alliances (Freem an 19 9 1; H aged o o rn
1995).
Fo r o rgan izatio n s in ra p id ly d evelo p in g fields, h eterogen eity in the p o rtfo lio o f
collaborators allow s firm s to learn fro m a w id e stock o f know ledge. O rgan ization s
w ith b ro ad er n etw o rk s are exp osed to m o re experiences, d ifferent com petencies,
and add ed o p p o rtu n itie s (B eck m an and H au n sch ild 2002). Su ch access creates an
en viron m en t in w h ic h “ creative abrasion,” the synthesis that is developed fro m
multiple p o in ts o f view , is m o re lik ely to occur. In this view , in n o vatio n occu rs at
the b o u n d a rie s betw een m in d sets, n o t w ith in the p ro vin cial te rrito ry o f one
fe o w le d g e a n d sk ill b ase” (L e o n a rd -B a rto n 1995: 62). B y h avin g access to a m o re
60 WALTER W. POWELL AND STINE GROP AL

varied set o f activities, experiences, and collaborators, com panies broaden the
resource and knowledge base that they can draw on. B y d evelopin g m ore multiplex
ties with individual partners, either through p u rsu in g m u ltiple collaborations or
expanding an existing R & D partnership into dow n stream developm ent, companies
increase the points o f contact between them . W hen relationships are deepened,
greater com m itm ent and m ore thorough know ledge sh arin g ensue. Organizations
with m ultiple and/or m ultifaceted ties to others are likely to have developed better
protocols for the exchange o f inform ation and the resolu tion o f disputes (Powell
1998). Parties that develop a broader bandw id th fo r co m m u n icatio n are, in turn,
m ore capable o f transferring com plex know ledge. In scien ce-driven fields such as
biotechnology, organizations that develop ties to different kinds o f organizations
and carry out m ultiple types o f activities w ith these o rgan izatio n s are central players
in industry networks (Powell et al. 2004). These centrally po sitio n ed organizations
are both capable o f pulling prom ising new entrants into the n etw ork and collabor­
ating with a wide assortm ent o f incum bents. M oreover, research shows that in
biotechnology, organizations lacking such connections fail to keep pace and fall by
the wayside (Powell et al. 2005).

3.3 V a r i e t i e s o f N e t w o r k s

The literature on networks em phasizes that they are m ost p ro n o u n ced in the domain
between the flexibility and autonom y o f m arkets and the force and control of
organizational authority (Powell 1990). N etw orks thus co m b in e som e o f the incen­
tive structures o f markets with the m on itorin g capabilities and administrative
oversight associated with hierarchies (M ow ery, O xley, and Silverm an 1996). For
our purposes, we include networks based on fo rm al contractual relations, such as
subcontracting relationships, strategic alliances or p articip ation in an industry-wide
research consortium , and inform al ties, based on co m m o n m em bership in a pro­
fessional or trade association, or even a looser affiliation w ith a technological
com m unity.
One can differentiate networks with respect to their duration and stability, as well
as whether they are forged to accomplish a specific task or evolve out o f pre-existing
bonds of association. Networks vary from short-term projects to long-term rela­
tionships, and the different temporal dimensions have important implications for
governance. Some networks are hierarchical, monitored by a central authority; while
others are more heterarchical, with distributed authority and strong self-organizing
features. Grabber and Powell (2004) focus on temporal stability and forms of
 N E T W O R K S OF I N N O V A T O R S 6l

governance to differentiate fo u r key types: in form al netw orks (based on shared

experience); project n etw ork s (sh o rt-term co m b in ation s to accom p lish specific
tasks); region al netw orks (w here spatial p ro p in q u ity helps sustain a co m m o n
co m m u n ity); and business netw orks (p u rp o sive, strategic alliances between two
parties). These types d o not represent essentialist categories; rather they m ay overlap
and interw eave w ith on e another. C o n sid er these fo rm s as useful coordinates to
locate netw orks w ith respect to d ifferent co m b in ato ry elem ents.
Several key concepts p ro vid e po ten t analytical tools that ap p ly across different
types o f netw orks and p erm it assessm ent o f their effects. First, consider the d iffer­
ences betw een stron g an d w eak ties (G ran ovetter 1973). In interperson al term s, a
strong tie is a perso n w ith w h o m y o u interact on a regular basis, w hile a w eak tie is an
acquaintance, o r a frien d o f a friend. Stro n g ties are im p o rtan t for social su p p o rt, but
m uch o f the novel in fo rm a tio n that a person receives com es fro m w eak ties. Stro n g
ties are based on co m m o n interests, consequ ently m ost in fo rm atio n that is passed
reinforces existin g view s. W eak ties in tro d u ce n ovelty in the fo rm o f d ifferent ideas
or tastes, and b y in tro d u cin g n ew in fo rm atio n they are, fo r exam ple, invaluable in
job searches an d o th er circu m stan ces w here a sm all am o u n t o f new in fo rm atio n is
highly useful. W eak ties have a lo n ger reach, but a m u ch n arro w er b an d w id th than
strong ties. T h e latter are m o re cohesive, an d o ften prove to be m o re effective at the
exchange o f co m p le x in fo rm a tio n . Figu re 3.1 illustrates the difference betw een
strong an d w eak ties.
M uch o f the research o n in terfirm netw orks extrapolates fro m interperson al
relations.2 In general, this is a plau sib le analytical m ove; how ever, it elides the
question o f w h eth er relatio n sh ip s at the firm level are dependent o n o n goin g
interperson al ties, an d w h eth er the business relationship w o u ld be h arm ed o r
severed i f the k ey p articip an ts w ere to depart. T h e extent to w h ich interorganiza-
tional ties are co n tin g en t u p o n relation s am o n g in d ivid u als is a key question for
scholarly research, as w ell as a critical challenge for business strategy (G u lati 1995?
Pow ell 1998).
A second n o tab le co n trast is the d istin ctio n betw een netw orks as brid ges and as
structural h oles (B u rt 1992). B rid ges are p o in ts o f con n ectio n betw een parties that

Fif«3,1 Strong and weak ties

62 W A L T E R W. P O W E L L A N D S T I N E G R O D A L

lack ties, such as when A knows B, and C know s B b u t n o t A . B is the bridge between
A and C, thus the gateway to a linkage between A and C . G ran ovetter (1973) argued
that bridges are the links that m ake w eak ties possible. B u rt (1992) deepened the
argum ent by m oving from the w ho question (i.e., w h ich p o sitio n in a network is best
situated) to the question o f how certain structural arrangem ents generate benefits
and opportunities. He coined the term structural holes as a potential connection
between clusters o f units that are not connected. The p o ssib ility o f m aking such a
connection provides leverage, or opportu nities fo r arbitrage. Those positioned to
take advantage o f structural holes can broker gaps in the social structure. See Figure
3.2 for illustration.
There is debate as to whether strong or w eak ties, o r brid ges o r structural holes,
offer greater opportunities for innovation (A h u ja 2000a; R u e f 20 0 2). Clearly, vari­
ation in network structures is associated w ith d ifferent content in relationships.
Strong ties between two parties m ay restrict in fo rm atio n gathering in terms of the
breadth o f search, but the inform ation that is exchanged is “ thick,” or detailed and
rich. Weak ties are thinner and less durable, but p ro vid e better access to non-
redundant inform ation. There is also disagreem ent as to w hether networks can be
designed or “ pruned” to produce “ o p tim al” shapes, w ith o u t triggering repercus­
sions. W hether location in a netw ork is h igh ly m alleable o r not, position in a
network both em powers and constrains opportu nities.
A third point o f contrast is between netw orks fo rm ed intentionally across a
market interface to accom plish a task and em ergent netw orks that grow out of
ongoing relationships. The form er m ay be considered an instrum en tal or strategic
relation, while the latter stems from m ore p rim o rd ial relations, such as common
ethnicity, friendship, or location. These different starting points m atter, but in the
fluid w orld o f networks, the point o f origin does not fix the evolu tion o f a relation­
ship. Consider two cases. There is a global trend tow ard vertical disaggregation in
m anufacturing, as firm s are relying on suppliers for design and com ponent inputs

Structural hole

Fig. 3 .2 Structural holes and redundant ties

 NETWORKS OP I N N O V A T O R S 63

in a variety o f indu stries (see W om ack, Jones, and R oos 1990, for autos; M cK en drick,
D oner, and H aggard 2000, fo r disk drives). O ften these o u tso u rcin g decisions are
driven by the need to reduce costs, save tim e, and enhance flexibility, w hile the large
firm concentrates 011 those activities in w hich it has som e form o f com petitive
advantage. But as m an y analysts have noted, there is no natural sto pp in g p o in t in
this relation (Sabel 1994; H elper, M cD u ffie, and Sabel 2000), T h e su b co n tractor can
end up in volved in design issues, d o in g critical R & D , o r becom e central to efforts to
im prove quality. W hat began as a choice to ou tsou rce can, in som e circum stances,
becom e either a deep, m u tu ally dependent co llab o ration o r a h ig h ly am b ig u o u s and
o p p o rtu n istic p artn ersh ip. H elper et al. (2000) and D yer and N ob eoka (2000)
illustrate the m arked trend fo r au to m o b ile su b con tractin g to evolve into in terd e­
pendent, bilateral relationships.
O r con sider the c o n tem p o ra ry life sciences, w here m an y R8cD partnerships
em erge out o f o n g o in g intellectual relation sh ips— co -au th o rsh ip s, m en to r-m e n tee
relationships, and co m m o n train in g (M u rray 2002). These in fo rm al person al rela­
tionships m ay, how ever, co m e to involve significant intellectual p ro p e rty in the fo rm
o f patents, an d thu s b ecom e h igh ly form alized contractual agreem ents between
organizations. We o ffer these exam ples as illu strations that netw orks forged out o f
strategic p u rp oses can take on stron g relational elem ents, w hile m o re person al ties
can becom e co n tractu al and h ig h ly specified. W hile it is possible to assign netw orks
to either a tran scatio n al (i.e., based o n a con sideration o f business o p p o rtu n ities
w ith ou t regard to p rio r social relations) o r relational (i.e., em bedded in o n goin g
social relation sh ips) category, it is in ap p ro p riate to assum e that relations rem ain
fixed. A s netw orks evolve, there is considerable give and take. Intense co m p etitio n
can render calcu lative strategic alliances m ore em bedded, w hile the prospect o f great
financial rew ard can tu rn a “ h an d sh ak e55 relationship betw een in d ivid u als into a
form al legal lin kage betw een firm s.
Figure 3.3 p ro vid es a ty p o lo g y o f different fo rm s o f netw orks, w ith the h orizo n tal
axis rep resen tin g degree o f p u rp osiven ess, ran gin g fro m in fo rm al to contractual.
The vertical axis represents the extent o f em beddedness, v ary in g fro m open, e p i­
sodic, o r flu id to recu rren t, dense con n ection s am o n g a fairly closed g ro u p (G ran -
ovetter, 1985). We illustrate each o f the fo u r cells w ith exam ples o f types o f
in n o vatio n n etw orks. In the lo w er left cell w e place in fo rm al netw orks, such as a
scientific in visib le college, that em erge o u t o f shared exp erien ce o r c o m m o n interest.
A lth o u gh these relation s ten d to be te m p o ra ry and sh ort-lived , the g ray arrow s are
intended to sh o w that these in fo rm al linkages can evolve into fo rm a l business
alliances o r m o re e n d u rin g p rim o rd ia l relations, w here p articip atio n is m o re c o n ­
stant a n d less flu id . T h e p rim o rd ia l n etw ork in the u p p er left cell is characterized by a
c o m m o n so cial iden tity, co n tin u o u s p articip atio n , an d close ties. A ll these features
are o ften fo u n d in p ro fe ssio n al netw orks, craft-b ased o ccu p atio n s, ethnic c o m m u ­
nities, an d in d u stria l districts. T h e u p p er rig h t cell is typ ified b y in volvem en t in a
common project. M e m b e rsh ip in such a network is typically restricted and o ften
64 W A L T E R W. P O W E L L A N D S T I N E G R O D A L

Embeddedness
Supply Chain
Primordial

Type:
Type:
Nodes dissimilar
Nodes similar
Common social identity Common work identity
Function:
Function:
*k0-% Multifunctional, craft based, Horizontal or vertical specialization
& Division of labor
J3 project-based
£ Structure:
Incremental innovation
<u
£ Dense, since the network exists Structure:
“O
<u prior to activity, work relationships Spider web
ІЛ are nested in ongoing personal The establishment of ties creates the
О
relations network The network can become a
Examples: form of social identification over
Film, construction, ethnic time, as for example in the Toyota
community, diamond trade supply chain network.
Examples:
Supply chain
Focal
firm

Invisible College Strategic

Type: Type:
Nodes both similar and dissimilar Nodes both similar and dissimilar
Common interest Identity different
Function:
Function:
Fast access to news and novel ideas
Division of labor
Discovery Recombination
Structure:
>* Structure:
Ties reinforce the structure of the
Purposive network, emerges through
as college
active tie creation
Examples:
Examples:
research collaboration, information
Common in high technology
sharing
industries like semiconductors,
Firm
biotechnology, and chemicals
University1
О ---------- О --------- О
Biotech Venture Pharma
Research Government
capital
institution

Caleulativeness

lnf0rmal Contractual

Fig. 3.3 Network typology

are ant р ѵ я і I ^ Г т ^ 8иР Р ^ ~ с^ аіп netw ork o r a large construction project

can evolve • ^ m eant t0 illustrate that su pply-ch ain relations
form al h u s i ° ^ ** °^ cupatl0nal com m unities and in d u strial districts or into
is the Strate^ r Pf r n m ips' T ^e m ost pu rp osive and instrum ental type o f network

discussion o f the relatiom hfn her l th* low er ri§ht celL We turn now t0 a
Р ween these fo rm s and the in n o vatio n process.
 3*4 E m p i r i c a l S t u d i e s o f t h e R o l e
of N e t w o r k s in I n n o v a t i o n

3 4 ,1 Formal Ties
M ost em p irical studies o f the relationship betw een netw orks and in n o vatio n focus
on fo rm al ties established am o n g organizations. T h is stream o f research d ocum en ts
a strong p o sitive relation sh ip betw een alliance fo rm atio n and in n o vation , across
such diverse indu stries as ch em icals (A h u ja 2 0 0 0 a), b io tech n o lo gy (Pow ell et aL
19 9 6 ,19 9 9 ; W alker, K o gu t, and Shan 1997; B au m , Calabrese, and Silverm an 2000),
telecom m u n icatio n s (G o d o e 2000), and sem icon d u ctors (Stuart 1998, 2000). The
d iversity o f the research contexts suggests the effects o f n etw ork structure m ay be
generalizable. N evertheless, m ost research has focused on high tech n o logy in d u s­
tries, and uses patents as a p ro x y fo r in n o vation . M o re direct m easures o f in n o vative
outputs are needed. Som e o f the im p o rtan t them es that em erge fro m this research
h ighlight specific tie characteristics, tech nological uncertainty, and netw ork evo lu ­
tion. In ad d itio n , researchers have em phasized the increased benefits in the fo rm o f
resources and k n o w led ge that alliances p ro vid e to entrepreneurial firm s.

Tie characteristics. O ne line o f research focuses on h o w different types o f ties

influence the benefits d erived fro m alliances. V in d in g (2002) identified 548 D an ish
m an u factu rin g firm s that d eveloped on e o r m o re new p ro d u cts over a tw o -year
period. In in terview s w ith a subset o f the com pan ies, he finds that the im p act o f a
co llab o ration o n in n o va tio n is sign ifican tly related to both the type o f partn er and
the pattern o f p re vio u s co llab o rative relationships. T h e im p o rtan ce o f p rio r in ter­
action w ith p artn ers p o in ts to the significance o f relation sh ip b u ild in g, and h o w
elem ents such as tru st an d cogn itive u n d erstan d in gs requ ire tim e to develop.
D o m estic p artn ers w ere fo u n d to have a greater p o sitive im p act on in n o vative
p erfo rm an ce than fo reign partn ers, p o ssib ly d u e to the h igh er costs, b oth p sych o ­
logical an d fin an cial, associated w ith m o re distant collaborations. V in d in g s research
em phasizes ben efits d erived fro m stron g local ties. Sim ilarly, in a ten -year (19 8 0 -9 0 )
case stu d y o f the R & D p o rtfo lio o f a N o rw eg ian telecom m u n icatio n s organ izatio n ,
G o d o e (2000) rep o rts co m p arab le results w ith respect to stron g ties. H is analysis
suggests that rad ical in n o vatio n s w ere m o re likely to em erge fro m in tim ate and
p ro lo n ged in teractio n . B u t in the N o rw eg ian case, the affiliation s w ere n o t local, b u t
instead based o n m em b ersh ip in in tern atio n al telecom m u n icatio n s association s.
Pow ell et aL (1999) em ph asize that experien ce w ith co llab o ration an d cen trality in
the n etw o rk d erived fro m a d iverse set o f ties are im p o rtan t determinants o f
innovation a m o n g b io te c h n o lo g y firm s over the period 1988-99. Their analyses
suggest th a t c e n t r i i t y an d exp erien ce resulted in m o re patenting. The most conse*
in tern s o f patenting were R&D partnerships. The diversity o f
66 W A L T E R W. P O W E L L A N D S T I N E G R O D A L

network ties also had a positive influence on rates o f patenting. Powell et al. (1999)
found that while network experience had a positive in flu en ce on patenting, the rate
o f increase dim inished with additional experience, suggesting possible declining
returns to network connectivity. The question o f w hether there are limits to con­
nectivity needs to be investigated further.3 These results suggested a “ cycles of
learning” process in which R & D collaborations generate attention that attracts
other partners, who collaborate in developing n ovel ideas. T h is enhanced diversity
o f affiliations increases a firm's experience at m an agin g co llab o ration s and transfer­
ring knowledge, and increases their centrality in the in d u stry network. Greater
centrality is associated with a higher rate o f patenting, and both centrality and
higher volum es o f patenting trigger subsequent R & D partnerships, restarting the
cycle for centrally placed firms.
M ost research has looked at the presence or absence o f a fo rm al collaboration,
Ahuja (2000a), however, developed а m ore nuanced analysis inclu ding both direct
and indirect ties, and the level o f indirectness. D raw in g data fro m a sample of 97
firm s in the global chem icals industry, he used the n u m b er o f patents as a measure of
innovative output, while collaboration was m easured th ro u gh form al ties. More
distant connections through affiliates o f partners w ere coded as w eak or indirect ties.
The results show that both direct and indirect ties have a positive influence on
innovation, though the im pact o f indirect ties is sm aller than the im pact o f direct
ties. The num ber o f direct ties also negatively m oderates the im pact o f indirect ties.
In contrast to B u rt’s (1992) argum ents about the arbitrage opportu nities available
through non-redundant contacts, A huja shows that a n etw ork w ith m any structural
holes can reduce innovative output, as m easured b y rates o f patenting. A key
advantage o f close-knit networks m ay be due to their su p erio r ability to transfer
tacit knowledge (Van W ijk, Van den Bosch, and V olberda 2003). In an analysis o f the
exchange o f inform ation across project team s in a large m u ltinational computer
company, Hansen (1999) also illustrates that com plex know ledge is transferred most
easily through tightly knit networks.

Entrepreneurial firms. One active area o f research concerns the effects o f networks
on survival chances o f newly founded firm s. Larson s (1992) ethnographic study of
how a startup firm grew and prospered by draw ing on external resources and
support for key business functions illustrates h o w relationships are forged and
sustained as startup firm s grow. W hile not explicitly lo o k in g at innovative output,
Larson added insight into the signal im portance o f netw orks in obtaining resources
necessary to fuel a startup firm s success. Shan, Walker, and K ogu t (1994) examined
whether biotechnology startup firm s' cooperative relationships w ith other firms had
a positive effect on patenting. Their results offer su p p o rt for the argum ent that
collaborative relationships increased innovation, because form al cooperative rela­
tionships explained innovative output, while innovative ou tp u t did not account for
the pattern of alliances. The salience o f alliances for yo u n g and sm all firm s is further
 N E T W O R K S OF I N N O V A T O R S 6/

em phasized in S tu a rts (2000) stu d y o f in n o vatio n in the sem icon d u ctor industry.
H is dataset includes 150 firm s, follow ed by the co n su ltan cy firm D ataquest over the
period i985“ 9i. D raw in g on sales figures, patterns o f strategic alliances, and
patenting activity, Stu art show s that firm s possessing tech n ologically sophisticated
alliance partn ers patented at a su bstantially greater rate than those that lacked such
ties. Firm s establishin g strategic alliances w ith large partners also grew at a h igher
rate than firm s w ith o u t access to such partners. T h e returns to netw orks w ith regard
to patentin g w ere greatest fo r both you n g and sm all firm s.
B au m , C alabrese, and Silverm an (2000) pu rsue a sim ilar question, asking h o w the
co m p o sitio n o f a startu p firm 's alliance p o rtfo lio affects its p erform an ce. U sing data
on 142 b io tech n o lo g y firm s fo u n d ed in C an ad a betw een 1991 and 1996, they find a
positive effect o f alliance fo rm atio n on startup in n o vation . N etw o rk efficiency,
defined as the d iversity o f in fo rm atio n and capabilities p er alliance, show ed a large
positive effect on the n u m b er o f biotech patents. A lliances w ith direct co m petitors
had a negative effect on in n o vation , how ever. These results w ere m o d erated w hen
the rival b io tech n o lo g y firm had a larger share o f the relevant m arket o r i f the rival
b io tech n o lo gy firm w as h ig h ly innovative. O f the vario u s p erfo rm an ce m easures
used, the n u m b er o f patents and the volu m e o f R & D expen ditu res w ere m ost
significantly in flu en ced b y rates o f alliance fo rm atio n .

Network dynamics . D raw in g on C o h en and LevinthaFs (1990) ideas about ab so rp ­

tive capacity, Pow ell et al. (1996) argu ed that firm s utilize external collab oration s to
stay abreast in rap id ly d evelo p in g tech nological fields. B u t o rgan ization s cannot be
passive recipients o f new know ledge. “ W hat can be learned is cru cially affected by
w hat is already k n o w n ” (Pow ell et al. 19 9 6 :12 0 ). To un d erstan d the “ n ew s” generated
externally, o rgan izatio n s have to m ake “ n ew s” internally. In this fashion , the rate o f
acquisition o f skills and resources fro m the ou tsid e is closely linked to the generation
o f expertise internally. In their w o rk on the glo bal b io tech n o lo g y indu stry, they find
that firm s that d evelop experien ce at m an agin g co llab o rative R & D relation sh ips
garner faster access to cen trally p o sitio n ed organizations. A s experience at co llab o r­
ating grow s, firm s w id e n the n etw ork o f o rgan izatio n s w ith w h o m th ey partner. A s a
firm 's exp erien ce w ith co llab o ratio n an d its d iversity o f partn ers increase, the m o re
central an d visib le the firm becom es in the indu stry. T h is cen trality leads, in tu rn , to
grow th in the size o f the firm , an d to the ab ility to co ord in ate m o re alliances, creatin g
a feedback cycle. T h is cycle o f learn in g has been sh ow n to be associated w ith po sitive
fin an cial p e rfo rm an ce (P ow ell et a l 1999), and a greater ab ility to collab orate w ith
diverse k in d s o f o rgan izatio n s, w h ich perm its firm s to retain a leadersh ip p o sitio n in
the industry (Powell et al. 2005).
The general picture that emerges from research in organizational sociology and
business strategy is one in which networks and innovation constitute a virtuous
cycle* External linkages facilitate innovation, and at the same time innovative
Dutputs attmct further collaborative ties. Both factors stimulate organizational
68 W A L T E R W. P O W E L L A N D S T I N E G R O D A L

growth, and appear to enhance fu rther in n ovation . A h u ja (2000b) and Stuart

(2000), for exam ple, dem onstrate that firm s w ith m an y p rio r patents are more likely
to form alliances than firm s lacking patents, suggesting a recursive process of
innovation and growth in which collaborative ties p la y a central role. Further
attention needs to be given, however, to such issues as the effects o f the duration
o f linkages, experience w ith collaboration, and the consequences o f broken ties on
rates o f innovation.

Technological uncertainty. Two additional aspects o f the in n o vatio n process involve

the relationship between strategy and alliance fo rm atio n and the level o f techno»
logical uncertainty in the field. Eisenhardt and Sch o on h o ven (1996) studied the
population o f sem iconductor firm s launched in the U S betw een 1978 and 1985, and
found that the m ore risk-taking a com pany's strategy, the m o re alliances a company
form ed. One explanation is that as firm s gain cred ib ility fo r developing pioneering
technologies, access to financial and other resources fo r developing innovative
technology is secured through alliances. A n alternative v ie w is that alliances are
necessary to share the attendant risks in h igh -velocity environ m en ts. Sarkar, Echam-
badi, and H arrison s (2001) analysis o f m anagers in a range o f high-tech industries
revealed that an active strategy o f alliance fo rm atio n enhanced performance, as
m easured by market share, sales grow th, m arket d evelopm ent, and product innov­
ation. They also report that m anagers w ho perceived the environm ent as more
uncertain were m ore likely to pursue alliances. In ad d ition , sm aller firm s derived
m ore value from network linkages than larger firm s, p resu m ab ly because smaller
com panies viewed the technological landscape as m ore uncertain.
R osen kopf and Tushm an (1998) exam ined the role o f technical communities in
the flight sim ulation industry, where cooperative technical organizations play a
critical role in developing standards and advancing the state o f the art. In a study
covering the years 1958-92, they found long periods m arked b y increm ental change
punctuated by shorter eras o f ferm ent. T h ey sh ow that the rate o f founding of
technical networks increases during periods o f discontinuity, and stabilizes into
core cliques when ferment declines and a d om in an t design em erges. Subsequent
technological change disrupts dom inant cliques, and triggers the form ation o f new
networks, restarting the cycle. Thus, both tech nological strategy and industry
evolution are linked to patterns o f netw ork fo rm atio n , w ith external networks
assum ing greater im portance during periods o f technical discontinu ity and for
firm s with m ore risk-taking strategies. The im portan ce o f in d u stry technical com­
mittees in standards setting has also been em phasized in the com puter industry
(Farrell and Salom er 1988) and videocassette recorders (C u su m an o et al. 1992).
The overall conclusion o f this group o f studies is that netw orks provide access to
m ore diverse sources of inform ation and capabilities than are available to firms
lacking such ties, and, in turn, these linkages increase the level o f innovation inside
firms. Younger and smaller firm s m ay benefit m ore fro m collaborative relationships
 N E T W O R K S OF I N N O V A T O R S 69

than do larger firms* M o st notably, firm s w ith a central location w ith in netw orks
generate m o re in n o vative o u tp u t. B oth direct and indirect ties p ro vid e a positive
co n trib u tio n to in n o vation , but the effect o f indirect ties is m oderated by the
prevalence o f direct ties. T h e evidence for the benefits o f stru ctu ral holes is not
u n ifo rm ; w here stru ctu ral holes m igh t be beneficial is in the search fo r new in fo r­
m ation, b u t the know ledge transfer process appears to be facilitated b y closer-kn it
netw orks. From the vie w o f the d yn am ics o f co llab o ration , successful external
relations app ear to beget m o re ties, w hich fuel firm grow th and in n o vation . Clearly,
there are lim its to this cycle, but research has not addressed this qu estion in depth
thus far.
The m a jo rity o f the studies review ed in this section have been carried out using
patents as the dep end en t variable and fo rm al relationships as the independent
variable. Patents p ro vid e a m easure o f n o velty that is extern ally validated through
the patent exam in atio n process, hence th ey are a useful in d icato r o f know ledge
creation (G rilich es 1990). B u t patents have som e lim itation s. Som e kinds o f in n o v ­
ations are not patented, and there is variatio n in the extent o f patentin g across
industries. (See C h ap ter 14 b y M alerb a on in ter-in d u stry variatio n in in n o vation
processes.) O n the other h an d, the focus o f m an y o f these studies— sem icon d u ctors,
chem icals, biotechnology-— is in fields w here paten tin g is co m m on place, and c o m ­
petitors in these sectors are active patentors. T h e attention to these h ig h -tech n o lo gy
industries raises qu estion s, how ever, as to the generalizability o f the results to oth er
less k n o w led ge-in ten sive industries.
O ne stu d y that speaks to differences across indu stries is Row ley, Behrens, and
K rackh ard t’s (2000) analysis o f stron g an d w eak ties in the steel and sem icon d u ctor
industries. T h is stu d y m ad e a notable effort to d istin gu ish betw een stron g affilia­
tions, w here alliances entailed significant resource co m m itm en ts an d regular in ter­
actions, and m o re “ a rm s-le n g th ” tran sactions, w here there w as a rap id exchange,
and the relatio n sh ip w as characterized b y less freq u en cy and depth. For exam ple,
equ ity alliances, jo in t ventures, and R & D partn ersh ips w ere categorized as stron g
ties, w h ile licensin g, patent agreem ents, and m arketin g relations were classified as
w eak ties. R ecogn izin g that w eak ties serve as brid ges to n ovel in fo rm atio n , w hile
strong ties are u sefu l fo r b o th social co n tro l and the exchange o f tacit know ledge,
they find d ivergen t results. In the steel in d u stry stron g ties are positively associated
w ith p erfo rm an ce; w h ile in sem icon d u ctors weak ties are m o re efficacious. T h e y
suggest these fin d in g s reflect the im p o rtan ce o f search and p ro d u ct in n o vatio n in
sem icon d u ctors, an d a fo cu s o n im p ro vem en ts in the p ro d u ctio n process fo r ste e l
Much of the research on buyer-supplier relations and subcontracting has
focused on more traditional industries, such as automobiles or textiles. To be sure,
these industries make considerable use o f technological advances, but they are less
science-driven. As a consequence, the sources of relevant knowledge are not as
widely dispersed. Strong ties thus tend to predominate over weak ties. But the
content of those ties can evolve, changing from contractual to relational. Consider
JO W A L T E R W. P O W E L L A N D S T I N E G R O D A L ___________ ______

subcontracting relations, particularly one of the m o re notable examples—the

Toyota auto production network. Researchers stress h o w the d en sity o f overlapping
ties that connect this chain o f production facilitates know ledge sharing, mutual
learning, and fast responsiveness (D yer 1996; D yer and N o b eo k a 2000). But the trust
and reciprocity that characterize this dense network are the outcome of a long
developm ental process. In the 1950s and early 1960s, w hen Jap an ese firm s competed
on the basis o f lowest cost, relationships w ith su bcontractors w ere hierarchical and
asymmetric. As firms increasingly competed on the basis of q u ality and innovation,
however, complex multitiered supply relationships underwent significant change.
These relationships can remain hierarchical in two key respects: the larger lead firm
often has a significant financial stake in a su p plier o r affiliate, and it initiates the
production process. But the asymmetry has been sharply reduced. Suppliers, in an
effort to rem ain com petitive, m ake significant investm ents in new equipment,
constantly upgrade workers’ skills, and take on more critical aspects of the assembly
process (Helper et al. 2000). In turn, the larger firms offer long-term contracts, share
employees and provide technical assistance, and make financial investments to fund
equipment upgrades.
Too explicit a focus on formal, contractual linkages, however, neglects the myriad
informal ties that connect organizations. All kinds of informal interactions take
place between organizations, including participation in ad hoc industry commit­
tees, or executive education programs, conferences, trade association activities, and
the like. Personnel mobility and common educational backgrounds may also foster
informal linkages across firms. Such informal connections maybe the basis on which
m ore formal, contractual alliances are forged. Indeed, the success of formal affilia­
tions may hinge on the strength of informal ties. Thus we turn to a discussion of
noncontractual relations.

3.4*2 Informal Ties

Inform al patterns o f affiliation have long been a central top ic in sociology and
anthropology, where studies of friendship networks, advice and referral networks,
and communities are common. There is also a well -established strand of research in
organization theory that points out how informal relations within organizations are
often not closely aligned with form al authority (Dalton 1959; Blau 1963). A small line
o f work focuses on the impact of informal networks in large, multinational com­
panies (Ghoshal and Bartlett 1990; Hansen 1999). Relatively few studies, however,
link informal ties to the innovation process, and there is scant research on informal
m terorganizational relations.
^c^°^ars have o f t e n a r g u e d t h a t t h e s h a r i n g o f c o m p l e x i n f o r m a t i o n i s e n h a n c e d
у em bedded t ie s , which suggests t h a t i n f o r m a l ties have the potential t o make a
 N E T W O R K S OF I N N O V A T O R S J \

significant co n trib u tio n to in n o vation . There is a stron g sense am o n g researchers

that in fo rm al relations u n d ergird fo rm al ties. Powell et a l (1996) argue that, in the
life sciences, “ beneath m ost fo rm al ties lie a sea o f in fo rm al ties.” N evertheless, m any
organizations are largely unaw are o f the extent to which form al activities are
buttressed b y in fo rm al con n ection s (C ross, B orgatti, and Parker 2002).
O ne o f the key studies o f in fo rm al netw orks am o n g firm s w as Von H ippePs (1987)
w ork on the sh arin g o f p ro p rie ta ry in fo rm atio n am o n g U S steel m in i-m ill p ro d u ­
cers. Based on in terview s w ith plant m anagers and other engineers w ith direct
know ledge o f m an u factu rin g processes, he fo u n d that the trad in g o f p ro p rie ta ry
know ledge w ith b o th co op eratin g and rival firm s w as co m m on place. H e w as in itially
surprised that p ro p rie ta ry know ledge w as so “ leaky,” but he cam e to recognize that
in fo rm atio n exchange w as h ig h ly reciprocal an d co n d itio n ed o n expectations that
requests for help w o u ld be m et. M u ch o f the in fo rm atio n that w as shared focused on
p ro d u ctio n p ro b lem s, m atters o f p o llu tio n control and safety, and issues dealing
w ith in d u stry-w id e concerns. B u t w h en relationships am o n g engineers in rival firm s
were p articu larly d o se , m o re p ro p rie ta ry in fo rm a tio n w as exchanged. Von H ippel
also fo u n d that engineers h ad stron g n o rm s o f m em b ersh ip in a pro fession al
co m m u n ity that cu t across firm s, an d that in fo rm a tio n trad in g w as a m eans to
secure rep u tation an d status in that co m m u n ity. H e pro vid es n u m ero u s exam ples o f
h o w the sh arin g o f co m p le x in fo rm a tio n b y engineers con trib u ted to the p ro d u c tiv ­
ity o f m in i-m ills.
The cluster o f in d ivid u als that share a sim ilar set o f skills and expertise has been
dubbed a “ c o m m u n ity o f p ractice” (W enger 1998), o r a “ netw ork o f practice”
(B row n and D u g u id 20 0 1). S im ilar in som e respects to a technical co m m u n ity, o r
a sophisticated h o b b y club, these lo ose g ro u p s are engaged in related w o rk practices,
though they do n o t necessarily w o rk together. Su ch flu id gro u p s are im p o rtan t to the
circulation o f ideas. S axen ian (1994) o b served am ple sh arin g o f p ro p rie ta ry k n o w ­
ledge am o n g engineers in Silicon Valley, m an y o f w h o m have as stron g a c o m m it­
m ent to their peers w ith in the sam e o ccu p atio n al grou p as to their com panies.
Saxenian argues that in fo rm a l kn ow led ge sh aring, w id ely institu tion alized as a
profession al p ractice in S ilico n Valley, is one o f the cru cial factors co n trib u tin g to
its fertile in n o vative clim ate. C o h en and Fields (1999) stress that p ro fession al ties in
Silicon V alley are fo rged in co m p le x co llab o ratio n s betw een entrepreneurs, scien ­
tists, firm s, and asso ciation s, fo cu sed o n the p u rsu it o f in n o vatio n an d its co m m er­
cialization. T h is co llab o rative process generates an d refines the intangible raw
m aterial o f tech nical ch an ge— ideas.
Kreiner and Schu ltz (1993) analyze the im p o rtan ce o f in fo rm al ties th ro u gh in-
depth interviews with university researchers and industry research directors in the
D anish b io tech field. T h e y stress that successful co llab o rative R & D alliances w ith in
the Danish biotech industry are often based on informal ties. A barter economy,
w here m aterials, la b o ra to ry tests, ch em icals, etc. are exch anged, w as p ervasive in this
sector. T h e y sh o w th at n o rm s o f sh arin g in fo rm a tio n o n the fro n tier o f research aid
72 W A L T E R W. P O W E L L A N D S T I N E G R O D A L

in the form ation o f m ore form al networks. A s in the m in i m ills, information

exchange is not under m anagerial control, even th o u gh such reciprocal flows can
be channeled b y m anagerial actions.
Many studies o f inform al relationships stress the significance o f trust. Tsai and
Ghoshal (1998) studied the association between intrafirm netw orks and innovation in
fifteen business units o f a m ultinational electronics com pany. They found, not
surprisingly, that social ties led to a higher degree o f trustw orthiness am ong business
units. Trust increased resource-exchange and com bination betw een the business units,
which contributed to product innovation. The im portance o f trust also loom s large in
Uzzi’s (1997) analysis o f the difference between “ arm ’s-length” ties (“ a deal in which
costs are everything” ) and embedded ties (“ you becom e friends w ith these p e o p le -
business friends. You trust them and their work. T h ey re part o f the fam ily” ). Uzzi
conducted interviews and ethnographic observations at tw enty-three wom en s better-
dress firms in the N ew York C ity apparel industry. H is stu d y is notable not only for the
quality o f his data, but also for his attention to the perform ance consequences of
different kinds o f exchange relations. Uzzi fou nd that organizational performance
increases with the use o f embedded ties to netw ork partners, as these ties were superior
at conveying complex, context-dependent knowledge. H e argued, however, that a
balance between a firm ’s embedded ties and a firm ’s arm ’s-length ties needed to be
struck, because a network structure com prising only arm ’s-length ties or embedded
ties decreased organizational perform ance.
The significance o f a balanced netw ork structure, m ix in g fo rm al and informal
affiliations, is also em phasized in R u e f’s (2002) analysis o f entrepreneurship. He
found that individuals positioned in heterogeneous netw orks, com prising both
strong and weak ties, are m ore likely to be regarded as in n ovative by peers, in
com parison to entrepreneurs in m ore h om ogeneou s netw orks. Rosenkopf has
found a sim ilar interweaving o f form al and in form al relations in her research on
industry-w ide expert com m unities in the areas o f flight sim u lation and mobile
phones. Rosenkopf, M etiu, and G eorge (2001) analyze jo in t participation by cellular
service firm s in technical com m ittees, finding that such m em bership facilitates
subsequent form al interfirm alliance form ation. The effect o f participation in
technical com m ittees decreases when firm s have already established prior alliances,
suggesting that the effect o f inform al ties is m ore catalytic w hen firm s do not already
have established alliance partners.

34.3 Multi-Party Relationships

Most studies of networks and innovation have examined either dyadic ties, or a focal
firm in the context of an overall network. Rosenkopf and Tushman’s (1998) work
on expert communities emphasized the importance of studying multi-party
 N E T W O R K S OF I N N O V A T O R S 73

relationships that connect technical profession als across organizations. A kera (20m )
uses archival data to m ap the im p o rtan ce o f the IB M user grou p Share in the early
days o f co m p u tin g , o fferin g a p o rtrait o f a large m u lti-p arty collaboration. Share was
form ed soon after IB M 's release o f the first m ain fram e co m p u ter in 1953, to enable
IB M 's costum ers to sw ap p ro gram s, and collaborate on p ro g ram m in g so that
duplicative effort w as avoided. Som e o f the m ain in n o vation s that cam e ou t o f
Share p ro vid ed the basis fo r both system s p ro g ram m in g and o peratin g system s,
w hich to d ay fo rm the b ackbon e o f m o d ern com pu ter use. Share conveyed in fo r­
m ation about h ard w are changes that substantially im p ro ved the design o f IB M
com puters and periph erals. O ne o f the m ain co n trib u tio n s o f the Share netw ork
was the creation o f technical standards, but a secon d -ord er effect w as also the
d iffu sion o f kn ow ledge across com panies and betw een users.
A m odern an alogu e to Share is the L in u x co m m u n ity, fo u n d ed as a grou p o f users
tryin g to d evelop an alternative to the o p eratin g system su pplied b y M icro so ft. T h e
netw ork o f L in u x p ro g ram m ers has proven effective in d evelopin g softw are in a
h igh ly distrib u ted fashion . In the b egin n in g, m ost p ro gram m ers h ad never m et each
other and o n ly k n ew each o th er v irtu ally — b y the u sern am e they used w hen coding.
The L in u x c o m m u n ity has a v e ry m odest o rgan izatio n al structure, relying on a
co m b in atio n o f in terp erson al netw orks and an in d iv id u a ls rep u tation as a skilled
p ro gram m er to serve as the ad m issio n ticket to the n etw ork (O 'M a h o n y 2002; W eber
2003). O ne difference betw een the Share n etw ork and the L in u x n etw ork is that
corp orate interests d ro ve the Share netw ork, w h ile L in u x has been p rim a rily d riven
b y the end users. N evertheless, in both cases, the d eco m p o sab ility o f p ro g ra m m in g
tasks is an im p o rta n t facto r in facilitatin g distribu ted netw orks.
A n o th er large n etw o rk that has been w id ely studied is the n etw ork o f scientists,
often term ed in visib le colleges (C ran e 1972). A n invisible college is an in fo rm al
netw ork o f researchers w h o fo rm aro u n d a co m m o n p ro b lem o r p arad igm . B y
stud yin g in visib le colleges, C ran e (1972) h o p ed to u n d erstan d h o w know ledge
grow s and h o w the stru ctu re o f scientific co m m u n ities affects the exp an sion o f
know ledge. T h ere are n o w n u m ero u s studies o f scientific netw orks, m a p p in g the
structure o f co -a u th o rsh ip an d citation s (N ew m an 2003 pro vid es a go od o verview ),
th ou gh few attend exp licitly to the issue o f in n o vatio n . D av id (2 0 0 1), how ever,
develops a fo rm a l m o d el to sh o w that the liberal sh arin g o f know ledge w ith in the
scientific c o m m u n ity is a m a jo r d river o f sch o larly in n o vation . O ne o f the historic
characteristics o f scien tific co m m u n ities is that in fo rm a tio n an d research results
have been d istrib u ted o p e n ly am o n g m em b ers o f the relevant co m m u n ity. T h e shift
tow ard in creasin g research co m m ercializatio n b y universities has led som e scholars
to qu estio n w h eth er the in n o vative benefits o f in visib le colleges w ill persist, o r i f
co m m ercial interests w ill b lo ck in fo rm a l know ledge sh arin g a m o n g scientists
(Pow ell an d O w e n -S m ith 1998; O w en -Sm ith 2003). C h ap ter 8 b y M o w e ry and
Sampat o n u n iv e rsity -in d u stry interfaces offers a m o re d etailed d iscu ssio n o f the
ro fe-o f u n iversities in the in n o v a tio n process.
 W A L T E R W. P O W E L L A N D S T I N E G R O D A L
74

Another line o f research that has attended to scientific and technological net­
works, dubbed A cto r-N etw o rk T h eory (A N T ), exam ines h o w particular definitions
or configurations o f science and technology triu m p h o ver alternative conceptions,
A cto r-N etw o rk T heory is unique in its treatm ent o f artifacts and technologies, as
well as people and organizations, as m em bers o f a n etw ork (C allo n 1998; Latour
1987)* The p rim ary contribution o f A N T to the relationship betw een networks and
innovation is to show that not only can netw orks facilitate in n o vation , but they also
constrain it by determ ining the kind o f in n o vation s pro d u ced , their subsequent
interpretation, and their final use (C allon 2002).
A related line o f w ork that looks at netw orks as system s o f activity is the markets-
as-networks approach, developed by Scan dinavian m arketin g researchers (Hakans-
son and Snehota 1995), This approach exam ines the m u ltiple relationships among
organizations, and shows how these different aspects o f interorganizational rela­
tionships transform and evolve over tim e. For exam ple, su p p lier networks may
change frequently, with different elem ents o f p ro d u ctio n b ein g either outsourced
or insourced (Waluszewski 1995). The resources that are exchanged among the
partners in a production netw ork are constantly changing. W hat determines
whether an entity is a resource depends on the situation , and its use in combination
with other resources. Resources are, thus, always p o lyvalen t in both use and value.
The participants in a production network, both in d iv id u ally and collectively, de­
velop bonds characterized by trust and com m itm ent. These bonds also have an
organizing effect on networks, as they shape the identities o f actors, and account for
different levels o f com m itm ent am ong participants. T h is rich vein o f qualitative
research has not explicitly focused on innovation, how ever.
The various studies o f m u lti-p arty netw orks tend to em phasize the processual
aspects o f collaboration. This attention to content is w elcom e, but it sometimes
comes at the expense o f m easuring the output o f relationships, particularly how the
sharing and processing o f inform ation b y m em bers o f a netw ork can determine the
generation of novelty. A fuller understanding o f the in n o vatio n process needs to
exam ine the topic o f inform ation sharing, a subject to w h ich we now turn.

3-5 K n o w l e d g e T r a n s f e r

The role o f knowledge transfer is clearly central to the innovation process. Research
has highlighted two different aspects o f the know ledge-transfer process, each o f which
influences innovation, albeit in different respects. O ne explanation for the exchange
o f infoim ation through networks emphasizes the im portance o f complementary
 N E T W O R K S OF I N N O V A T O R S 75

assets in the d ivisio n o f in n o vative lab or (M ow ery, O xley, and Silverm an 1996)* I f
firm A is good at p ro d u cin g a specific co m p o n en t and firm B is capable o f using that
com p o n en t to p ro d u ce an engine, they collaborate in a jo in t p ro d u ctio n in which
their capabilities reinforce one another. In biotech n ology, fo r exam ple, sm all firm s
w ith close ties to u n iversity scientists m ay excel at d ru g d evelopm ent, but lack the
skills and resources to m anage o r fu n d costly clinical trials. B y w o rk in g closely w ith a
research h osp ital and an established firm that has a lim ited pip elin e o f new m e d i­
cines, the parties collab orate in a d ivisio n o f lab or that is m u tu ally rew arding, and
can result in the p articip an ts learn in g fro m one another, and accom p lish in g tasks
they could not do in d ivid u ally.
A second fo rm o f know ledge sh arin g occu rs w hen existing in fo rm atio n w ithin a
netw ork is reco m b in ed in novel w ays. Indeed, n o velty is often the unanticipated
result o f reco n figu rin g existin g know ledge, problem s, and solu tion s (N elson and
W inter 1982; Flem in g an d Soren son 2001). A s a consequence o f such collisions o r
tran spositions, firm s can generate so m eth in g they w ere unable to create on their
ow n. B oth fo rm s o f kn ow led ge tran sfer d epend on som e m an n er o f successful
exchange o f ideas, how ever.
A n o ft-u sed d istin ctio n is d raw n betw een tacit and explicit know ledge (C o w an ,
D avid, an d F o ra y 2000). Interest in tacit know ledge stem s fro m Polanyi ’s (1956)
argum ent that w e freq u en tly k n o w a go od deal m o re than w e can express verbally.
E xplicit kn ow led ge is h ig h ly codified , as in blueprin ts, recipes, m anuals, o r in the
form o f train in g. Tacit kn ow led ge lacks such extensive co d ificatio n (N on aka and
Takeuchi 1995). V aluable, p ro d u ctive know ledge often d em an d s considerable effort
to acquire, an d su ch kn ow led ge is freq u en tly altered in the process o f acq u isition and
application. Perh aps the m o st v iv id exam ple is the co n tin u in g effort o f U S a u to ­
m akers to acquire, u n d erstan d , and im plem en t the Jap an ese system o f lean p ro d u c ­
tion (W om ack, Jon es, an d R o o s 1990; D yer an d N o b eo k a 2000). K n ow led ge o f
co m p lex p ro d u c tio n tech nologies is rarely obtained in a fu lly digestible fo rm ;
un d erstan d in g in evitab ly entails learn in g b y using. T h e d istin ction betw een co d ified
and tacit is k e y becau se the latter d em an d s co n sid erab ly m o re trial-an d -erro r
learning to a p p ly the n ew k n o w led ge in a different setting.
M a n y studies p o in t to the relatively easy tran sferab ility o f exp licit know ledge in
contrast to tacit kn ow ledge. S im o n in (1999) show s that kn ow led ge tran sfer w ith in
alliances is n egatively affected b y b oth the nature o f know ledge an d differences in
organ izatio n al culture. H e observes im p o rtan t differences in know ledge exchange
betw een lo n g an d sh o rt-lived alliances. O ld er alliances d evelop a co m m o n language
and shared m en tal m o d els betw een partn ers, suggesting a learn in g cu rve w ith in
alliances w h ere the negative effects o f lack o f experien ce an d know ledge co m p le x ity
subside as the alliance m atu res. T h u s, as an alliance ages an d p articip an ts develop
relatio n sh ip -sp ecific u n d erstan d in g, there is the o p p o rtu n ity to co n vey m o re subtle
fo rm s o f in fo rm a tio n m o re effectively. C o m p le x tacit k n o w led ge can b eco m e m o re
exp licit as p a rtn e rs d evelo p a w id er b an d w id th o f co m m u n icatio n .
7б W A L T E R W. P O W E L L A N D S T I N B G R O P A L _ -------------------------------------_ _ _

• „ Kmitina factor in the transfer o f knowledge, then the

^ t o th , type o f
S ” t d C * i > widely dispersed a , a low cos, (B oiso, m S ) thus th.
M h o o d that explicit

к“ * Г а п Т 1,а іп а a large ,aci, component, the degree o f difficulty and the

Tosts Of transfer are high. Consequently, the expected gam s real,zed from tb .
„form ation are uncertain, as the cos, o f obtaining inform ation m ay exceed ,0
value This suggests tha, when knowledge involves a m oderate level o f complex,,,
the benefits derived from transfer may be greatest. Figure 3.4suggests a hypothetical
inverted U-shaped relationship between innovation and codification Here we
assume that there is variability in the cost o f inform ation transfer, and that the
greatest value may be derived when novel ideas are transm itted without too much

Szulanski’s (1996) analysis o f the transfer o f internal benchmarking efforts m

eight companies suggests key dimensions along which knowledge transfer can be

The costs of

Fig. 3.4 Knowledge codification and innovation

 N E T W O R K S OF I N N O V A T O R S 77

distinguished. H e dem onstrates that relationships between sender and receiver are
im portan t, in that b o th parties need m utual awareness o f state-of-th e-art practices.
O bviously, c o m m u n ica tio n is critical to in fo rm atio n exchange. But even when
relationships fu n ctio n well, som e know ledge is causally am b igu ou s, or stic k y and
thus not easily tran sferred. M oreover, in fo rm atio n exchange is hin dered w hen the
parties have differential levels o f ab sorp tive c a p a c ity that is, the ab ility to recognize
the value o f new in fo rm atio n , assim ilate it, and ap p ly it to com m ercial ends (C ohen
and Levinthal 1990). T h is cap acity is essential to in n o vative c a p a b ility F o r exam ple,
internal success w ith R 3 cD and R8cD expen ditures p o sitively affect a firm 's ab ility to
exploit the o p p o rtu n ities presented in external relations (C o h en and Levinthal
1990; Pow ell et al. 1996).
The p ro d u ctive tran sfer o f know ledge is also essential w hen tw o o r m o re o rgan iza­
tions are able to co m b in e their d ifferent capabilities, and create a pro d u ct o r service
that th ey w o u ld not be able to construct on their ow n. A g o o d illustration is the
Italian m o torcycle in d u stry w h ere the locus o f in n o vation is b ro ad ly dispersed,
and th rough co o p eratio n , the particip an ts benefit fro m specialization an d variety
generation (L ip p a rin i, L o ren zo n i, and Z o llo 2001). Because all the participan ts
p ro vid e valuab le inpu ts, there is a h igh co m m itm en t to know ledge generation.
The lead firm s d evelop relational capacities aim ed at p o o lin g the skills o f specialized
participan ts, h elp in g the overall flo w o f in fo rm atio n and resources in the netw ork.
A parallel analysis o f the Italian packagin g m ach in ery in d u stry stresses the creation
o f a su p p lier n etw ork in w h ich specialized roles are h ig h ly co m p lem en tary
(Loren zon i an d L ip p a rin i 1999). O ver tim e, m an agers o f the core com pan ies d e­
veloped a specialized su p p lier n etw ork and each p articip an t focused on a narrow ,
but h igh ly co m p etitive set o f core com petencies. T h is n etw ork structu re enabled
every step o f the su p p ly ch ain to specialize in im p ro vem en ts o f their specific
co m pon en t, th ereb y in creasin g the responsiveness o f the particip an ts to m arket
conditions.

3 .6 G o v e r n a n c e and In c e n t iv e Issu es

M any studies o f interfirm networks draw data from a single point in time and thus
do not examine how collaborations unfold over time. Even studies that do look at
dynam ics tend to do so from the perspective o f a dyad. Initially, the choice by a
young firm o f whom to partner with is often driven by resource needs. A s both firm s
and the network m ature, various dyadic choices increasingly reflect structural
properties o f the network. Thus, the existing network structure shapes search
78 W A L T E R W. P O W E L L A N D S T I N E G R O D A L

behavior. Consequently, networks both enable in fo rm atio n to becom e knowledge

and determ ine the nature o f knowledge (K ogut 2000). B u t w e do not, as yet, havea
parallel understanding o f the m anagem ent and governance o f netw orks to accom­
pany analyses o f structure and topology. C oncern s w ith h o w the parties in a
relationship adapt to changing circum stances, o r attend to the incentives to adjust
the relationship to make im provem ents rem ain largely u nexam in ed.
N ot surprisingly, then, m any studies assum e that as lo n g as a relationship
persists, the participants are achieving their goals, and v ie w term ination as a s ip
o f failure. Such a view misses the poin t that the relationship m ay have completed
its goal or outlived its usefulness. Inkpen and Ross (2001) find evidence that
the term ination o f alliances m ight not signify failure, b u t sim p ly be a sign of the
conclusion o f collaborative activity (for exam ple, a n ew pro d u ct is launched).
Nevertheless, some collaborations seem to persist even th o u gh they have stagnated
or outlived their usefulness. Inkpen and Ross suggest several reasons why interfirm
alliances m ay persist beyond their optim al d u ra tio n .4 A lliances can be difficult
and expensive to form ; thus, once established, there m ay be reluctance to disband
them. M oreover, as m ore firm s pursue alliances, a b an d w agon effect is created
and m any firm s jum p on it out o f fear o f being left w ith o u t a partner. There are
also challenges and costs associated w ith m anaging a partnership. I f the relationship
is poorly coordinated, the costs can outw eigh the benefits. Alliances can also
becom e synonym ous with a firm 's values, m aking them d ifficult to discontinue.
Finally, there m ay be costs associated w ith closing an alliance. A ll o f these factors
m ay contribute to alliances existing beyond the period w hen they create value for
a firm.
Several studies have pointed to problem s o f stagnation that can occur in some
long-term associations. A lthough this w ork does not deal d irectly with rates of
innovation, focusing instead on viability and su rvival, the general point is apt.
W hen the participants in a netw ork becom e too tigh tly knit and information
circulates only am ong a small group, netw orks can becom e restrictive and ossified.
Inform ation that cycles back and forth only am on g the sam e participants can lead to
lock-in or sclerosis. W hen networks turn inward and becom e restricted in terms of
access to new members, the possibility of “group think" increases. Grabber's (1993)
study o f steelm aking in the Ruhr illustrates h ow a h igh ly cohesive, homogeneous
region became so overem bedded that no producers opted for alternative strategies.
This cognitive lock-in eventually led to the decline of steelmaking in Germany.
Powell (1985) shows how the failure o f editors to ren ew and update their networks
leads to a decline in the quality o f a b oo k publisher's list and reputation. Fortes
and Sensenbrenner (1993) illustrate how ethnic community networks can become
restrictive and subject successful entrepreneurs to ostracism when they deviate
too m uch from com m unity standards. In sum , the ties that bind economic
actors together can become the ties that blind, thwarting recognition of preferable
alternatives.
 3*7 S u m m a r y

In terorganization al netw orks have grow n considerably in im p o rtan ce over recent

decades. N etw orks co n trib u te sign ifican tly to the in n ovative capabilities o f firm s by
exp osin g them to novel sources o f ideas, enabling fast access to resources, and
enhancing the transfer o f know ledge. F o rm al collab oration s m ay also allow a
d ivisio n o f in n o vative lab o r that m akes it possible fo r firm s to accom plish goals
they could not pu rsu e alone. Research on alliances d ocu m ents that investm ents in
m utual learn in g and a p o rtfo lio o f diverse collab oration s are associated w ith in ­
creased patenting. W h ile patentin g is o n ly an inpu t to the in n o vatio n process, the
strength o f these em p irical results h igh lights the im po rtan ce o f access to h etero­
geneous contacts.
We have argu ed that the nature o f know ledge, conceptualized in term s o f tacitness
or explicitness, is an im p o rtan t factor in d eterm in in g w hether m em bers o f a n etw ork
can effectively share in fo rm a tio n and skills. N etw orks that are rooted in a d ivisio n o f
innovative lab o r logic m ay fin d it easier to transfer tacit know ledge in the fo rm o f
finished inputs, w hile netw orks involved in the co -creatio n o f novel ideas m ay
succeed o r fail on the basis o f their ab ility to con vey and transfer ideas that are not
easily codified.
A n o th er central challenge to netw orks o f in n ovators is d evelopin g the capacity to
sim u ltan eou sly enhance the flo w o f in fo rm atio n am o n g cu rren t p articipan ts and be
open to new entrants. T h e tw in tests o f increasing cohesion w ith in the n etw ork and
recognizing p ro m isin g sources o f new ideas are d ifficult to su rm o u n t. Som e research
suggests that a m ixtu re o f stron g and w eak ties affords the p ro p er b len d o f reliability
and novelty.
M uch m o re research is needed, how ever, to ascertain h o w m ixtu res o f thick,
reliable affiliatio n s can be co m b in ed w ith n ovel linkages to new com ers. M oreover,
since som e affiliatio n s are essentially person to person ties, a greater u n d erstan d in g
is needed o f h o w relations a m o n g in d ivid u als are aggregated to p ro d u ctive relations
betw een co rp o rate actors. Or, in cases w here this issue is neglected, as in studies o f
contractual relation s betw een firm s w here the in fo rm al relational u n d erp in n in gs are
not analyzed, the o p p o rtu n ity to exam in e the in tertw in in g o f the careers o f in d iv id ­
uals and the strategies o f firm s is m issed. E q u ally im p o rtan t, the n etw ork literature
has n o t fo cu sed exp licitly on d ifferent m easures o f in n o vative o u tp u t, w h eth er it is
new p ro d u cts o r services, new m o d es o f organ izin g p ro d u ctio n , o r m o re rapid
response to co m p etitive d em an d s. T h e standard m easures are based o n either
patents, an in p u t to the in n o va tio n process, o r p ro b lem so lvin g, w ith o u t sufficient
attention to either the tim eliness o r the o p tim ality o f the so lu tio n . T h e n etw ork
literature is still relatively yo u n g, how ever, so it m a y be prem atu re to expect such
soph isticated answ ers. We look fo rw a rd to fu tu re research that o ffers a m o re
compelling an alysis of the sp ecific w ays in w h ich netw orks sh ape in n o vative o u tp u ts.
go W A L T E R W. P O W E LL AN D STINE G R O D A L

N otes __ _____ ___ ____ ________________ _ _ _ _ _ _

l The five that form ed the core o f the hum an sequencing com pon en t were the Whitehead
Institute, Washington University, Baylor College o f M edicine, the Joint Genome Institute
(a cluster ot three national laboratories at the D epartm ent o f Energy), and the Sanger
Institute in England.
2. Notable exceptions include Ahuja 2000#, Row ley et al. 2000, Powell et al. 2005.
3. Owen-Sm ith and Powell (2003) found that US universities w ith strong ties to a limited set
o f com m ercial partners had "fertile" patent portfolios, w ith fertility measured by the
impact o f patent citations. Those universities with few ties also had les fertile patents and
patented much less. The optim al strategy for research universities, w ith respect to patent
volum e and impact, appears to be one o f diverse ties to a w ide array o f industrial partners.
Diversity mitigates possible capture from too close relations w ith com m ercial firms.
4. Inkpen and Ross (2001) do not specify how to m easure the appropriate duration of
relationships. Instead, they assume that som e relationships becom e stale and dutiful
over time, and no longer generate benefits that outw eigh the costs o f sustaining them.

R e f e r e n c e s *

*A h u ja , G. (2000a), "C ollaboration Networks, Structural H oles, and Innovation: A Longi­

tudinal Study," A d m in istra tiv e Science Q u a rterly 45: 425-55.
------ (2000b), "T h e Duality o f Collaboration: Inducem ents and O pportunities in the For­
m ation o f Interfirm Linkages," Strategic M a n a g e m e n t fo u r n a l 21: 317-43.
A kera , A. (2001), "T h e IBM User G roup, Share," T ech n o logy a n d C u ltu re 42: 710-36.
B aum , J. A. C , C alabrese , X , and S ilv er m a n , B, S. (2000), "D o n 't G o It Alone: Alliance
Network Com position and Startups' Perform ance in C anadian Biotechnology," Strategic
M a n a g em en t Jo u rn a l 21: 267-94.
B eckm an , C , and H aunsch ild , P. (2002), "N etw o rk Learning: The Effects o f Partners
Heterogeneity o f Experience on Corporate A cquisitions," A d m in is t ra tiv e Science Quarterly
4 7 : 92-1724.
B lau , P. M. (1963), Consultation A m ong Colleagues," in D y n a m ic s o f B ureaucracy, Chicago:
University o f Chicago Press, 157-69.
Bo 1so t , M. H. (1998), K n o w ledge Assets: S e cu rin g C o m p e titiv e A d v a n ta g e in the Information
E conom y , Oxford: Oxford University Press.
B rown, J. S., and D uguid , P. (20m), "K now ledge and O rganization: A Social Practice
Perspective," O rgan ization Science 12(2): 198-213.
B rusco , S. (1982), rh eb m ilian Model: Productive D ecentralization and Social Integration,"
C a m b rid g e Jo u rn a l o f Econom ics 6 :16 7-8 4 .
♦ B urt , R. S. (,992), Stru ctu ra l H oles , Cam bridge, M ass.: H arvard U niversity Press.
C a i .lo n , M. (1998), Actor-Network Theory— The M arket Test,” in J. Law and J. Hassard
(eds.), A ctor N etw o rk T h eory a n d A fter, Cam bridge, M ass.: Blackwell, 186-96.

Asterisked items are suggestions for further reading.

 N E T W O R K S OF I N N O V A T O R S 8l

------ (2002), “ From Science as an Economic Activity to Socioeconomics o f Scientific Re­

search,” in R Mirowski and E. M. Sent (eds.), Science Bought and Sold, Chicago: University
o f Chicago Press.
C h esbro u g h , EL (2003), The Era of Open Innovation, Boston: Harvard Business School
Press.
C ohen , S.> and F ie l d s , G. (1999), “ Social Capital and Capital Gains in Silicon Valley”
California Management Review 41(2): 108-30.
C ohen , W., and L e v in t h a l , D. (1990), “Absorptive Capacity: A New Perspective on Learn­
ing and Innovation,” Administrative Science Quarterly 35:128—52.
C owan , R., D a v id , P. A., and F o ray , D. (2000), “ The Explicit Economics o f Knowledge
Codification and Tacitness,” Industrial and Corporate Change 9: 211—53.
C r a n e , D. (1972), Invisible Colleges: Diffusion of Knowledge and Scientific Communities,
Chicago: University of Chicago Press.
C ro ss , R., B o r g a t t i , S., and P a r k e r , A. (2002), “ Making Invisible Work Visible,” California
Management Review 44(2): 25-46.
C usum ano , M., M ylo n a d is , and R osenbloom , R. (1992), “ Strategic Maneuvering
and Mass-Market Dynamics: The Triumph o f VH S over Beta,” Business History Review
66: 51-94.
D alto n , M. (1959), “ Power Struggles in the Line,” in Men Who Manage, New York: Wiley,
71-109.
D a v id , P. A. (2001), “ Cooperation, Creativity, and Closure in Scientific Research Networks:
Modeling the Simpler Dynamics o f Epistemic Communities,” Conference at Centre Saint-
Gobain, Paris.
*D y e r , J. H. (1996), “ Specialized Supplier Networks as a Source o f Competitive Advantage:
Evidence from the Auto Industry,” Strategic Management Journal 17(4): 271-91.
------ and N o beo ka , K. (2000), “ Creating and Managing a High-Performance Knowledge-
Sharing Network: The Toyota Case,” Strategic Management journal 21: 345-67.
E c c les , R. (1981), “ The Quasifirm in the Construction Industry,” Journal of Economic
Behavior and Organization 2: 335-57.
E isen h a r d t , K. M., and S ch oo nh oven , С. B. (1996), “ Resource-based View o f Strategic
Alliance Formation: Strategic and Social Effects in Entrepreneurial Firms,” Organization
Science 7:136-50.
Fa r r e l l , J., and S a lo m er , G. (1988), “ Coordination through Committees and Markets,”
RAND Journal of Economics 19(2): 235-52.
F l em in g , L., and S o ren so n , O. (2001), “ Technology as a Complex Adaptive System:
Evidence from Patent Data,” Research Policy 30:1019-39.
*F reem an , C. (1991), “ Networks o f Innovators: A Synthesis o f Research Issues,” Research
Policy 20: 499-514.
G a lu n ic , D. C , and R odan , S. (1998), “ Resource Recombinations in the Firm: Knowledge
Structures and the Potential for Schumpeterian Innovation,” Strategic Management Jour­
nal 19:119 3-20 1.
G hoshal , $., and B a r t l e t t , C. (1990), “ The Multinational Corporation as an Interorgani-
zational Network,” Academy o f Management Review 15:561-85.
G odoe, H, (2000), “ Innovation Regimes, R&D and Radical Innovations in Telecommuni­
cations” Research Policy 29:1033-46.
G o m es -C a s s e r e s , B. (1996)r The Alliance Revolution: The New Shape of Business Ш ш іщ
Cambridge* Mass.: Harvard University Press.
82 W A L T E R W. P O W E L L A N D S T I N E G R O D A L

G r a b b e r G (1993), “ The Weakness of Strong Ties: The Lock-in of Regional Development in

1 the Ruhr Area," in G. Grabher (ed.), The Embedded Firm, London: Routledge.
____ and PoWE1Xi w . W. (2004), “Introduction,” in Critical Studies in Economic Institutions:
Networks, London: Edward Elgar.
* G r a n o v e t t e r , M . (1973), “ The Strength o f Weak Ties,” A m e r ic a n Jo u r n a l o f Sociology 78:
13б0™80.
____(i 985)( “ Economic Action and Social Structure: The P roblem o f Embeddedness;1
A m erica n Jo u rn a l o f Sociology 91: 481-510.
G r il ic h e s , Z. (1990), “ Patent Statistics as Econom ic Indicators: A Survey,” Journal of
E con om ic L itera tu re 28:1661-707*
G u lati , R. (1995), “ Does Familiarity Breed Trust? The Implications of Repeated Ties for
Contractual Choice in Alliances,” Academy of Management Journal 38: 85-112.
H agedoorn , J. (1995), “ Strategic Technology Partnering During the 1980s: Trends,
Networks, and Corporate Patterns in Non-Core Technologies,” Research Policy 24:
2О7-3І.
------ (1996), “ Trends and Patterns in Strategic Technology Partnering since the Early Seven­
ties,” R e v ie w o f In d u stria l O rgan ization 11: 6 0 1-16 .
------ (2002), “ Inter-Firm R8cD Partnerships: A n O verview o f M ajo r Trends and Patterns
since i9 6 0 ” Research P o lic y 31: 477-92.
------ , and D u y s t e r s , G. (2002), “ External Sources o f Innovative Capabilities: The
Preference for Strategic Alliances or M ergers and A c q u isitio n s” Jo u r n a l o f Management
Stu dies 39(2): 167-88.
H akansson , H „ and Snehota , I. (1995), D e v e lo p in g R e la tio n s h ip s in B usin ess Networks,
London: Routledge.
H an sen , M . T (1999), “ The Search-Transfer Problem : The Role o f Weak Ties in Sharing
Knowledge across Organization Subunits,” A d m in is tra tiv e S c ie n c e Q u a rte rly 44: 82-111.
H elper , S., M ac D u ffie , J. R, and Sa b el , C. (2000), “ Pragm atic Collaborations: Advancing
Knowledge While Controlling O pportunism ,” In d u s tr ia l a n d C o rp o ra te C h a n g e 9(3): 443*
89-
H ic k s , D. M., and K atz , J. S. (1996), “ W here is Science G oin g?” S cien ce , Technology and
H u m a n Values 21(4): 379-406.
I n kpen , A. C , and R o ss, J. (2001), “ W hy do Som e Strategic A lliances Persist beyond their
Useful Life?” C a lifo rn ia M a n a g e m en t R e v ie w 44:132-48.
K ogut , B. (2000), “ The Network as Knowledge: Generative Rules and the Emergence of
Structure ” Strategic M a n a g em en t Jo u rn a l 21: 405—25.
K r ein er , K., and Schultz , M . (1993), “ Inform al C ollaboration in R & D , The Formation of
Networks across Organizations,” O rg a n iz a tio n S tu d ies 14 :18 9 —209.
L am brigh t , W. H. (2002), M anaging ‘ Big Science’ : A Case Study o f the Hum an Genome
Project, Washington, D C: Pricewaterhouse C oopers Endow m ent for the Business of
Government.
Larson , A, (1992), Network Dyads in Entrepreneurial Settings: A Study o f the Governance
o f Exchange Processes” A d m in istra tiv e Science Q u a rterly 37: 76 -10 4 .
Latour , B. (1987), Science in A ctio n , C a m b rid g e , M ass.: H a r v a r d U n iv ers ity Press .
L eonard -B arton , D. (1995), W ellsprings o f K n o w led g e: B u ild in g a n d S u s ta in in g the Sources of
In n o v a tio n , Boston: HBS Press.
L i n k , A. (1996), “ Research Joint Ventures: Patterns from Federal Register Filings,” Review of
In d u stria l O rgan ization II (October): 617-28.
 N E T W O R K S OF I N N O V A T O R S 83

------ ( i 999)j “ Public/Private Partnerships in the United S tates” Industry and Innovation 6(2):
19 1-2 17 .
L ip p a r in i , A ., L o ren zo n i , G ., and Z ollo , M . (2001), “ D ual N etw ork Strategies: M anaging
K now ledge-based and Efficiency-based N etw orks in the Italian M otorcycle In d u stry ”
W orking paper,
Lo ren zo n i , G., and L ip p a r in i , A . (1999), “ The Leveraging o f Interfirm Relationships as a
D istinctive O rganizational C a p a b ility ” Strategic Management Journal 20: 317-38.
M c K e n d r ic k , D. G ., D o n n er , R. R, and H a g g a r d , S. (2000), From Silicon Valley to
Singapore: Competitive Advantage in the Hard Disk Drive Industry, Stanford, Calif.: Stan ­
ford U niversity Press.
M e r r il l , S. A ., and C oo per , R. S. (1999), “ Trends in Industrial Research and D evelopm ent:
Evidence fro m N ational D ata Sources,” in Securing America's Industrial Strength, W ash­
ington, D C : N ation al A cadem y Press, 9 9 -116 .
M erto n , R. K. (1957), “ The R ole-Set: Problem s in Sociological Theory,” British journal of
Sociology 8 : 110 -2 0 .
M o w ery , D. C. (ed.) (1988), International Collaborative Ventures in US Manufacturing,
C am bridge, M ass.: Ballinger.
------ (i999)> “A m e rica s Industrial Resurgence? A n O verview ,” in U.S. Industry in 2000:
Studies in Competitive Performance. W ashington, D C : N ational A cadem y Press, 1-16 .
------ O x l e y , J. E., and S il v e r m a n , B. S. (1996), “ Strategic A lliances and Interfirm K n o w ­
ledge Transfer,” Strategic Management Journal 17: 77-9 1.
M u r r a y , F. (2002), “ In n o vatio n as C o -evo lu tio n o f Scientific and Technological N etw orks:
E xp lo rin g T issue Engineering,” Research Policy 3 1:13 8 9 -4 0 3 .
N ational Science B o ard (1998), Science and Engineering Indicators— 1998. A rlington, Va.:
N ational Science Fou nd ation .
N elson , R., and W in t e r , S. (1982), A n Evolutionary Theory of Economic Change, Cam bridge,
M ass.: H arvard U n iversity Press.
N ew m an , M . (2003), “ The Structure and Fun ction o f C o m p lex N etw orks,” SIAM Review 45:
167-256.
N o n aka , L, and T a k e u c h i , N . (1995), The Knowledge-Creating Company, N ew York: O xford
U niversity Press.
N oteboom , B. (1999), Inter-Firm Alliances: Analysis and Design, Lon don: Routledge.
O’M a h o n y , S. (2002), The Emergence of a New Commercial Actor: Community Managed
Software Projects, Ph.D . dissertation, D epartm en t o f M anagem ent Science and Engineer­
ing, Stan ford U niversity.
O w en -S m it h , J. (2003), “ F ro m Separate System s to a H ybrid O rder: A ccum ulative A d v an ­
tage A cross P ublic an d P rivate Science at Research O ne Universities,” Research Policy 32(6):
10 8 1-10 4 .
-------and P o w ell , W. W. (2003), “ The E xp an d in g R ole o f U n iversity Patenting in the Life
Sciences,” Research Policy 32(9): 16 9 5 -7 11.
P io r e , M ., and S a b e l , C (1984), The Second Industrial Divide, N ew York: Basic Books.
Po lanyi , M. (1956), Personal Knowledge— Towards a Post-Critical Philosophy, Chicago:
University o f Chicago Press.
Po r tes , A ., and S e n s e n b r e n n e r , J. (1993), “ Em beddedness and Im m igration : N otes o n the
Social D eterm in an ts o f E co n o m ic A c tio n ” American Journal of Sociology 9 8 :13 2 0 -5 0 .
P o w b u , W. W* (1985), Getting Into Print: The Decision-Making Process in Scholarly Publish­
ing, Chicago: University o f Chicago Press.
84 W A L T E R W. P O W E L L A N D S T I N E G R O D A L ____________________

Powell , W. W. (1990), “ Neither M arket N or H ierarchy: N etw ork Form s o f Organization,” in

В. M. Straw and L. L. Cum m ings (eds.), Research in Organizational Behavior, Greenwich,
Conn.: JAI Press, 12: 295-336.
____ (1996), “ Inter-Organizational Collaboration in the B io tech n o lo gy Industry, Journal of
Institutional and Theoretical Economics 120(1): 197-215-
____ (1998), “ Learning from Collaboration: Knowledge and N etw orks in the Biotechnology
and Pharmaceutical Industries,” California Management Review 40(3): 228-40.
____ an(j B r a n t l e y , P. (1992), “ Com petitive C ooperation in Biotechnology: Learning
Through Networks?” in R. Eccles and N. Nohria (eds.), Networks and Organizations,
Boston: Harvard University Press.
* ____ K oput , K. W „ and S m ith -D o err , L. (1996), “ Interorganizational Collaboration
and the Locus o f Innovation in Biotechnology,” Administrative Science Quarterly 41(1):
II6-45.
_________ and O w en -S m ith , J, (1999)1 “ N etw ork Position and Firm Perform ance” in S.
Andrews and D. Knoke (eds.), Research in the Sociology of Organizations Greenwich,
Conn.; JAI Press, 16 :129 -59 .
------ and O w en -S m ith , J. (1998), “ Com m ercialism in U niversities: Life Sciences
Research and Its Linkage with Industry,” Journal of Policy Analysis and Management 17
(2): 253-77.
------ , W h ite , D. R„ K oput , K., and Ow en -S m ith , J. (2005), “ N etw ork Dynam ics and Field
Evolution: The Growth o f Interorganizational C ollaboration in the Life Sciences,” Ameri­
can Journal of Sociology.
R oberts , E. B., and Liu, W. K. (2001), “A lly or Acquire? H ow Technology Leaders Decide”
Sloan Management Review 43: 26-34.
R osen bloom , R. S., and S pen cer , W. J. (1996), “ The Transform ation o f Industrial Research ”
Issues in Science and Technology 12(3): 68-74.
*R osenkopf , L., and T ushman , M . (1998), “ The C oevolu tion o f C o m m u n ity Networks and
Technology: Lessons from the Flight Sim ulation Industry,” Industrial and Corporate
Change 7: 311-46 .
------ M etiu , A., and G e o r g e , V. P. (2001), “ From the Bottom Up? Technical Committee
Activity and Alliance Form ation,” Administrative Science Quarterly 46: 748—72.
R owley , T., B ehrens , D., and K ra ckh ard t , D. (2000), “ Redundant Governance Struc­
tures: An Analysis o f Structural and Relational Em beddedness in the Steel and Semicon­
ductor Industries,” Strategic Management Journal 21: 369—86.
R uee , M . (2002), “ Strong Ties, Weak Ties and Islands: Structural and Cultural Predictors of
Organizational Innovation,” Industrial and Corporate Change 11: 427—49.
S ab el , C. F. (1994), Learning by M onitoring: The Institutions o f Econom ic Development,”
in N. Smelser and R. Swedberg (eds.), Handbook of Economic Sociologyy Princeton:
Princeton University Press, 137-65.
S ar k a r , K. B., Ech am badi , R., and H a r riso n , J. S. (2001), “A lliance Entrepreneurship and
Firm Market Performance, Strategic Management Journal 21: 369—86.
S axen ian , A. (1994), Regional Advantage: Culture and Competition in Silicon Valley and
Route 128, Cambridge, Mass.: Harvard University Press.
S h a n , W., W a lk er , G., and K ogut , B, (1994), “ Interfirm C ooperation and Startup Innov­
ation in the Biotechnology Industry,” Strategic Management Journal 15:387-94 .
S im m e l , G. (1954), Conflict and the Web of Group Affiliations^ Glencoe, 111.: The Free Press.
 N E T W O R K S OF I N N O V A T O R S 85

S im o n in , B, L. (1999), "A m b igu ity and the Process o f Know ledge Transfer in Strategic
A llian ce s” Strategic Management Journal 20: 595-623.
St u a r t , T. E, (1998), "N etw o rk Positions and Propensities to C o lla b o ra te ” Administrative
Science Quarterly 43; 668-98,
------ (2000), "In tern rganizational A lliances and the Perform ance o f Firm s: A Study o f
G row th and In n ovation Rates in a H igh-tech nology In d u stry ” Strategic Management
Journal 21: 7 9 1-8 11.
S z u la n sk i , G. (1996), "E x p lo rin g Internal Stickiness: Im pedim ents to the Transfer o f Best
Practice w ith the F ir m ” Strategic Management Journal 17: 27-43.
T s a i , W., and G h o sh al , S. (1998), "S o cial C apital and Value Creation: The Role o f Intrafirm
N etw orks,” Academy of Management Journal 41: 464-76.
*U z zi, B. (1997), "S o cial Structure and C om petition in Interfirm N etw orks: The Paradox o f
E m b ed d ed n ess” Administrative Science Quarterly 42: 464-76.
V an W i j k , R ., van den B osch , F. A. J., and V o lberd a , H. W. (2003), "K n o w led ge and
N etw orks,” in M . Easterby-Sm ith and M . A. Lyles (eds.), Handbook of Organizational
Learning and Knowledge Management, O xford : Blackwell.
V in d in g ,A. L. (2002), "In tero rgan ization al D iffu sion and T ransform ation o f Know ledge in
the Process o f Product In n o vatio n ” Ph.D . Thesis, A alb o rg University.
V on H ip p e l , E. (1987), "C o o p eratio n between R ivals: In form al K n o w -h o w Trading,”
Research Policy 16: 29 1-30 2.
------ (1998), "E co n o m ics o f P rodu ct D evelopm ent by Users: The Im pact o f'S tic k y " Local
In form ation ” Management Science 44: 629-44.
*W a l k e r , G. B „ K o gu t , B „ and S h a n , W. (1997), "S o cial Capital, Structural H oles and the
Form ation o f an In d u stry N etw ork,” Organization Science 8 :10 9 -2 5 .
W a l u sz e w sk i , A. (1995), "G lu lam ,” in H. H akansson and I. Snethota (eds.), Developing
Relationships in Business Networks, Lon don: Routledge.
W e b e r , S, (2003), The Success of Open Source, C am bridge, M ass.: H arvard U n iversity Press.
W e n g e r , E. (1998), Communities of Practice, N ew York: C am bridge U niversity Press.
W o m ack , J. P., J o n es , D. T „ and R o o s , D. (1990), The Machine That Changed the World: The
Story of Lean Production, N ew York: H arp er 8c Row,
 C H A P T E R 4

INNOVATION
PROCESSES

KEITH PAVITT

4.1 In t r o d u c t io n 1

T h is c h ap te r co n c ern s in n o v a tio n p ro c e sse s w ith in fir m s , fo c u s in g m ain ly on

in n o v a tio n w ith in large c o rp o ra tio n s in a d v a n c e d c o u n t r ie s .2 W h a t d o w e know
a b o u t the h isto ric a l e v o lu tio n o f th ese in n o v a t io n p ro c e sse s a n d a b o u t the key
ch allen ges fa c in g “ in n o v a tio n m a n a g e rs " w ith in m o d e r n in d u s tr ia l corp oration s?3
T h e ch a p te r d raw s o n e m p iric a l stu d ies o f in n o v a t io n p ro c e sse s, b e a rin g in m in d the
d ifficu ltie s fo r g en era lizatio n p o se d b y th e h ig h ly c o n tin g e n t n a tu re o f innovation.
S e c tio n 4 .2 p resen ts a sh o rt in tro d u c tio n to th e m a n y th e o rie s a n d e m p iric a l studies
o f in n o v a tio n 4 an d su ggests a sim p le fr a m e w o r k fo r d is a g g re g a tin g the many
in n o v a tio n a c tiv itie s w h ic h tak e p la c e at th e fir m le v e l. T h r e e b r o a d , overlapping
su b p ro cesses (n o t stages) o f in n o v a tio n are id e n tifie d : th e p r o d u c t io n o f knowledge;
the tr a n sfo rm a tio n o f k n o w le d g e in to a rtifa c ts— b y w h ic h w e m e a n products,
system s, p ro cesses, an d services; an d th e c o n tin u o u s m a t c h in g o f th e latter to market
n eed s an d d e m a n d s.5 Sectio n s 4.3 th ro u g h 4.5 e x a m in e k e y a sp e c ts o f each o f these
three su b p ro cesses, sh o w in g h o w each h as e v o lv e d h is to r ic a lly a n d w h y they pose
su ch d ifficu lt p ro b le m s to r in n o v a tio n -re la te d m a n a g e rs , e n tre p re n e u rs, research­
ers, an d w o rk ers. T h e c h a p te r id en tifies th ese m a n a g e m e n t d iffic u ltie s a n d points to
so m e o f the strategies firm s h ave d e p lo y e d to m eet th e se c h a lle n g e s.
 INNOVATION PROCESS 87

4.2 C o r p o r a t e In n o v a t io n P r o c e s s e s

Since there is m o re than o n e process o f in n o vation , there is no easy w ay to organize

this chapter. In n o vatio n processes d iffer in m an y respects accordin g to the econ om ic
sector, field o f know ledge, type o f in n o vation , h istorical period and co u n try con-
cerned. T h e y also vary w ith the size o f the firm , its corp orate strategy or strategies,
and its p rio r experience w ith in n ovation . In other w ords, in n o vation processes are
“ contingent.” 6 These d ifficulties are co m po u n d ed b y the fact that there is no w idely
accepted th eo ry o f firm -level processes o f in n o vation that satisfactorily integrates
the cognitive, organ izatio n al, and econ om ic dim en sion s o f in n ovation processes in
firm s. E co n o m ists tend to concentrate on the econom ic incentives for, and the
effects of, in n o vatio n (largely ign o rin g w h at happens in betw een). O rganizational
specialists focu s on the structu ral and p ro ced u ral correlates o f in n ovative activities
and processes, so cio logists on their social determ inants and consequences. M a n ­
agerial specialists address the practices m ost likely to lead to com petitive success,
psych ologists m ay exam in e the p h en o m en on o f creativity o r the w ays in w hich
people's visio n is restricted to one o r other set o f o p p ortu n ities. R ich accounts and
data resources have com e fro m these and oth er lines o f w o rk over the past decades.
E m p irical evidence and theoretical u n d erstan d in g have been am assed b y h istorians,
from su rvey researchers and those concerned w ith b ib liom etrics, patenting, and
other quan tifiab le d im en sio n s o f inn ovation . A grow in g n u m ber o f “ in n o vation
studies” show s little allegiance to an y p articu lar discipline, and w id ely disparate
theories and m ethods coexist in relevant jou rn als and h andbooks.
Joseph S ch u m p eter is considered a p io n eer in the econ om ic analysis o f in n o v ­
ation, h avin g concen trated m o re effort on this top ic than an y other econom ist in the
first h a lf o f the tw entieth century. H is insights have guided the subsequent d evelo p ­
m ent o f the field, and h elped to explicate the vital role o f in n o vatio n in grow th and
com petitiveness. B u t Sch u m p eter's early w o rk in p articu lar (see Fagerberg 2003;
Fagerberg, C h . 1 in this v o lu m e ), stressed the role o f in d ivid u als, rather than
o rgan ization s, in the in n o va tio n process, h igh ligh tin g the character and d eterm in ­
ation o f o u tstan d in g in d ivid u als, and d efin in g in n o vatio n s as “Acts o f W ill” rather
than “A cts o f Intellect.” H is interpretation reflected the nature o f the evidence
available to h im at the tim e. A s the au th o rs o f the S A P P H O study (R oth w ell et al.
1974) fo u n d , the vast m a jo rity o f earlier studies o f successful an d unsuccessful
in n o vatio n w ere p erson al m em o irs and anecdotes o f the exp loits o f in d ivid u al
scientists, in ven to rs o r m an agers and contained little or no system atic co m p ariso n
o r analysis. P ro ject S A P P H O an d m a n y sim ilar, subsequent studies w ere in part
attem pts to o verco m e the “ in d iv id u a list” bias o f p erson al m em o irs. T h e y set ou t to
exam in e a w id e r set o f o rgan izatio n al factors, as w ell as the skills and experien ce o f a
w id e r ran ge o f in d iv id u a ls p a rticip a tin g in each in n o vation .
88 KEITH pavitt

In order to understand this rich but poten tially co n fu sin g m osaic o f knowledge
about innovation processes, I suggest the fo llo w in g general fram ew ork:
(i ) Innovation processes involve the exp loratio n and exp loitatio n o f opportun­
ities for new or im proved products, processes o r services, based either on
an advance in technical practice (“ k n o w -h o w ” ), o r a change in market
dem and, or a com bination o f the two. In n o vatio n is therefore essentially a
m atching process. The classic paper on this subject is M o w e ry and Rosenberg

(t 979 )*
(2) Innovation is inherently uncertain, given the im p o ssib ility o f predicting
accurately the cost and perform ance o f a new artifact, and the reaction of
users to it. It therefore inevitably involves processes o f learn in g through either
experim entation (trial and error) o r im pro ved un d erstan d in g (theory). Some
(but not all) o f this learning is firm -specific. T h e processes o f competition in
capitalist markets thus involve pu rposive exp erim en tation through competi­
tion am ong alternative products, system s, processes, and services and the
technical and organizational processes that deliver them .
In organizing the evidence, it is useful to divid e in n o vatio n into three, partially
overlapping processes, consistent with the tw o features described above.7 Each
process is closely associated with contributions fro m particu lar academ ic discip­
lines. Each process also has undergone m ajor h istorical transform ations as the
process o f innovation has evolved.
The production o f scientific and technological knowledge: a m ajo r trend, since the
industrial revolution, has been for the pro d u ction o f scientific and technological
knowledge to have becom e increasingly specialized, b y discipline, by function and
by institution. H istory and social studies o f science, tech n o lo gy and business have
contributed significantly to o u r understanding o f this tran sform ation .
The translation of knowledge into working artifacts: in spite o f the explosive growth
in scientific knowledge in recent years, th eo ry rem ains an insufficient guide to
technological practice. This reflects an underlying trend for grow in g complexity of
technological artifacts,8 and in the know ledge bases u n d erp in n in g them. Techno­
logical and business history have m ade m ajo r co n trib u tio n s here as have the
cognitive sciences m ore recently.
Responding to and influencing market demand: this involves a continual process of
m atching working artifacts with users' requirem ents. T h e nature and extent o f the
opportunities to transform technological know ledge into useful artifacts vary
am ongst fields and over tim e, and determ ine in part the nature o f products,
users and m ethods o f production. In the com petitive capitalist system , corporate
technological and organizational practices со -evolve w ith m arkets. Social change
and Innovations in m arketing and m arket research have contributed to complex
problem s and equally com plex solutions to the challenge o f m atching techno­
logical opportunities with m arket needs and o rgan ization al practices. These
 INNOVATION PROCESS 89

processes are central concerns o f scholars in m anagem ent, econ om ics and
m arketing studies.

I n o w discuss the im p licatio n s o f change over tim e in the structure and nature o f
each o f these features o f in n o vatio n processes. Such change presents considerable
challenges to the m o d ern in n o vatio n m anager and to the co rp oratio n as a whole.

4.3 T h e P r o d u c t io n o f S c ie n t if ic a n d

T e c h n o l o g ic a l K n o w l e d g e

All the im provem ents in m achinery, however, have by no m eans been the inventions o f
those w ho had occasion to use the m achines. M a n y ...h a v e been m ade by the m akers
o f the m achines, w hen to m ake them becam e the business o f a peculiar trade: and som e
b y . . .th o se w ho are called philosophers, or m en o f speculation, w hose trade is not to do
anything but to observe everything: and w ho, u pon that account are often capable o f
com bining together the pow ers o f the m ost distant and dissim ilar o b je c ts.. . , Like every
other employment, . . it is subdivided into a number of different branches, each of which affords
occupation to a peculiar tribe or class of philosophers; and this subdivision of employmen t
in philosophy as well as in every other business, improves dexterity and saves time. (Sm ith
1937 (orig. 1776): 8, m y italics)

A d am Sm ith s id en tificatio n o f the benefits o f specialization in the p ro d u ctio n o f

know ledge has been a m p ly co n firm ed by experience. P rofession al education, the
establishm ent o f lab orato ries, and im provem en ts in techniques o f m easurem ent and
exp erim en tation have increased the efficien cy o f discovery, invention , an d in n o v ­
ation. In creasin g ly d ifficu lt p ro b lem s can be tackled an d solved.9 N ew and useful
fields o f k n o w led ge have been d eveloped, pu n ctu ated by the perio d ic em ergence o f
fields w ith rap id rates o f tech n ological advance and o fferin g rich o p p o rtu n ities fo r
co m m ercial exp lo itatio n . F o r exam ple, progress in such activities as m etal cuttin g
and fo rm in g , o r the use o f n ew p o w er sources, has been in fo rm ed b y an d has draw n
u p on ph ysics, ch em istry, an d biology, as w ell as a variety o f related engineerin g
disciplines. Today, the c o o rd in a tio n o f increasing specialization rem ains a fu n d a ­
m ental task o f the large co rp o ratio n .
T h ree fo rm s o f co rp o rate sp ecialization have d eveloped in p a ra lle l First is the
d evelo p m en t in large m a n u fa ctu rin g firm s o f R & D lab orato ries specialized in the
p ro d u c tio n o f k n o w led ge fo r co m m ercial exp loitatio n . Second is the d evelopm en t of
a m yriad o f sm all firm s p ro v id in g co n tin u o u s im p ro vem en ts in specialized p ro d u ­
cers’ go od s. A th ird tren d o f sp ecialization is the ch an gin g ^ d ivision o f labor”
between private k n o w led g e d eveloped an d app lied in business firm s, an d public
90 KEI TH P A V I T T

knowledge developed and dissem inated b y universities an d sim ilar in stitu tm *

Taken together, all these form s o f specialization have co m b in ed to m ake a hetero­
geneous and path-dependent pattern o f technical change that places great demands
on corporations for coordination o f processes w ith in their b o u n d aries and between
these organizations and others external to the firm . These processes of change,
exam ined below in m ore detail, have intensified and b ro ad en ed the challenges facing
managers, “ intrapreneurs” and entire co rp o ratio n s.10

4.3.1 Functional Specialization and Integration:

Industrial R&D Laboratories
A m ajor source o f innovation in the twentieth cen tu ry w as the industrial R&D
laboratory, which remains im portant in the tw enty-first century. It emerged first in
Germ any in the chem ical industry and in the U S A in the electrical industry, for two
reasons. It was part of the more general process of functional specialization of the
large m anufacturing firm (M ow ery 1995; see also the ch ap ter b y Bruland and
Mowery in this volume), which itself emerged from the exploitation of economies
of scale and speed made possible by radical innovations in materials processing and
forming, and in power sources (Chandler 1977). But the industrial research labora­
tory also provided these firms with a means for exploiting the rich veins of useful
knowledge emerging from fundamental advances in chemistry and physics. In
addition, these new in-house laboratories served as a “monitoring post” for estab­
lished firm s seeking to acquire new technologies from other firm s.
Mowery (1995) has shown for the USA that a growing proportion of industrial
R & D in the twentieth century was integrated within large manufacturing firms,
rather than in independent companies. Until about 10 years ago, business-funded
R & D in all OECD countries was almost exclusively performed within innovating
firm s (not only manufacturing firms, since telecommunications and some other
services have long undertaken R&D). Mowery explain ed this lack of vertical disinte­
gration by the difficulties of writing contracts for an activity whose output is
uncertain and idiosyncratic, and pointed out that integration of R&D within the
firm reflected important operating advantages as well. Thus competitive advantage
can be gained by the effective combination of specialized and often tacit knowledge
across functional boundaries, within the individual firm . Accumulated firm-specific
experience is very im portant.11
is that one
A robust conclusion em erging from research on in n o vatio n processes
o f the most important factors differentiating successful from unsuccessful innov-
ation has been the degree of collaboration and feedback between product design and
other corporate functions, especially manufacturing and marketing within the firm
(Rothwell 1992). M any product designs turn out to be technically difficult (even
 INNOVATION PROCESS 91

im possible) to m an u factu re, an d /or fail to take into account often elem entary user
requirem ents (Forrest 19 9 1).
From a co rp o rate strategy perspective, the im po rtan ce o f such in trafirm co llab o r­
ation has increased the im p o rtan ce o f cross-fu n ction al integration sp an n in g d ep art­
m ental b o u n d aries. Japanese au tom ob ile firm s pioneered the use o f “ h eavyw eigh t”
project m anagers, em po w ered to control resources across the firm , rep ortin g d ir­
ectly to the sen ior m an agem en t team at the sam e level as the departm ental m anager
(C lark and F u jim o to 1992). These project m anagers were in turn linked to in n o v ­
ation processes w ith in cu stom ers and key suppliers, enabling fast, project-based
innovation. These “ h eavyw eigh t” project m anagers o ccasion ally clash w ith fu n c­
tional bosses, som e o f w h o m are u n w illin g to “ give u p ” control over their resources
and object to project-led m anagem ent. M a n y large firm s n o w p ro vid e form al
training in p ro ject m an agem en t to their p rofession al project m anager, covering
such issues as m an agem en t o f fast-m o vin g project team s responsible fo r integrating
research o u tp u ts, con cep tu al and detailed design, and vario u s engineering fu n c­
tions, w hile at the sam e tim e resp o n d in g to ch an gin g o r em erging cu stom er req u ire­
m ents d u rin g the p ro d u c tio n process.
M an y w riters stress the im p o rtan ce o f person al contacts an d exchanges across
functions w ith in the firm to deal w ith tacit elem ents o f both pro d u ct design and its
successful tran sfer to m an u factu re and m arket. There is no perfect o r fo o lp ro o f
process fo r en su rin g effective co o rd in atio n . Indeed, so called “ best practice” can be
p ositively h arm fu l w h en its ap p lication is taken too far. Excessive use o f “ h eavy­
w eigh t” project m an agers can lead to the loss o f such benefits from integration as
econom ies o f scale, and cost red u ction s fro m the use o f co m m o n co m pon en ts (e.g.
in au to m o b ile d evelopm en t; see L e o n a rd -B arto n 1995). Firm s can find it h ard to
decide w h at to d o w ith a h eavyw eigh t p ro d u ct m anager (and the associated staff)
w hen a p ro d u ct d evelo pm en t is failing. Failure to grasp this d ifficu lt nettle can lead
to the p ro b lem o f “ escalation” o r an in ab ility to term inate a failin g p ro je c t.12
M an agin g the trad e-o ffs betw een project and fu n ction al m anagem ent, and o ver­
co m in g the inh eren t difficu lties in project-b ased m anagem en t, present m a jo r d iffi­
culties to sen io r tech n o lo g y m anagers.

4.3.2 Technological Convergence and Vertical Disintegration

in Production Techniques
Even in industries with heavy investments in product innovation, however, some
“vertical disintegration” (outsourcing of specific activities to supplier firms) in
manufacturing process innovation has occurred since the nineteenth century,
often stimulated by technological advances. Rosenberg (1976) highlighted the emer­
gence o f specialized machine tool firms in the nineteenth-century US economy as a
92 KEI TH PAVI TT

result o f advances in metal cutting and m etal-fo rm in g techniques that produced

technological “ convergence” in operations that w ere co m m o n to a number of
m anufacturing processes. For exam ple, b orin g accurate circu lar holes in metal was
an operation com m on to the m aking both o f sm all arm s an d o f sew ing machines.
Although the skills associated w ith such m ach in in g o p eratio n s w ere often craft-
based and tacit, their outputs could be codified and standardized. T h e demand for
such com m on operations grew sufficiently that m arkets fo r m achines to perform
them became large enough to sustain the grow th o f specialized firm s who designed
and m ade such machines. Large m anu factu ring cu stom ers could therefore buy
machines that incorporated the latest im provem ents (d raw in g on the feedback
from m any users) and were far superior to what they could do b y themselves.
Sim ilar processes o f technological convergence and vertical disintegration have
been com m on features o f tw entieth-century capitalist d evelopm en t (see Table 4.1).
New opportunities have emerged from breakthroughs that h ave created applications
spanning m ultiple product groups. Exam ples include: m aterials shaping and
form ing, properties o f materials, com m on stages o f co n tin u o u s processes, storage
and m anipulation o f inform ation for controlling vario u s business functions such as
m anufacturing operations and design.
Lundvall (1988) and other writers show that the links betw een the (often small)
firm s providing these specialized production inputs and their (m ain ly large) cus­
tom ers are often “ relational” rather than arm s-len g th , and include considerable
exchange o f inform ation and personnel related to the developm ent, operation and
im provem ent o f the specialized inputs. M anaging the o u tso u rcin g o f these critical
inputs has becom e a m ajor challenge to m anagers o f large firm s (Q uinn 2000). For

Table 4.1 Examples of technological convergence and vertical disintegration

Underlying technological advance Technological convergence Vertical disintegration

Metal cutting ft forming Production operations Machine Tool Makers

Chemistry ft metallurgy Materials analysis ft testing
Chemical engineering Process control
Computing Design
Repeat operations
N e w M a te ria ls
Building Prototypes
ICT
Application software
Production systems
Contract Research
Instrument Makers
Plant Contractors
CAD Producers
Robots Makers
Rapid Prototyping Firms
Knowledge-Intensive
Business Services
C o n tra c t m anufactu re
 INNOVATION PROCESS 93

exam ple, logistics and IC T system s often d iffer betw een co m po n en t su ppliers and
integrator com pan ies, creating seriou s (albeit often technically sim ple) problem s for
co m m u n icatio n and transactions am o n g these firm s. M ore fundam entally, the
choice o f w h ich activities to ou tsou rce and w hich to retain in-h ouse defines the
“ core co m p eten ce” o f the m o d ern co rp o ratio n , m o ld in g the bou n d aries o f the firm
(H am el an d P rah alad 1994; D avies et al. 2001),

4*3.3 Industrial Linkages with Universities

A s in n ovative activities in business firm s have becom e m ore profession alized, and
u n iversity research m ore specialized, universities n o w p lay an im p o rtan t role in
p ro vid in g the train ed researchers for firm s in som e sectors to p erfo rm their in n o v a ­
tive activities. A t the sam e tim e, firm s have fo u n d it im p o rtan t to have effective
processes in o rd er to benefit fro m progress in those lon ger-term research p ro gram s
in universities in fields that have possible im pacts on their cu rrent and future
activities.
The range o f in teraction s betw een firm s and universities is considerable. At one
extrem e, there is so m eth in g close to the so-called (but relatively rare) “ linear m odel.”
H ere, fu n d am en tal research b y a u n iversity scientist leads to a discovery, its practical
im portan ce is recognized b y a business firm , w h ich m ay collaborate w ith the
u n iversity scientist in o rd er to exp loit it. T h is happens m ost often in science-based
industries in clu d in g the chem ical, biotech nology, an d ph arm aceu tical sectors,
w here the focu s is o n the d isco very o f interesting and useful synthetic m olecules.
A t the other extrem e, the p ro visio n o f trained researchers, fam iliar w ith the latest
research tech n iq u es and integrated in international research netw orks, is im p o rtan t
to firm s. It is ran ked b y m an y indu strialists as the greatest benefit p ro vid ed b y
universities (M a rtin an d Salter 1996). T h us, even i f u n iversity research in m echan ical
engineering has few er direct ap p lication s than research in chem istry, it still pro vid es
m echanical engineers train ed (fo r exam ple) in those sim u latio n and m o d ellin g
techniques that are in creasin gly im p o rtan t in the design an d d evelopm ent o f
au tom o b iles and aero-engines.
In betw een are a v ariety o f other, often com plem en tary, processes w h ich have to be
m an aged in o rd er to lin k u n iversity research w ith in d u strial in n o vatio n , in clu d in g
direct in d u strial fu n d in g o f u n iversity research, u n iversity-b ased consultants, and
exchanges o f research person n el. T h ree co m m o n features o f u n iv e rsity -firm links
emerge from the literature.
(1) T h e im p o rta n ce o f p erso n al and o ften in fo rm al contacts. In fo rm al relatio n ­
sh ip s give p ractitio n ers en try p o in ts into the academ ic w o rld , people w h o
th ey ca n ask a b o u t w h ere the im p o rtan t d evelopm ents lie an d w h o the
relevan t p eo p le are. Su ch relation sh ips give researchers insigh t in to the
94 KEITH PAVITT

problem s that are confronting industry, h o w the leading edge o f corporate

practice is developing, and so on. The in fo rm al relationships can result in
form al outputs that can in turn trigger m ore in fo rm al contacts. For instance,
industrial publications in the scientific literature can be seen as signals to the
wider academ ic com m u nity o f fields and p ro b lem s o f in du strial interest that
w ould benefit from m ore intense personal exchanges (H icks 1995).
(2) M uch university research that is useful to indu strialists also is valued by
academics. Som e academ ics erect a d em arcation betw een industrially rele­
vant applied w ork and m ore fundam ental research, and such a demarcation
m ay apply to the w ork o f som e university grou p s p layin g a role in technology
transfer to local businesses. But it does not ap p ly to a go od deal o f more
advanced research activities. U S-based studies b y M an sfield (1995) and Marin
et al. (1997) suggest that a high p ro po rtion o f in d u strially significant research
is publicly funded, perform ed in the academ ically prestigious research uni­
versities, and published in high quality academ ic jou rn als.
(3) The practical benefits o f m ost university research em erge from processes that
are roundabout and indirect. Probably the m ost frequent contribution is the
provision o f graduates trained by leading researchers, and often conversant
with emergent research methods and approaches. Such individuals can be
“ carriers” o f new theoretical insights, new techniques and observations, and
new skills, all o f which industrial firm s find difficult to provide themselves.
Over time, they will turn these capabilities to solving the problem s that their
employers face— or provide support by m eans o f form in g spin-offs or joining
consultancies that provide innovation-supporting services to industrial clients.
Over the past twenty years, governm ents have begun to expect greater direct useful­
ness from university research. O ften this has been su p ported b y em pirically ques­
tionable assum ptions and theories,13 or an incom plete und erstand in g o f the indirect
benefits actually valued by industrialists. For entirely different reasons, certain fields
o f university research— m any fields o f b iotech n ology and som e o f software and
related activities now provide an increasing stream o f inventions with potential
industrial application. These are reflected in increases in u n iversity licensing activity,
in university-founded sp in -o ff firms, and increases in private fu n din g o f university
research. In Chapter 8, M ow ery and Sam pat discuss the nature and im plications of
these recent developments.
These university-industry relationships can be extrem ely difficult for firms to
manage. 4 M anagers often com plain that universities operate on extended “ time
lines with little tegard for the urgent deadlines o f business. Th erefore, they argue,
universities should not be placed on the critical path o f any im portant projects.
Universities, in turn, som etim es find themselves in the in vid io u s position o f being
viewed as a low-cost perform er o f industrial projects, often w ith the encouragement
o f governm ent and research council program s o f “ tech n o logy transfer.” At worst,
 G L O B A L I Z A T I O N OF I N N O V A T I O N 325

States, hom e o f a n u m ber o f MNEs w hich have a relatively low p ro po rtion o f their
R& D and patenting activity ab ro ad .5
On average, firm s fro m E U countries obtain a larger share o f patents from their
foreign subsidiaries than is true o f U S or Japanese com panies (Table 12.6). D uring
the 1969-95 p eriod, the share o f total patents o f EU firm s attributable to foreign
affiliates grew from 26.3 to 32.5 per cent. European firm s tend to concentrate a
considerable share o f their international R8cD activities in the U S (over 50 per cent
o f their foreign R & D investm ent on average, w ith G erm an, British, and Swiss firm s
showing the highest concen tration o f their foreign activities in the United States).
The foreign patenting activity o f U S firm s also increased during this period, but
rem ained below 10 per cent.6 A lth ough U S foreign R & D activities are relatively low
com pared to E U firm s, they are m uch larger than Japanese com panies, w hose
offshore patenting declined from 2.1 per cent in 19 6 9 -77 to approxim ately 1 per
cent o f their total patentin g activity in 1987-95.
O verall, M N E s have increasingly internationalized their innovative activities,
with a few relevant exceptions (m ost notably, Japanese M N E s). The im portance o f
R & D activities o f foreign affiliates has grow n in m ost host econom ies over the 1990s.
R & D b y foreign firm s is especially high in the U K , Ireland, Spain, H ungary, and
Canada, and low est in Japan , w ith other countries (including the U S, France, and
Sweden) in interm ediate positions. N evertheless, m ost R & D and patenting activities
are still largely concentrated in the M N E s' hom e countries, and in a few host
countries. Well over 90 per cent o f the R & D expenditures o f m ost M N E s is located

Table 12.6 Share of US patents of the world's largest firms attributable to

research in foreign locations by main area of origin of parent firms,
1 9 6 9 - 1 9 9 5 (% )

Nationality of parent firm 1969-77 1978-86 1987-95

US 5.4 6,9 8,3

Japan 2.1 1,2 1.0
European countries3 26.3 25.6 32.5
Total alt countries1*' 10.3 10.7 11,3
Total all countries excluding Japan 11,1 13.0 16.2

Notes*
a Germany, UK, Italy, France, Netherlands, Belgium, Luxembourg, Switzerland, Sweden,
Denmark, Ireland, Spain, Portugal, Greece, Austria, Norway, Finland.
b Total induces all the 784 world's largest firms recorded by the University of Reading
database, base year 1984.
Source: Cantwell and Jarme (2000).
96 KEITH PAVITT

role in innovation is likely to be low where large firms are selling to a mass market
of users who lack strong technological capabilities (e.g. many consumer products).
• Innovation processes will differ greatly between large and small firms in other
respects. Innovations in large firms involve a larger number of people in special­
ized functions, with shifting responsibilities over time. Innovation processes in
large firms are also more likely to involve recognizable procedures, whether formal
or informal. In small firms, there are fewer resources to apply to such issues.
Decisions related to the recognition of opportunities, the allocation of resources,
and the coordination of functional activities, are more likely to reflect the compe­
tencies and behavior of senior managers.16
The heterogeneity and contingent nature of innovation means that there can be no
simple “best practice” innovation model for firms or managers to follow. Each firm
proceeds on the basis of its prior experience and the technological trajectories
evident in the specific industry or product group. But the lack of global “best
practices” should not be taken to mean that innovation strategy does not matter,
nor that good management cannot make a difference to firms’ productivity, market
share, or profitability.

4.4 T r a n s f o r m a t i o n of Knowledge
into W orking A r t if a c t s

Scientific advances enable the creation o f artifacts o f increasing com plexity, em­
bodying an increasing num ber o f subsystem s and com pon en ts, and drawing on a
broadening range o f fields o f specialized know ledge. T h is increasing system com­
plexity is one consequence o f the grow ing specialization in know ledge production, it
has resulted in both better understanding o f cau se-effect relationships, and better
and cheaper m ethods o f experim entation. These advances have reduced the costs of
technological search, enabling greater com plexity in term s o f the num ber o f com­
ponents, parts or m olecules that can be successfully em bod ied in a new product or
service, Developm ents within IC T are accelerating this trend: digitalization opens
options for m ore com plex systems, and sim ulation techniques reduce the costs of
experim entation (Pavitt and Steinm ueller 2001),
M anagers involved in transform ing S & T know ledge into products, systems, and
services need to be aware o f tour sets o f specific trends in their industries. These are:
(a) technology trajectories and scientific theories; (b) w here relevant, government-
funded R & D program s; (c) systems integration; and (d) techniques and approaches
to m anaging uncertainty. I deal with each o f these issues below.
 INNOVATION PROCESS 97

4.4.1 Keeping Technological Practice (not too far)

ahead of Scientific Theory
In spite o f the sp ectacular increases in scientific know ledge over the past 200 years,
theory rem ains an insufficient guide to tech nological practice, T h is is p artly because
o f the increasing co m p lexity o f physical artifacts and the know ledge bases that
u nderpin them . T h e im p o rtan ce o f practice is reflected in the co n tin u in g d om in an ce
in industrial R8cD lab orato ries o f d evelopm ent acti vities— the design, b u ild in g and
testing o f speci fic art ifacts— co m p ared to research in fields on w hich they are based.
C onstant (2000) describes this as tech n o logy ad van cin g through the recursive
practices o f scientists and engineers, in volvin g “ alternate phases o f selection and o f
co rrob o ration b y u s e .. . . T h e result is stron gly co rrob o rated fo u n d atio n al kn ow -
ledge: know ledge that is im plicated in an im m ense n u m b er and variety o f designs
em bodied in an even larger p o p u latio n o f devices, artifacts, and practices, that is
used recursively to p ro d u ce new know ledge"5 (p. 221).
C o n tin u o u s in n o va tio n requires constant im provem en t in m ethods o f tech n o ­
logical search, b u t technical co m p lexity cannot ru n too far ahead o f scientific
u n d erstan d in g,17 T h e feedbacks in both direction s— between im provem en ts in
scientific un d erstan d in g and im pro vem en ts in technical p erform an ce— have been
well d ocu m en ted b y h istorian s and others in areas such as aerodyn am ics and
th erm o d yn am ics.18 M ah d i (2002) has recently developed a tax o n o m y o f tech n o ­
logical search that d ep en d s on three factors: (1) the degree to w h ich technological
problem s can be d eco m p o sed into sim p ler sub-tasks; (2) the level o f und erstand in g
of cause-effect relations; and (3) the costs o f exp erim en tation w ith possible
solutions.
A d van ces in the tech n ologies o f m easu rem en t and m an ip u latio n o f the increas­
ingly sm all, are a m a jo r sou rce o f im pro vem en ts in tech nological search. T h is has
been the case in the past few decades in m o lecu lar b io lo g y and m aterials, b oth o f
w hich have open ed m a jo r new o p p o rtu n ities fo r technical ch an ge.19 IC T can also
reduce the costs o f search and selection. M a jo r advances in large-scale co m p u tin g
and sim u latio n tech n o lo g y have reduced the costs o f exp lo rin g alternative technical
co n figu ration s, an d have created o p p o rtu n ities fo r in creasin gly co m p lex system s
m ade p ossib le th ro u g h the d igitalization o f data o f all sorts (Pavitt and Steinm ueller
20 0 1), In n o v a tio n m an agers an d engineers involved in tran sfo rm in g know ledge into
w o rk in g artifacts to d a y need to be aw are o f specialized IC T trends in their ow n
as well as new measurement and manipulation techniques elsewhere that
industries,
themselves frequently involve the application of advanced ICTs.
Nightingale (2000) has shown that these mechanisms have radically changed
experimental techniques in the pharmaceutical industry during the past ten years.
There has-been-a shift towards more fundamental science, for example, linking
ЬіпгЬ^тігяІ■■тргЬяпіятч to the expression of genes; and there has been a much
greater use of simulations (involving models, extensive data banks, etc,) to conduct
98 KEITH PAVITT

virtual experiments complementary to real ones. A third related development has

been the use of high throughput screening techniques.20

44*2 Government Funded Programs

Technological activities directly financed by governments have sometimes been of
major importance in opening and exploiting innovative opportunities. Successes
include ICT in the USA, where military-related programs played a major role in the
early development of computers, semiconductors, software, and the Internet. Mili­
tary programs have also had important technological sp in -o ffs into civil aviation in
the USA, while governments in Japan and France have successfully supported the
development of high-speed trains. But there have also been many disappointments.
Policies to support the development o f civilian nuclear power have on the whole
not been successful, nor have those tor the support o f high-volume, low-cost
residential construction (See Eads and Nelson 1971, as well as the case studies in
Nelson 1982), M ore recently, policies to encourage the development of renewable
energy technologies have met with mixed success. And controversy surrounds the
achievements o f a whole series o f EU programs.
Since the 1980s government programs for “ p rc-co m p etitive” collaborative R&D
in Europe (e.g. E SP R IT an d Eureka), the USA (e.g. Sem atech) and Japan (e.g. the 5G
IC O T Program) have proliferated. Thus, most major firms are presented with
opportunities for participation in such programs, Firms require methods for evalu­
ating their potential contribution to corporate goals, the financial and organiza­
tional costs ol participating, the risks involved in not participating, and the ways in
which government programs can complement or fit into the overall corporate
strategy (Floyd 1997).
It is difficult to generalize from the recent history o f government support tor
industrial innovation. All such programs involve technical lobbies successfully
putting pressure on governments for financial support, often in fields related closely
to m ilitary applications, or (often large-scale) infrastructure, such as transport,
energy, housing, and communications. This process can lead to neglect ol commer­
cial constraints and to premature commitments to particular designs. Economists
highlight the opportu nity costs ol these programs; but government support can also
speed up ci itical technological learning at a time when purely private markets are not
ready to take the risks. I he early development o f ICT in the USA suggests the
im poi tanee ot diveisity and experimentation in government support for techno­
logical pmgtess. But would this have worked for the development o f high-speed
trains, where the exists ot experiments are much higher and technical change is more
increm ental' And, as we shall see in the next section, everyone makes mistaken
assumptions about future developments in a complex and last-changing world.
 INNOVATION PROCESS 99

4.4.3 Multi-Technology Firms, Modularization

and Systems Integration
In ad d ition to in creasin gly co m p lex artifacts, specialization in know ledge p ro d u c ­
tion has increased the range o f fields o f know ledge that contribute to the design o f
each product. C o m p are the origin al m echanical lo om w ith the m an y fields o f
specialized know ledge— electrical, aero d yn am ic, softw are, m aterials— that are
now em bod ied in the co n tem p o ra ry design. O r observe the w ay in w hich m odern
autom obiles in creasin gly integrate plastic and oth er new m aterials, as well as
electronic and softw are con trol system s.
Firm s d esign in g these in creasin gly co m p lex produ cts find it difficult to m aster
advances in all the fields that they em body. H ence the grow in g im po rtan ce o f
m o d u lar pro d u ct architectures, w here co m po n en t interfaces are standardized,
and interdependen cies am o n gst co m pon en ts are decoupled. T his enables the o u t­
sourcing o f design and p ro d u ctio n o f com pon en ts and subsystem s, w ith in the
constraints o f overall p ro d u ct (o r system ) architecture.
Technological convergence has also p ro vid ed o p p ortu n ities for fu rth er vertical
disintegration betw een p ro d u ct design and m anu factu re. Sturgeon (1999), fo r
exam ple, describes recent grow th in contract m an u factu rin g in electronics, in
which specialized firm s take over pro d u ct design fro m other firm s, and assum e
resp on sib ility fo r detailed engineerin g and m anufacture. T h e technological conver­
gence here is based o n in creasin g au to m atio n o f rou tin e operatio n s (e.g. co m pon en t
in sertion ), and on g ro w in g use o f standard softw are tools.
C o n tract m an u factu rin g is gro w in g in o th er industries as w ell,21 givin g rise to
“ m o d u lar p ro d u ctio n netw orks.” These m o d u les are defined b y distinct breaks in the
value chain at p o in ts w h ere in fo rm atio n regarding p ro d u ct specifications can be
h igh ly form alized. T h is occu rs w ith in fu n ctio n ally specialized valu e chain nodes,
w here activities tend to be h ig h ly integrated and based on tacit linkages. Betw een
these n o d es linkages d ep en d o n the tran sfer o f codified in fo rm atio n .
A t first sight, these recent changes m igh t app ear to p o in t to a n eatly specialized
system fo r the p ro d u c tio n o f in n o vation s, w ith p ro d u ct and system s designers, their
su bcontractors fo r co m p o n en ts and subsystem s, and their m an u factu rers, w o rk in g
together th ro u gh arm s-len gth m arket relations— a trend foreseen b y Sturgeon
(1999). B u t this neglects the consequences o f im p o rtan t d istin ction s that need to
be m ad e betw een th e p ro p erties o f artifacts, the know ledge o n w h ich they are based,
and the degree to w h ich such k n o w led ge can be tran sfo rm ed into codified in fo r­
m atio n (G ran stran d et al. 1997).
B rie fly stated, the development and production o f increasingly complex artifacts
are, as w e have seen, based on the integration o f an increasing number o f fields o f
specialized kn o w led ge. T h ese fields advance at different sp eed s and their progress
cannot be traefadm leiy b y monitoring codified information, The division o f labour
Ь е Ш т і - т т р ш т г т production thu s cannot be mirrored by an equivalent
100 KEI TH PAVI TT

division o f labour in knowledge (B rusoni et a l 20 0 1). S o m e overlap between

com panies in knowledge com petencies is necessary to deal w ith the transfer of
tacit knowledge, to m anage unforeseen consequences o f system ic complexity, and
to resolve im balances between com ponents that are liable to result from their uneven
rates o f technical change. Sim ilarly arm 's-length relations betw een firm s will not be
as effective as form s o f "loose co u p lin g" w ith p erio d ic b o u ts o f integration, when
systems architectures and the tasks o f com ponent su ppliers are redefined by firms
specializing increasingly in systems design and integration.

4.4.4 “ Managing5’ Uncertainty?

Specialized R & D and related activities in business firm s have becom e institutional­
ized and predictable sources o f discoveries, inventions, in n o vation s, and improve­
ments. However, the process o f innovation is com plex, in volvin g m any variables
whose properties and interactions (and econom ic usefulness) are understood im­
perfectly. As a consequence, firm s are not able to explain fu lly and predict accurately
either the technical perform ance o f m ajor in n o vation s, o r their acceptability to
potential users (or in som e cases even w ho the poten tial users are). They cannot
accurately predict the technical and com m ercial o u tcom es o f their own innovative
activities, nor those o f other firm s.22 On average, research scientists and engineers
tend to be over-optim istic about the costs, benefits, and tim e periods o f their
proposed projects, and about market dem and for the p ro d u cts resulting from
them. But there is typically great variation in the ratio o f ex post outcom es to ex
ante estimates o f investment or profit w ithin any co rp orate p o rtfo lio o f projects
(Freem an 1982; M ansfield 1995), Indeed, com m ercially unsuccessful projects often
account for a disproportionate share of corporate R & D sp en d in g (Griliches 1990).
Business firm s (and others) rarely are capable o f defin ing the full array of possible
uses that m ay emerge for their innovations, especially radical ones. Examples of
inaccurate predictions about what turned out later to be spectacularly successful
technologies and innovations are legion. Early tw en tieth-cen tu ry pioneers of radio
com m unication conceived it as a system ot p o in t-to -p o in t com m unications, par­
ticularly between naval vessels— it was o nly m uch later that the m uch larger market
for mass radio com m unications was recognized. A fter W orld W ar II, the founder of
IB M foiesaw a world market for com puters in single figures. For the m ore recent
period between the 1960s and the 1980s, Schnaars and Berenson (1986) concluded
that only about h alf the m ajor new product fam ilies annou nced in the U SA turned
out to bt com m ercially successful. The recent experience o f inaccurate forecasts of
the potential markets for various generations o f the m obile phone, and various
functions associated with these (the unexpected success o f text messaging, for
instance) is equally instructive.
 INNOVATION PROCESS 101

C o rp o rate m anagers therefore face severe difficulties in deciding h o w to deal w ith

innovative activities, w h ich have som e o f the elem ents o f conventional investm ent
activities, but fo r w h ich severe u n certain ty m eans that co n tin u ou s feedback fro m the
m arket, past experience and exp erim en tation are essential. In practice, to p -d o w n
corporate visio n s can be a p o o r guide to in n o vatio n strategies. E ricsson's success in
opening up m obile telephony began w ith initiatives from m iddle-level technical
m anagem ent, rather than fro m the top. In the academ ic and business literatures, the
failures o f to p -d o w n visio n s are easily forgotten and the successes oversim plified.
For exam ple, P rah alad and H am el (1990) tell the sto ry o f C an on 's successful d iver­
sification fro m optics and p recisio n m echanics into electronics technology, and
from cam eras into p h o to co p yin g and co m p u ter periph eral produ cts; but they
ignore the firm 's failed d iversificatio n into record ing produ cts and electronic calcu ­
lators (San doz 1997).
The b road differences betw een search and selection activities have been recog­
nized fo r a lo n g tim e— in practice, w ith the d istin ction betw een co rp orate and
d ivisio n al R & D activities; and in th e o ry w ith the distinction betw een “ know ledge
b u ild in g” and “ strategic p o sitio n in g ” on the one hand, and “ business investm ent”
on the other (M itch ell and H am ilto n , 1988). H ow ever, as the recent h isto ry o f
corporate R & D show s, m ain tain in g balan ce and linkages betw een the tw o is not
an easy task. B rie fly stated, there is no one best w a y o f evaluating the costs and
benefits o f co rp o rate R M ) expen ditu res ex ante . R ule-based system s fail, because
they in evitab ly sim p lify, and m a y therefore neglect w hat tu rn out to be im p o rtan t
factors in a co m p lex system . Jud gem en t-b ased system s fail, because o f the im p o ssi­
b ility o f q u ick ly d istin gu ish in g go o d ju d g em en t fro m go od luck. O ne consequence is
periodic sw in gs in fash ion s an d m anagem en t practices. These often reflect struggles
for influence betw een fin an cially trained m anagers, w ho tend to prefer rule-based
system s, an d those w h o are tech n ically trained and prefer to rely on technical
judgem ents.

4.5 M a t c h in g of A r t i f a c t s ,

O r g a n iz a t io n a l P r a c t ic e s a n d

M a r k e t D e m a n d

T h e matching o f p ro d u cts, processes, systems, an d services (and organizational

practices) with actual an d p o ten tial market demand is a m a jo r responsibility fo r
managers o verseein g innovation in the successful co rp o ratio n . T h e co rp o ratio n
102 KEITH PAVITT

builds on its accum ulated knowledge o f prod u ct and process technologies, of

organizational practices, and o f users needs to ca rry o u t this function. Responding
to (and creating) market needs and dem ands, as w ell as m atching organizational
practices with technological opportunities, involves dealing w ith disruptive change.
This disruptive change interacts w ith one o f the negative consequences o f special­
ization: namely, the potential for tribal w arfare over the old and the new between
specialized functions and disciplines w ithin the firm .

4*5,1 Matching Technology and Organizational Practices

with Market Needs
Chandler (1977) has show n that the rise in the U S A at the end o f the nineteenth
century o f the large, m ulti-unit firm , and o f the co o rd in atin g function o f profes­
sional m iddle managers, depended critically on the d evelopm ent o f the railroads,
coal, the telegraph, and continuous flow produ ction . Sim ilarly, the later develop­
ment o f the m ultidivisional firm reflected in part the m ajo r opportunities for
product diversification in the chem ical indu stry opened up b y breakthroughs in
synthetic organic chemistry, A lthough it is co m m on p lace tod ay to argue that
technologies and organizational practices do, o r shou ld, co-evolve with market
dem ands, there is som e risk o P tec h n o lo g ic a l d eterm in ism ’' in C h andler s argument
that this process largely involves the adaptation o f corp orate organizational practices
to em erging market needs and technological opportu nities.
Technical advances often precede organizational and m arket advances, among
other reasons because o f the firm er kno wledge base and low er costs o f experimen­
tation associated with technical, as opposed to m arket or organizational, innov­
ation. This does not m ean that technology im poses one organizational “ best way" or
even a clear strategy towards the m arketplace. V ariety in the characteristics of
technologies, their continuous change and uncertain app lications also produces
variety and experim entation in organizational and m arketing practices. But this
variety and change does not mean that “ anything goes" in either organizational or
m arketing terms. It m ay be practically im possible for a firm that wishes to remain
com petitive to resist m aking use o f new technologies and know ledge in its own
future product and process developm ent— unless it w ishes to becom e a niche
producer like the m anufacturers o f clockw ork watches and analogu e record players.
But m oving into new technologies w ithout app ropriate changes in terms o f skills
and training, divisions o f labour and interrelations between parts o f the organiza­
tion, can be even m ore costly. O r consider the firm s investm ent decisions. A firm
applying conventional cost-benefit analysis and strict cost controls to all o f these
decisions will not prosper in the long term in a com petitive m arket governed by
 INNOVATION PROCESS 103

the exp loitatio n o f a rich, varied and rap id ly advan cin g b o d y o f technological
know ledge.23
The em p irical literatu re sum m arized in the first tw o colu m n s in Table 4.2 high-
lights the organ izatio n al and m arketin g practices that m ust be m ade consistent w ith
key features o f technologies:

* External linkages w ith potential custom ers, and w ith the im po rtan t sources o f
know ledge an d skills.
* Internal linkages in the key functional interfaces for experim entation and learning.
* The centralization o f resource allocation and m o n ito rin g activities needs to be
con son an t w ith the costs o f technological and m arket experim entation .

Table 4.2 Matching corporate technology and organizational practices with

market needs and demands

Corporate technology Matching organizational and Dangers in radical technological

—► marketing practices change

Inherent characteristics
1. Richness of la. Allocating resources for
opportunities exploring options
lb. Matching technologies
with product markets
2. Costs of specific 2. Degree of centralization
experiments in decision making
1a. Greater opportunities not
matched by resources for
exploring options
1b. Matching opportunities missed
in the marketplace
2. Reduced cost of experiments
not matched by
decentralization or market
testing

Supporting skills and networks

1. Specific sources of 1. Participation in specific 1. Difficulties in recognizing 8t
external knowledge professional knowledge joining new knowledge
networks networks
2. Accumulated 2. Learning and improving in 2a. Difficulties in recognizing £t
knowledge d f specific key functions and across responding to new customers'
customers' demands* key functional interfaces demands, distribution
distribution channels* channels, production methods,
production methods, * supply chains
supply chains. 2b. Difficulties in recognizing new
key ftmctional interfeces
3. Anticipating and 3. Scepticism and resistance from
V V\ 'Л, .J 1
organizing for the 1 established or potentially
obsolescent professional and
: . . groups ;
104 KEITH PAVITT

* Criteria for resource allocation need to be con son an t w ith levels o f technological
and market opportunity*
* Alignm ent o f professional groups, w ho possess po w er an d control, with fields of
future opportunity.
The richness o f the technological and m arket o p p o rtu n ities and the scale of
technical experim ents should determ ine the app ro priate share o f resources allocated
to technological search, as well as the degree o f centralization and fluidity in
organization structures. Supporting skills and netw orks will define the specific
com petencies to be accum ulated, professional netw orks to be joined and key
functions and functional interfaces w ithin and across w h ich learning must take
place within the firm.
The particular circum stances o f the in d ivid u al firm and project will obviously
define the basic skills required for com m ercial in n o vation . But the discussion in this
chapter leads to a clear-cut conclusion: in add ition to specialist skills, “ gatekeeper”
skills and general com m unication skills n ow are m ore im p o rtan t alm ost everywhere.
People who are capable o f com m unicating across o rgan ization al barriers, disciplin­
ary barriers, and professional barriers can be invaluable. In very sm all firms it maybe
satisfactory for one or two key individuals to possess the unique combination of
required skills. In larger firm s, the specific requirem ents m ay be hard to anticipate.
There is no single m anagerial planning prescription.
Differences am ongst technologies are reflected in differences in organizational
and m arketing practices. For exam ple, both ph arm aceutical and consum er electron­
ics firm s see rich technological and m arket o p p ortu n ities, and thus devote sub­
stantial resources to technological search. The m uch h igher costs o f experimentation
in pharm aceuticals mean that drug firm s tend to have centralized and formal
procedures for launching new products, whilst in con su m er electronics the situation
is m ore likely to be decentralized and inform al. Sim ilarly, as already mentioned,
both pharm aceutical and autom obile com panies have centralized decision struc­
tures, but the form er stress interfaces between corporate R & D and public research
in biom edical fields, whilst the latter concentrate on links between R8cD and
production.

4.5*2 Coping with Radical Change

The past 200 years have seen periodic step-jum ps in technological understanding
and perform ance in specific fields. D uring the past several decades, these discontinu­
ous advances have been m ore often than not based on m ajo r scientific break­
throughs. “ Radical” innovations have reduced considerably the costs o f key
econom ic inputs, and have therefore been w idely adopted and becom e the catalysts
for m ajor structural changes in the econom y. T h ey include steam power, electricity
 I NNOVATI ON PROCESS 105

m otorization, synthetic m aterials, and radio co m m u n icatio n s (Freem an and Lou^a

2001). The m ost celebrated co n tem p o rary exam ple, o f course, is the m assive and
con tin uin g reductions in the costs o f storing, m an ipu latin g, and tran sm itting in fo r­
m ation b rou gh t ab ou t b y im provem en ts in ICT.
Each w ave o f rad ically new technologies has been associated w ith the grow th o f
firm s that have m astered the new technologies and have pion eered in the d evelo p ­
m ent and co m m ercializatio n o f related products, processes, and services. In the
current jargo n o f co rp o rate strategy, these firm s have developed core com petencies
in the new technologies, w h ich have becom e a distinctive and sustainable co m p eti­
tive advantage.
Ever since Sch u m p eter associated the advent o f revo lu tio n ary technologies w ith
“ waves o f creative destru ction,” there has been debate about the relative role o f
incum bent large firm s and new entrants in exp loitin g them . O ver the past tw enty
years, m ost o f the an alytical w ritin g has been stacked against incum bents, although
recent em p irical studies p o in t to evidence in favo u r o f both (M ethe et a l 1996). O ver
tim e, the w eight o f the argu m en t has shifted som ew h at aw ay fro m em phasis on the
difficulties facing in cu m b en ts in m astering new fields o f tech nological know ledge
(C o o p er and Schendel 1976; Tushm an and A n d erson 1986; U tterback 1994).
M ore recent w o rk has em ph asized the difficulties faced b y incu m bent firm s that
m ust adapt established o rgan izatio n al practices to seize the o p p o rtu n ities opened b y
revo lu tio n ary tech n o logical changes. E xam ples include the o rgan ization al con se­
quences o f changes in p ro d u ct architectures (H en d erson and C lark 1990), resistance
from gro u p s w ith established com petencies (L eo n ard -B arto n 1995; Tripsas and
G avetti 2000), an d the u n expected em ergence o f new m arkets (C hristensen 1997;
Levinthal 1998).
C o n tra ry to a w id esp read assu m p tio n , the nature and d irection s o f rad ical new
technological o p p o rtu n ities are easily recognized b y the technically qualified; for
exam ple, m in iatu rizatio n , com pression , and digitalization are key trajectories in
ICT. T h e result in this case is that a g ro w in g n u m b er o f large firm s, in a grow in g
n um ber o f in du stries, are n o w tech nically active in IC T (G ran stran d et al. 1997;
M endon^a 2000). H ow ever, the difficult, costly an d u n certain task is that o f co m ­
b in in g rad ically n ew tech nical com petencies w ith existin g technical com petencies
and o rgan izatio n al practices, m an y o f w h ich m ay be threatened o r m u st be changed
in o rd er to e xp lo it p o ten tial m arket o p p ortu n ities. E xp erim en tatio n and d iversity
therefore are n ecessary fo r exp lo ratio n o f the directions and im p licatio n s o f radical
tech nological changes, but also fo r assessing the im p licatio n s o f these changes fo r
products, m arkets an d o rgan izatio n al practices.
T h e th ird c o lu m n o f Table 4.2 identifies so m e reasons w h y such exp erim en ts m ay
fail in in cu m b en t firm s. S o m e are a consequence o f the need to m o d ify com petencies
o r o rgan izatio n al practices; and som e o f the inevitable u ncertainties in the early
stages o f rad ically n e w tech n ologies. T h e lik elih o o d that established firm s w ill fail
increases w ith the n u m b e r o f practices an d com petencies that need to be changed.
106 KEITH PAVITT

H ere a com parison between the conclusions o f tw o recent indu stry studies is
instructive, Klepper and Sim ons (2000) have show n that firm s already established
in m aking radios were subsequently the m ost successful in the new ly developing
colour T V market. O n the other hand, H o lb ro o k and his colleagues (2000) have
shown that none o f the firm s established in designing and m akin g therm ionic valves
was subsequently successful in sem iconductors.
W ith the benefit o f hindsight, we can see that success in sem iconductors required
m ore changes in technological com petencies, organizational practices, and market
experim entation am ongst incum bents, than d id success in co lo u r TV. The valve
firm s required new com petencies and netw orks in q u an tu m physics, a much
stronger interface between product design and v ery d em an d in g manufacturing
technology, and the ability to deal with new sorts o f custom ers (com puter makers
and the m ilitary, in addition to consum er electronics firm s). Fo r the radio firms, the
shift to color T V required basically the sam e technological com petencies, augmented
by w ell-know n screen technologies. O therw ise, the cu stom ers and distribution
channels rem ained unchanged, as did the key netw orks and linkages both inside
and outside the firm.
According to Chandler's (1997:76) so-called “ c o n tin u ity " thesis, the population of
incum bent large firm s has rem ained stable in recent tim es, because o f their accumu­
lated skills and resources in adopting new technologies and adapting to them. This
thesis has been challenged by Lou^a and Mendon<;a (1999) and b y Freeman and
Lou^a (2001: 340-55), who argue that, on the contrary, a cohort o f new large firms
continues to join the population o f incum bent large firm s w ith each new wave of
technical change. O nly a m in o rity o f the largest firm s was able to rem ain at the top
through several waves. This suggests that the m icro-level evidence considered in this
Section, especially the factors listed in the third (righ t-h an d) colum n o f Table 4.2,
have had significant consequences for the evolu tion o f in d u stry structure.24 Radical
technological change, and the success or failure o f incum bent firm s in adapting to it,
thus m ay have im portant consequences for structural change in the econom y as a
whole.
Firm s in the vanguard o f developing and exploiting rad ically new technologies
must be distinguished from the m ore num erous firm s w ho adopt and integrate the
new technologies with their current activities. F o r these firm s, in-house competen­
cies in the new technologies are background; in other w ords, necessary for the
effective adoption o f advances made outside the firm . Paradoxically, the very fact
that radically new technologies allow step-jum p reductions in the costs o f a key
input sim ultaneously makes their adoption a com petitive im perative and an un­
likely source for sustained com petitive advantage am o n g adoptin g firms. For
exam ple, in the past m any factories had no choice but to adopt coal and steam—
and later electricity— as a source o f power, given their cost and other advantages. The
same is true today tor m any ICT-based m anagem ent practices. But in neither case
were these revolutionary advances by them selves a source o f sustainable competitive
 I NNOVATI ON PROCESS 10J

advantage fo r the ad o p tin g firm s. M u ch o f the em phasis by w riters on corporate

strategy— like B arn ey (1991) an d Porter (1996)— on the im po rtan ce o f establishing a
distinctive and sustainable advantage accordin gly does not, and cannot, ap p ly to the
m ajor tran sfo rm atio n s n o w in evitab ly h ap pen in g in m any com panies through the
ad option o f ICT. T h eir fram ew o rk can help understand C IS C O (a m ajo r U S supplier
o f equipm en t for the Internet), but does not help m uch w ith T E S C O (a m ajo r U K
superm arket chain, in creasin gly using the Internet).

4.5.3 Tribal Warfare

In his en u m eration o f the poten tial advantages o f increasing specialization in
know ledge p ro d u ctio n , A d am Sm ith describes the vario u s scientific disciplines
as 'Tribes.” C ertain ly an im p o rtan t elem ent in co n tem p o rary processes o f in n o v ­
ation is "trib a l” con flict betw een different profession al groups w ith specialized
know ledge. Som e such conflicts have been tou ch ed u p on above: finan cial versus
technological com petencies in the evalu ation o f R & D pro gram s; technical versus
m arketing com petencies in pro d u ct developm ent. But perhaps the m ost im p o rtan t
is the poten tial resistance o f a co m p an y’s current top m anagers and technical staff
(rooted in the successes o f the past), to the in tro d u ction o f new specialized co m p e­
tencies and m ethods, reflecting poten tial o p p o rtu n ities fo r tom orrow .
The difficulties in in tro d u cin g the new, in the face o f the tried and tested old, w ere
spelled out lo n g ago:

It must be considered that there is nothing more difficult to carry out, nor more doubtful o f
success, nor more dangerous to handle, than to initiate a new order o f things. For the
reformer has enemies in all those who profit from the old order o f things, and only lukewarm
defenders in all those who would profit by the new order, this lukewarmness arising partly
from fear o f their adversaries. , . and partly from the incredulity o f mankind, who do not
truly believe in anything new until they have had actual experience o f it. Thus it arises that on
every opportunity for attacking the reformer, his opponents do so with the zeal o f partisans,
the others only defend him half-heartedly. (Machiavelli 1950: 21-2)

W ell-docum ented c o n tem p o ra ry exam ples o f this process include IB M ’s early re­
luctance to enter the perso n al co m p u ter m arket and P o laro id ’s early com m itm en t
but subsequent failu re to d evelop a business based on d igital im agin g (Tripsas and
G avetti 2000). In b o th cases, a co m p an y w ith the technical resources to develop the
new tech nology, failed to d o so, due to resistance and scepticism fro m the established
p o w er structures. In these cases, it can p lau sib ly be argu ed that yesterday s "co re
com petencies,” becam e to d a y ’s "c o re rigid ities” (L e o n a rd -B arto n 1995).
B u t the n ew does n o t alw ays tu rn o u t to be better than the old. C o n servative
resistances in o il co m p an ies to investm ents in n uclear po w er in the 1970s tu rn o u t to
h ave been largely ju stified . So, m o re recently, w as scepticism ab ou t the d o t.co m
108 KEITH PAVITT

boom . A nd in the light o f IB M ’s subsequent success in system s integration and

software, resistance to heavy com m itm ent to the P C co u ld yet be seen as beneficial in
the longer run. This is what m akes decision m aking ab o u t radical innovations so
difficult. The political battle for influence often involves one-sided and distorted
analyses, reflecting the interests o f specific disciplines and fu nctions. Crucial factors
and key uncertainties m ay be ignored, con sciou sly o r otherw ise. T h e successes and
failures only becom e clear well after the sm oke o f battle has cleared.
For today’s corporate m anager there can be no sim p le tools o r m odel to neutralize
the uneasy, politicized task o f dealing w ith radical in n o vation s. G o od judgement,
experience, trial and error learning rem ain the o n ly feasible “ too lk it” available to
today’s innovative corporations.

4.6 C o nclu sio ns

Despite spectacular im provem ents in the scientific know ledge base, and slower but
steady im provem ents in organizational know -how , in n o vation processes are neither
tidy, nor easy to delineate or m anage. Increasing specialization in the production of
artifacts and in knowledge has also increased levels o f co m p lexity— in artifacts
themselves, in the knowledge on w hich they are based, and in the organizational
form s and practices for their developm ent and com m ercial exploitation. As a
consequence— and contrary to som e o f the prediction s o f Schum peter (1962) and
Penrose (1959):

* innovations— especially radical innovations— rem ain unpredictable in their tech­

nical and com m ercial outcom es;
* technical entrepreneurship is not a general-purpose m anagem en t skill; at a time of
radical breakthroughs and new opp ortu n ities it is in large part specific to a
particular technological field and often to a particu lar place;
* m ajor innovation decisions are a largely political process, often involving profes­
sional groups advocating self-interested outcom es under conditions o f uncer­
tainty (i.e. ignorance), rather than balanced and careful estim ates o f costs, benefits
and measurable risk.

As a consequence, established large firm s have som etim es fo u n d it difficult to deal

with the radically new. In the future, they w ill co n fro n t new challenges. Increasing
com plexities in products, systems and the u nd erlying know ledge base are leading
firm s to experim ent with m odular product architectures and greater use o f ICTs and
the outsourcing o f com ponent design and produ ction . Large innovating firms are
therefore likely to becom e less self-sufficient in their processes, not m ore so.
 I NNOVATI ON PROCESS 109

Finally, increasing specialization in the p ro d u ction o f artifacts, and their u n d er­

lying know ledge bases, has m ade in n ovative processes increasingly path-dependent.
As a consequence, several aspects o f in n o vation processes are contingent on sector,
firm , and tech n o lo gy field. These include: the know ledge base u n d erlyin g innovative
o pp ortu n ities; the links betw een scientific th eo ry and technological practice; p o ssi­
bilities fo r know ledge-based diversification ; m ethods o f research budget allocation;
degree o f centralization; and the critical skills, interfaces and netw orks that need to
be developed. O n ly tw o in n o vatio n processes rem ain generic: co ordin atin g and
integrating specialized know ledge, and learning u n d er con dition s o f uncertainty.

N otes

1. Keith Pavitt had prepared several drafts o f this chapter, but passed away before completing
it. Amendments to the original have been made by Christopher Freeman, Mike Hobday,
Ian Miles, and the editors.
2 This chapter focuses mainly on large firms within the USA, Europe, and Japan. It does not
consider issues arising in connection with entrepreneurial attitudes and motivations, such
as their willingness to take risks (Drucker (1985) examines some o f the links between
entrepreneurship and innovation, and Roberts (1991) and Oakey (1995) examine the
process o f innovation within SMEs). For studies o f innovation processes within firms
based in the industrializing countries (or “ latecomer firms” ) see Hobday (1995), Kim
(1997), and Fagerberg and Godinho in this volume.
3. Arguably, “ innovation managers” (in practice, if not formally identified by job title) can
be found at all levels o f the firm (for an overview o f innovation management in firms see
Tidd et al. 2001). Hamel (2000) examines “ strategic innovation” (or innovation in
strategy approach) while Schonberger (1982) and Robinson (1991) deal with kaizen (or
continuous improvements) to current vintages o f capital equipment and organization.
4. For critical assessments o f firm-level models o f innovation, ranging from the linear model
to chain-link models and more recent interactive/contingent models, see Rothwell (1992),
Forrest (1991) and Mahdi (2002).
5. The term “ stages” is avoided here, as it implies linearity. Research has consistently shown
that the process o f innovation within the firm is anything but linear (Kline and Rosenberg
1986; Tidd et al. 2001; Van der Yen et al. 1999). The three sub-processes o f innovation,
although distinctive, overlap considerably and often occur concurrently.
6. For a review o f models o f innovation and the importance o f contingency, see Mahdi (2002:
ch. 2).
7. As noted earlier, most innovation processes are overlapping and intertwined, terms like
“ stages” or “ phases” impose an unrealistic linearity on the various innovation processes.
8. Recall that by artifacts we mean services and systems as well as more tangible material
artifacts such as items o f equipment.
9. The classic texts on this are Rosenberg (1974b Price (1984). and Mowery and Rosenberg
(1989b
ПО KEI TH PAVI TT ______________ ______ ______

10. See PInchot (1997) for a discussion o f the intrapreneur, that is, an innovator operating
within a large corporation.
11. See Georghion et al. (1986).
12. For a discussion of escalation using examples from the UK stock exchange and other
major ICT project failures, see Flowers (1996). In the public sector, or state-dependent
industries, the problem can be acute. The experience o f many countries that established
agencies to develop nuclear power, for instance, suggests that these are often extremely
difficult to restrain.
13. Three very different lines o f argument have been mobilized over recent years. First is the
view that the output of university research is a free good available to everybody—as
assumed by orthodox economics. In contrast, a second view argues that publicly funded
university research is a form o f conspicuous intellectual consumption reflecting techno­
logical and economic achievements, but not contributing to them (as suggested by Kealey
1996). A more nuanced argument is the “ mode 2” claim that the locus o f useful scientific
discovery is moving from universities to 4contexts o f application” (as persuasively
argued, if somewhat overstated, by Gibbons et al. 1994).
14. Universities also face management challenges here, o f course— not least in reconciling
the established structures of academic recognition with the contexts o f industrially
relevant research.
15. Georghiou et a l (1986).
16. Community Innovation Survey data (see the chapter by Smith in this volume for further
discussion) are now beginning to provide us with systematic data on many aspects of the
innovative effort and sources o f innovative ideas for firms o f different types and in
different sectors.
17. For example, in the absence of theory and/or cheap methods for constructing and testing
prototypes, the costs o f search and selection become prohibitively high— see Martin
(2000) on why Japanese swords did not improve over a period o f more than 500 years,
18. See e.g. Rosenberg and Nelson (1994) on the origins o f the engineering disciplines in US
universities.
19. A similar conclusion has been reached by Becker and Murphy (1992). They argue that the
degree of specialization in tasks is limited not by the extent o f the market, as in Smith's
famous formulation, but by the costs o f coordinating specialized activities. These
coordination costs are reduced by increases in general knowledge.
20. In addition, some prestigious academic institutions, such as Stanford University in the
USA, and the Ecole des Mines in Paris, are developing research programs in “ bio-inform­
atics, growing out of the challenges to the complex results o f the Human Genome
project,
21. Sturgeon lists apparel and footwear, toys, data processing, offshore oil drilling, home
furnishings and lighting, semiconductor fabrication, food processing, automotive parts,
brewing, enterprise networking, and pharmaceuticals. In addition, Prencipe (1997)
demonstrates that outsourcing has increased in the production o f aircraft engine com­
ponents.
22. Rosenberg (1994) has pointed out that in the nineteenth century, the Western Union
turned down an opportunity to purchase Bell's patent for the telephone, which it
regarded as an inferior product to the telegraph.
23. See e.g. the history of the UK General Electric Company under Arnold Weinstock (Aris
1998). ;
 I NNOVATI ON PROCESS 111

24. Note also, the first point in this column (іа). This relates also to the attitude and behavior
o f the financial system. Some recent work on financial capital has returned to the original
Schumpeterian emphasis on “credit creation” for the finance o f innovation at various
stages o f the successive technological revolutions. These factors which affect the growth,
composition and fluctuation o f demand and hence the influence o f demand upon
innovation at the firm level are further considered in the chapters by Verspagen and
O’Sullivan in this volume.

R eferen c es*

A r i s , S. (1998), Arnold Weinstock and the Making of GEC, London: Aurum Press.
B a r n e y , J. (1991), “ Firm Resources and Sustained Competitive Advantage” Journal of
Management 17: 99-120.
B e c k e r , G., and M u r p h y , K. (1992), “ The Division o f Labour, Co-ordination Costs, and
Knowledge,” The Quarterly Journal of Economics 107:1137-60.
* B r u s o n i , S., P r e n c i p e , A., and P a v i t t , K. (2001), “ Knowledge Specialization, Organiza­
tional Coupling and the Boundaries o f the Firm: Why Firms Know More Than They
Make” Administrative Science Quarterly 46 (4): 597-621.
C h a n d l e r , A. (1977), The Visible Hand: The Managerial Revolution in American Business,
Cambridge, Mass.: Belknap.
------ (1997), “ The United States: Engines o f Economic Growth in the Capital Intensive and
Knowledge Intensive Industries,” in A. Chandler, F. Amatori, and T. Hikino (eds.), Big
Business and the Wealth ofNationsyCambridge: Cambridge University Press.
C h r is t e n s e n , C. (1997), The Innovator's DilemmayBoston: Harvard University Press.
C l a r k , K., and F u jim o t o , T. (1992), Product Development Performance, Boston: Harvard
Business School Press.
C o n s t a n t , E. (2000), “ Recursive Practice and the Evolution o f Technological Knowledge,”
in J. Ziman (ed.), Technological Innovation as an Evolutionary Process, Cambridge: Cam-
bridge University Press.
C ooper, A., and Sch en d el , D. (1976), “ Strategic Responses to Technological Threats,”
Business Horizons (Feb.), 61-9.
D a v ie s , A., T a n g , P., Brady , T„ H obday , M., Rush , H „ and G a n n , D. (2001), Integrated
Solutions: The N ew Economy between M anufacturing and Services, Brighton: SPRU.
D rucker , P. F. (1985), Innovation and Entrepreneurship, New York: Harper & Row.
E ads , G „ and N elson , R. R. (1971), “ Government Support o f Advanced Civilian Technolo­
gies: Power Reactors and the Supersonic Transport,” Public Policy (Summer): 405-28.
Fagerberg , J. (2003), “ Schumpeter and the Revival of Evolutionary Economics: an Ap­
praisal of the Literature,” Journal o f Evolutionary Economics 13:125-59.
F loyd , C. (1997), M anaging Technology fo r Corporate Success, Hampshire: Gower.
F lowers , S. (1996), Software Failure: Management Failure: Amazing Stories and Cautionary
Tales, Chichester: Wiley.

* Asterisked items are suggestions for further reading.

112 KEITH PAVITT

FmREST, J. E. h m % "Models of the Process of Technological ЫвтшйтГ ТвЛяЛ/у

Analysis and Strategic Management з(4 ): 439"~52-
♦ Preemam, C (1982)1 The Economics of Industrial Innovation, London: France Pinter*
____m £ Lou^Ai E (2001)1 As Time Goes By: Emm the Industrial Revolutions to the Infor­
mation Revolution, Oxford: Oxford University Press*
G eokghiou, L , M etcalfe l 5., G ibbons M., Ra y , X , and Evans , l {1986), Рой-Іпптоікт
Performance, Basingstoke and New York: MacMillan*
G ibbons, M,, Limoges, C,, N owotny , H., Sch w artzm an , S., S c o t t , P., and T row, M.
(1994), The New Production of Knowledge, the Dynamics of Science and Research in Con­
temporary Societies, London: Sage,
♦ G r a n s t r a n d , a , P a t e l , E, and Р а ѵ іт т , К. (1997), "Multi-technology Corporations: Why
They Have 'Distributed* rather than 'Distinctive Core* Competencies,” California Man­
agement Review 39:8-25,
G riliches , Z. (1990), "Patent Statistics as Economic Indicators,” Journal o f Economic
Literature 28:1661-707.
Hamel , G. (2000), Leading the Revolution, Boston: Harvard Business School Press,
------and Prahalad , С. K, (1994), Competing fo r the Future, Boston: Harvard Business
School Press.
♦ H enderson , R „ and C lark , K. (1990), "Architectural Innovation: the Reconfiguration of
Existing Product Technologies and the Failure o f Established Firms,” Administrative
Sciences Quarterly, 35: 9-30,
♦ H icks, D. (199$), "Published Papers, Tacit Competencies and Corporate Management of the
Public/private Character of Knowledge,” Industrial and Corporate Change 4 : 401-24
H obday, M, (1995), Innovation in East Asia: the Challenge to Japan, Aldershot: Edward
Elgar.
♦ H olbrook, D., C ohen , W , Hounshell , D., and K lepper , S. (2000), "The Nature,
Sources, and Consequences of Firm Differences in the Early History o f the Semiconductor
Industry” Strategic Management Journal 21:1017-41.
Kealey , T. (1996), The Economic Laws o f Scientific Research, London: MacMillan.
K im , L. (1997), Imitation to Innovation: the Dynamics o f Korea*s Technological Learning,
Boston: Harvard Business School Press.
K lepper , S., and Simons, K* (2000), "Dominance by Birthright: Entry o f Prior Radio
Producers and Competitive Ramifications in the U.S. Television Receiver Industry,”
Strategic Management Journal 21:997-1016.
♦ K line , S. J., and Rosenberg , N. (1986), “An Overview of Innovation” in R. Landau and N.
Rosenberg feds.), The Positive Sum Strategy: Harnessing Technology fo r Economic Growth,
Washington, DC: National Academy Press, 275-304.
L eonard -Barton , D. (1995), Wellsprings o f Knowledge, Boston: Harvard Business School
Press.
Levinth al , D. (1998), The Slow Pace o f Rapid Technological Change: Gradualism and
Punctuation in Technological Change,” Industrial and Corporation Change 7: 217-47*
Lou<;a, E, and МвнвомдА, S. (1999), "Steady Change: The 200 Largest US Manufacturing
Firms in the Twentieth Century” Working Paper No. 14/99, CISEP-ISEG, UTL, Lisboa
L undvall , B. (1988), "Innovation as an Interactive Process: from User-Producer Inter­
action to the National System of Innovation ” in G. Dosi, C. Freeman, R. Nelson,
G. Silverberg, and L. Soete (eds.), Technical Change and Economic Theory, London: Frances
Pinter.
 INNOVATION PROCESS Щ

M a c h ia v e l l i , N. (1950 edn.), The Prince, New York; Modern Library College Editions.
M ah d i , S. (2002), “ Search Strategy in Product Innovation Process: Theory and Evidence
from the Evolution o f Agrochemical Lead Discovery Process/' D.PhiL Thesis, Unpub­
lished, SPRU, University o f Sussex,
M a n s f ie l d , E. (1995)? Innovation, Technology and the Economy, Aldershot; Edward Elgar.
M a r t i n , G. (2000), “ Stasis in Complex Artefacts," in J. Ziman (ed.), Technological Innovation
as an Evolutionary Process, Cambridge; Cambridge University Press.
M a r t in , B., and S a l t e r , A. (1996), “ The Relationship between Publicly Funded Basic
Research and Economic Performance: a SPRU Review," Report for HM Treasury, SPRU,
University o f Sussex,
M endon ^A, S, (2000), The ICT Component of Technological Diversification, M.Sc. disserta­
tion, SPRU, University o f Sussex.
M eth e , D., Sw a m in a t h a n , A., and M it c h e l l , W. (1996), “ The Underemphasized Role o f
Established Firms as Sources o f Major Innovations," Industrial and Corporate Change 5:
ll8l-203.
*M it c h e ll , G., and H a m ilto n , W, (1988), “ Managing R&D as a Strategic Option," Research
Technology Management 31:15-24 .
M ow ery , D. (1995), “ The Boundaries o f the U.S. Firm in R&D," in N, R. Lamoreaux and
D. M. G. Raff (eds.), Coordination and Information: Historical Perspectives on the Organiza­
tion of Enterprise, Chicago: University o f Chicago Press o f NBER.
*—— and R o se n b e r g , N. (1979), “ The Influence o f Market Demand upon Innovation: a
Critical Review o f Some Recent Empirical Studies," Research Policy 8:102-53.
------------- (1989), Technology and the Pursuit of Economic Growth, Cambridge: Cambridge
University Press.
N a r in , F., H a m ilt o n , K., and O l iv a st r o , D. (1997), “ The Increasing Linkage Between U.S.
Technology and Public Science," Research Policy 26:317-30.
N elso n , R. R. (ed.) (1982), Government and Technical Progress: A Cross-Industry Comparison,
New York: Pergamon.
N ig h t in g a l e , P. (2000), “ Economies o f Scale in Experimentation: Knowledge and Technol­
ogy in Pharmaceutical R 8 c D Industrial and Corporate Change 9: 3U-59*
O a k e y , R. (1995), High-Technology New Firms, London: Paul Chapman.
Pa v it t , K., and S t e in m u e l l e r , W. (2001), “ Technology in Corporate Strategy: Change,
Continuity and the Information Revolution," in A. Pettigrew, H, Thomas and R. Whit­
tington (eds,), Handbook of Strategy and Management, London: Sage.
P en ro se , E. T. (1959), The Theory of the Growth of the Firm, Oxford: Basil Blackwell.
Р ш сн о т, G. (1997), “ Innovation Through Intrapreneuring," ch. 26 in R. Katz (ed.), The
Human Side of Managing Technological Innovation, New York: Oxford University Press.
P o rter , M. (1996), “ What is Strategy?" Harvard Business Review, Nov./Dec,: 61-78.
P ra h a la d , C. K „ and H a m e l , G. (1990), “ The Core Competence o f the Corporation,"
Harvard Business Review, 90(3): 79-91.
P r e n c ip e , A. (1997), “ Technological Competencies and Product's Evolutionary Dynamics: a
Case Study from the Aero-engine Industry," Research Policy 25:1261-76.
P r ic e , D. d e S o lla (1984), “ The Science/Technology Relationship, the Graft o f Experi­
mental Science, and Policy for the Improvement o f High Technology Innovation,
Research Policy 13(1), 3-20.
Q u in n , J. B, (2000), “ Outsourcing Innovation; The New Engine o f Growth," Sloan Manage­
ment Review (Summer), 13-28.
114 KEITH PAVITT

R o b e r t s , E . (1991), Entrepreneurs in High Technology: Lessons from M IT and Beyond, Oxford:

O xford University Press.
R obinson , A. (1991), Continuous Improvement in Operations, Cambridge, Mass.: Productiv­
ity Press.
R o s e n b e r g , N . (i974)> “ Science, Invention, and Econom ic G row th, Econom ic Journal
84(333)» 90-108.
____ (1976), “ Technological Change in the M achine Tool Industry, 18 4 0 -19 10 ,” in Perspectives
on Technology, Cam bridge: Cam bridge U niversity Press.
____ (1994), Exploring the Black Box: Technology, Economics and History Cambridge:
Cam bridge University Press.
------and N elson , R. (1994), “American Universities and Technical Advance in Industry”
Research Policy 23:323-48.
R othw ell , R. (1992), “ Successful Industrial Innovation: C ritical Success Factors for the
1990s,” Research Policy 22(3), 221-39.
* ------F reem an , C , H orsley , A ., J e r v is , V., R o bertso n , A ., and T ow nsend , J. (1974),
“ SA PPH O Updated— Project SA PPH O Phase II,” Research Policy 3(3), 258-91.
S andoz , P. (199 7), Canon, London: Penguin.
Sa lter , A., D’ e ste , P„ M a r t in , B., G eu n a , A ., S co tt , А ., Р а ѵ іт т , K., P atel , P., and
N igh tin g ale , P. (2000), Talent Not Technology: Publicly Funded Research and Innovation
in the UK, CVCP, SPRU, and H EFCE.
S ch n aars , S., and B erenson , C. (1986), “ Grow th M arket Forecasting Revisited: A Look
Back at a Look Forward,” California Management Review 28: 71-88.
Sch onberger , R. (1982), Japanese Manufacturing Technique: Nine Hidden Lessons in Simpli­
city, New York: Free Press.
S ch u m peter , J. (1962), Capitalism Socialism and Democracy.; 3rd edn., N ew York: Harper
Torchbooks Edition (originally published 1942 by H arper and Brothers).
S m ith , A. (1937), An Inquiry into the Nature and Causes of the Wealth of Nations, New York:
M odern Library Edition.
Sturgeon , T. J. (1999), “ Turn-key Production N etworks: In d u stry O rganization, Economic
Developm ent, and the Globalization o f Electronics C on tract M anufacturing,” Unpub­
lished Ph.D, dissertation, University o f C alifornia at Berkeley.
*T idd , J., B essant , J., and Р а ѵ іт т , K. (2001), Managing Innovation: Integrating Techno­
logical, Market and Organizational Change, 2nd edn., Chichester: Wiley.
T r ipsa s , M ., and G a v e t t i , G. (2000), “ Capabilities, C ogn ition and Inertia: Evidence from
Digital Imaging,” Strategic Management Journal 2 1:114 7 - 6 1.
T ushm an , M . and A nderson , P. (1986), “ Technological D iscontinuities and Organizational
Environm ents, Administrative Science Quarterly, 31: 439—65.
U t terba ck , J. (1994), Mastering the Dynamics of Innovation, Boston, Mass.: Harvard
Business School Press,
V an de Y en , A., Polley , D. E., G arud , R., and V en t k a t a r a m a n , S. (1999), T h e Innovation
Journey, New York: O xford University Press.
 C H A P T E R 5

ORGANIZATIONAL
INNOVATION

ALICE LAM

5.1 In t r o d u c t io n

O r g a n iz a t io n a l creation is fu n d am en tal to the process o f in n o vation (Van de

Ven et al. 1999). In n o va tio n constitutes p art o f the system that p rodu ces it. T h e
system is itse lf “ o rgan izatio n ” o r “ organizin g,” to pu t it in W eick s (1979) term . The
ability o f an o rgan izatio n to in n ovate is a precon d itio n fo r the successful utilization
o f inventive resources an d n ew technologies. C onversely, the in tro d u ction o f new
tech n o logy o ften presents co m p lex o p p o rtu n ities and challenges fo r organizations,
leading to changes in m an agerial practices an d the em ergence o f n ew organizational
form s. O rgan izatio n al an d tech n ological in n o vation s are intertw ined. Sch um p eter
(1950) saw organ izatio n al changes, alon gside n ew pro d u cts an d processes, as w ell as
n ew m arkets as facto rs o f “ creative d estruction.”
In a general sense, the term “ organ izatio n al in n o vatio n ” refers to the creation o r
ad o p tio n o f an id ea o r b eh avio r new to the o rgan izatio n (D aft 1978; D a m a n p o u r and
Evan 1984; D a m a n p o u r 1996). T h e existin g literatu re o n organ izatio n al In n o vatio n is
indeed v e ry d iverse an d n o t w ell integrated into a coherent theoretical fram ew ork,
T h e p h en o m en o n o f “ o rgan izatio n al in n o vatio n ” is subject to different interpret-
ations w ith in the d ifferen t strands o f literature. T h e literatu re can be b ro ad ly
classified in to three d ifferen t stream s, each with a d ifferent focu s an d a set o f different
Иб ALICE LAM

questions which it addresses. O rganizational design theories focus predominantly

on the link between structural form s and the p ro p en sity o f an organization to
innovate (Burns and Stalker 1961; Lawrence and L o rsch 1967; M intzberg 1979). The
unit o f analysis is the organization and the m ain research aim is to identify the
structural characteristics o f an innovative organization, o r to determ ine the effects
o f organizational structural variables on pro d u ct and process innovation. This
strand o f literature has been m ost influential and well integrated into the literature
on technological innovation (e.g. Teece 1998). T h eories o f organizational cognition
and learning, by contrast, tend to focus on the m icro -level process o f h ow organiza­
tions develop new ideas for problem solving. T h ey em phasize the cognitive founda­
tions o f organizational innovation w hich is seen to relate to the learning and
organizational knowledge creation process (A gyris and S ch o n 1978; N onaka 1994;
N onaka and Takeuchi 1995). This cam p o f research pro vid es a micro-lens for
understanding the capacity o f organizations to create and exp loit new knowledge
necessary for innovative activities. A third strand o f research concerns organiza­
tional change and adaptation, and the processes u n d erlyin g the creation o f new
organizational form s. Its m ain focus is to understand w h eth er organizations can
overcom e inertia and adapt in the face o f radical en viron m en tal shifts and techno­
logical changes, and whether organizational change occu rs prin cip ally at the popu­
lation level through selection (e.g. H annan and Freem an 1977; 1984; Rom anelli and
Tushm an 1994). In this context, innovation is considered as a capacity to respond to
changes in the external environm ent, and to influence and shape it (Burglem an 1991;
Child 1997).
W hile there are im portant em pirical overlaps betw een these three different
strands o f research, they rem ain theoretically distinct. T h e separation o f these
research streams has prevented us from developing a clear vie w o f “ organizational
innovation,” and o f how its different dim ensions are in terrelated .1 T h is chapter seeks
to understand the interaction between organization and in n o vation from the three
different but interdependent perspectives. Section 5.2 exam ines the relationship
between organizational structures and innovation, d raw in g on the various strands
o f w ork in organizational design theories. Section 5.3 looks at organizational innov­
ation from the m icro-level perspective o f learning and organizational knowledge
creation. It argues that organizations w ith different structu ral fo rm s vary in their
patterns o f learning and knowledge creation, engendering d ifferent types o f innova­
tive capabilities. Section 5.4 discusses organizational ad ap tation and change, focus­
ing on whether and how organizations can overcom e inertia in the face of
discontinuous technological changes and radical shifts in environm ental condi­
tions. The chapter concludes b y discussing the lim itation s and gaps in the existing
literatures, and the areas for future research.
 ORGANIZATIONAL INNOVATION UJ

5.2 O r g a n iz a t io n a l S t r u c t u r e
and I n n o v a t i o n

C o n ven tio n al research on o rgan ization al innovativeness has explored the d eterm in ­
ants o f an o rgan izatio n 's p ro p en sity to innovate. A lth o u gh researchers have analysed
the influence o f in d ivid u al, organ ization al, and environ m en tal variables (K im b erley
and Evan isko 1981; B ald rid ge and B u rn h am 1975), m ost o f the research has focused
on o rgan ization al structu re (W olfe 1994)* W ith in the field o f o rgan ization al design
theories, there has been a lo n g trad ition o f investigating the links betw een e n viro n ­
m ent, structures, and o rgan ization al perform an ce. Several studies have show n h ow
certain o rgan izatio n al structures facilitate the creation o f new products and
processes, especially in relation to fast ch an gin g environ m en ts. T h e w o rk o f
m icro -eco n om ists in the field o f strategy also em phasizes the su p erio rity o f certain
organizational fo rm s w ith in p articu lar types o f business strategies and p ro d u ct
m arkets (Teece 1998). M o re recently, there has been a significant shift in the focus
o f theoretical e n q u iry aw ay fro m p u rely fo rm al structures tow ards a greater interest
in organ izatio n al processes, relationships, and b ou n d aries (Pettigrew and Fenton
2000). T h e gro w in g influence o f econ om ic so cio lo g y and the in tro d u ction o f
“ n etw o rk " concepts into the o rgan ization al design field denotes such a shift. The
relationship betw een n etw ork structu re and in n o vation is dealt w ith b y Pow ell and
G rod al (C h . 3 in this vo lu m e).

5.2.1 Contingency Theory: Context, Structure, and

Organizational Innovativeness
The classical theory o f organizational design was marked by a preoccupation with
universal forms and the idea o f “ one best way to organize.” The work o f Weber (1947)
on the bureaucracy and o f Chandler (1962) on the multidivisional form, was most
influential. The assumption o f “ one best w ay” was, however, challenged by research
carried out during the 1960s and 1970s under the rubric o f contingency theory w h ich
explains the diversity o f organizational forms and their variations w ith reference to
the demands o f context. Contingency theory argues that the most “ appropriate
structure” for an organization is the one that best fits a given operating contingency,
such as scale o f operation (Pugh et al. 1969; Blau 1970), technology (W oodward 1965;
Perrow 1970) or environment (Burns and Stalker 1961; Lawrence and Lorsch 1967).
This strand o f research and theory underpins our understanding o f the relationships
between the nature o f the task and technological environments, structure and
p erfo rm an ce. However, only some o f the studies deal specifically w ith the question
o f h o w stru ctu re is related to innovation.
Il8 A L I C E LAM

Burns and Stalker's (1961) polar typologies of “mechanistic” and “organic”

organizations (see Box 5.1) demonstrate how the differences in technological and
market environment) in terms of their rate of change and complexity, affect organ-

Box 5.1 B u m s and Stalker; m echanistic and organic structures

Burns and Stalker ( 1961) set out to explore whether differences in the technological
and market environments affect the structure and management processes in firms.
They investigated twenty manufacturing firms in depth, and classified environments
into “ stable and predictable” and “ unstable and unpredictable ” They found that firms
could be grouped into one of the two main types, mechanistic and organic forms, with
management practices and structures that Bums and Stalker considered to be logical
responses to environmental conditions.
The Mechanistic Organization has a more rigid structure and is typically found
where the environment is stable and predictable. Its characteristics are:
(a) tasks required by the organization are broken down into specialized, functionally
differentiated duties and individual tasks are pursued in an abstract way, that is
more or less distinct from the organization as a whole;
( b) the precise definition o f rights, obligations, and technical methods is attached to
roles, and these are translated into the responsibilities o f a functional position, and
there is a hierarchical structure o f control, authority, and communication;
(c) knowledge o f the whole organization is located exclusively at the top of the
hierarchy, with greater importance and prestige being attached to internal and
local knowledge, experience, and skill rather than that which is general to the whole
organization;
{d) a tendency for interactions between members o f the organization to be vertical, Le,
between superior and subordinate.
The Organic Organization has a much more fluid set o f arrangements and is an
appropriate form to changing environmental conditions which require emergent and
innovative responses. Its characteristics are:
(a) individuals contribute to the common task o f the organization and there is
continual adjustment and redefinition o f individual tasks through interaction
with others;
(b) the spread o f commitment to the organization beyond any technical definition, a
network structure o f control authority and communication, and the direction of
communication is lateral rather than vertical;
(c) knowledge may be located anywhere in the network, with this ad hoc location
becoming the center of authority and communication;
(d) importance and prestige attach to affiliations and expertise valid in industrial and
technical and commercial milieu external to the firm.
Mechanistic and organic form s are polar types at the opposite ends o f a continuum
and, m some organizations, a m ixture o f both types could be found.
Source. Bums and Stalker (1961).
 ORGANIZATIONAL INNOVATION 119

izational structu res and in n o vatio n m anagem ent. T h eir study found that firm s
could he grou p ed into one o f the tw o m ain types: the fo rm er m ore rigid and
hierarchical, suited to stable conditions; and the latter, a m ore flu id set o f arran ge­
m ents, ad ap tin g to co n d itio n s o f rapid change and innovation. N either type is
inherently righ t o r w ro n g , b u t the firm 's environ m en t is the contin gency that
prom pts a structu ral response. Related is the w o rk o f Law rence and Lorsch (1967)
on prin ciples o f o rgan izatio n al differentiation and integration and h o w they adapt
to different en viron m en tal conditions, in clu din g the m arket, tech nical-econ om ic
and the scientific su b -en viro n m en ts, o f different industries. W hereas B urn s and
Stalker treat an organ izatio n as an undifferen tiated w hole that is either m echanistic
or organic, Law rence and Lorsch recognize that m echanistic and organic structures
can coexist in different parts o f the sam e o rgan ization ow in g to the different
dem ands o f the fu n ctio n al su b -en viron m en ts. T h e w ork o f these earlier authors
had a p ro fo u n d im p act on o rgan ization al th eo ry and pro vid ed useful design
guidelines fo r in n o vatio n m anagem ent. B u rn s and Stalker's m odel rem ains h igh ly
relevant fo r o u r u n d erstan d in g o f the co n tem p o rary challenges facing m any o rg an ­
izations in their attem pts to m ove aw ay fro m the m echan istic tow ards the organic
form o f o rgan izin g, as in n o vatio n becom es m ore im po rtan t and the pace o f en viro n ­
m ental change accelerates. Law rence and Lorsch ’s suggestion that m echanistic
and organic structu res can coexist is reflected in the co n tem p o rary debate about
the im p o rtan ce o f d evelo p in g hybrid m odes o f organizations— “ am b id extrou s
o rgan izatio n s"— that are capable o f copin g w ith both evo lu tio n ary an d revo lu tio n ­
ary tech n ological changes (T u shm an and O ’ R eilly 1996; see also Sections 5.4.2
and 5.4.3).
Another important early contribution is the work of Mintzberg (1979) who
synthesized much o f the work on organizational structure and proposed a series of
archetypes that provide the basic structural configurations of firms operating in
different environments. In line with contingency theory, he argues that the success­
ful organization designs its structure to match its situation. Moreover, it develops a
logical configuration o f the design parameters. In other words, effective structuring
requires consistency of design parameters and contingency factors. T h e 4 configura­
tional hypothesis" suggests that firms are likely to be dominated by one o f the five
pure archetypes identified by Mintzberg, each with different innovative potential:
simple structure, machine bureaucracy, professional bureaucracy, divisionalized
form, and adhocracy. The characteristic features of the archetypes and their innova­
tive implications are shown in Table 5.1. The main thrust of the argument is that
bureaucratic structures work well in stable environments but they are not innovative
and cannot cope with novelty or change. Adhocracies, by contrast, are highly organic
and flexible forms o f organization capable o f radical innovation in a volatile envir­
onment (see also Section 5.3*3)*
Contingency theories account for the diversity of organizational forms in differ­
ent technological and task environments. They assume that as technology and
120 ALICE LAM
Table 5.1 Mintzberg's structural archetypes and their innovative potential!
O rg an izatio n Key features
archetype
Simple stru ctu re An organic type centrally controlled
by one person but can respond quickly
to changes in the environment, e.g.
small start-ups in high-technology.
Machine A mechanistic organization
bureaucracy characterized by high level of
specialization, standardization, and
centralized control. A continuous
effort to routinize tasks through
formalization of worker skills and
experiences, e.g. mass production
firms.
Professional A decentralized mechanistic form
bureaucracy which accords a high degree of
autonomy to individual professionals.
Characterized by individual and
functional specialization, with a
concentration of power and status in
the “a u th o rized experts." Universities,
hospitals, law, and accounting firms
are typical examples.
Divisionalized A decentralized organic form in which
form quasi-autonomous entities are loosely
coupled together by a central
administrative structure. Typically
associated with larger organizations
designed to meet local environmental
challenges.
Adhocracy A highly flexible p ro ject-b ased
organization designed to deal with
instability and complexity. Problem­
solving teams can be rapidly
reconfigured in response to external
changes and market demands. Typical
examples are professional
partnerships and software
engineering firms.
Sources: Mintzberg (1979); Tidd et al. (1997: 313-14); Lam (2000),
?
'■
■' ■.... . .............n :-
Innovative potential
Entrepreneurial and often
highly innovative, continually
searching for high-risk
environments. Weaknesses ait
the vulnerability to individual
misjudgement and resource
lim its on g ro w th .
Designed for efficiency and
stability. Good at dealing with
routine problems, but highly
rigid and unable to cope with
novelty and change.
The individual experts may be
highly innovative within a
specialist domain, but the
difficulties of coordination
across functions and
disciplines impose severe limits
on the innovative capability of
the organization as a whole.
An ability to concentrate on
developing competency in
specific niches. Weaknesses
include the "centrifugal pull"
away from central REtD towards
local efforts, and competition
between divisions which inhibit
knowledge sharing.
Capable of fast learning and
unlearning; highly adaptive
and innovative. However, the
unstable structure is prone to
short life, and may be driven
over time toward the
bureaucracy (see also Section
5,3.3).
 O R G A N I Z A T I O N A L I NNOVATI ON 121

product m arkets becom e m o re co m p lex and uncertain, and task activities m ore
heterogeneous and u n predictable, organizations w ill adopt m ore adaptive and
flexible structures, and they w ill do so by m o vin g aw ay from bureaucratic to organic
form s o f organ izin g. T h e u n d erlyin g difficulties in achieving the “ m atch,” how ever,
are not addressed in this stran d o f research. C o n tin gen cy theories neglect the
possibility that the factors identified as m ost im po rtan t in this th eo ry are susceptible
to different in terpretation s b y o rgan ization al actors (D aft and W eick 1984), and
ignores the influence o f other factors such as m anagerial choice (C h ild 1972; 1997) o r
institutional pressures (Pow ell and D iM ag g io 1991). These aspects w ill be discussed
in Sections 5.3 and 5.4.

5.2.2 Industrial Economics: Strategy, Structure,

and the Innovative Firm
The w o rk o f m icro -eco n o m ists in the field o f strategy considers organizational
structure as both cause and effect o f m anagerial strategic choice in response to
m arket o p p o rtu n ities. O rgan ization al fo rm s are constructed fro m the tw o variables
o f “ strategy” and “ structure.” T h e central argu m en t is that certain organizational
types o r attributes are m o re lik ely to yield su p erior in n o vative perfo rm an ce in a
given en viron m en t becau se th ey are m o re suited to reduce tran saction costs and
cope w ith alleged capital m arket failures. T h e m u ltid ivisio n al, o r M -fo rm ,
for exam ple, has em erged in response to increasing scale and co m p lexity o f enter­
prises and is associated w ith a strategy o f diversification into related p ro d u ct
and tech n ological areas (C h an d ler 1962). It can be an efficient in n o vato r w ithin
certain specific p ro d u ct m arkets, b u t m ay be lim ited in its ab ility to develop new
com petencies.
The theory o f “ the innovative enterprise” developed by Lazonick and West (1998)
is rooted in the Chandlerian framework, inasmuch as it focuses on how strategy
and structure determine the competitive advantage of the business enterprise. It
also builds on Lawrence and Lorsch’s (1967) conceptualization of organizational
design problems as differentiation and integration. This theory postulates that, over
time, business enterprises in the advanced economies have to achieve a higher degree
of “organizational integration” in order to sustain competitive advantage. Japanese
firms are said to have gained a competitive advantage in industries such as electron­
ics and automobiles over the USA because of their superior organizational capacity
for integrating shop-floor workers and enterprise networks, enabling them to plan
and coordinate specialized divisions o f labor and innovative investment strategies.
Lazonick and West also argue that those US firms (e.g. Motorola and IBM) that have
been able to sustain their competitive advantage also benefit from a high degree o f
organizational integration. The “organizational integration hypothesis directs
122 ALICE LAM

o u r attention to the social structure o f the enterprise and its internal cohesiveness
as a critical determ inant o f corporate strategy and in n o vative performance.
But this interpretation itself is insufficiently attentive to the contingency view-
point__the Japanese model of organizational integration works well in established
technological fields in which increm ental in n o vatio n is im portant* but not
necessarily in rapidly developing new fields w here rad ical in n o vation is vital for
com petitiveness.
Teece (1998) explains the links between firm strategy, structure, and the nature of
innovation by specifying the underlying properties o f tech nological innovation and
then proposing a related set o f organizational requirem ents o f the innovation
process. H is fram ew ork suggests that both the fo rm al (governance modes) and
inform al (cultures and values) structures, as w ell as firm s1 external network,
pow erfully influence the rate and direction o f their in n o vative activities. Based on
four classes o f variables including firm bou n daries, internal fo rm al structure, in­
ternal inform al structure (culture), and external linkages, the au th or identifies four
archetypal corporate governance m odes: m u ltip rod u ct integrated hierarchy, high-
flex silicon valley type, virtual corporation, and con glom erate. H e argues that
different organizational arrangem ents are suited to different types o f competitive
environm ents and differing types o f innovation. Teece (1998: 15 6 -7 ) illustrates the
argum ent by distinguishing between two m ain types o f in n o vation , namely “ au­
ton om ous” and “ system ic” innovation, and m atchin g them with different
organizational structures. A n autonom ous in n ovation is one that can be introduced
to the market w ithout m assive m odification o f related p ro d u cts and processes. An
exam ple is the introduction o f pow er steering w h ich did not initially require any
significant alternatives to the design o f cars or engines. T h is can often be advanced
rapidly by sm aller autonom ous structures, such as “ v irtu a l” firm s, accomplishing
necessary coordination through arm s-len gth arrangem ents in the open market. By
contrast, the m ove to front-w heel drive required the com plete redesign o f many
autom obiles in the 1980s. This type o f change is system ic in n o vation which favors
integrated enterprises because it requires com plex co o rd in atio n am ongst various
subsystems, and hence is usually accom plished under one “ roof.” These propos­
itions, however, have yet to be em pirically verified (Teece 19 9 8 :14 6 -7 ).
The work by micro-economists highlights the interaction between market and
organizational factors in shaping innovative performance, although it devotes little
attention to the internal dynamics and social processes within organizations. Many
of the empirical predictions within this literature on the relationship between firm
strategy and structure, and innovative performance have yet to be verified, and pose
intriguing opportunities for future research.
 ORGANIZATIONAL INNOVATION 1.23

5.3 O r g an izatio n al C o gn itio n ,

L e a r n in g and In n o v a t io n

5.3.1 The Cognitive Foundations of

Organizational Innovation
The “ structural perspectives” discussed above treat in n o vation as an o u tp u t o f
certain structu ral features; b u t som e organizational researchers regard in n ovation
as “ a process o f b rin g in g new, pro b lem -so lvin g, ideas into use” (A m abile 1988;
Kanter 1983). M ezias and G lyn n (1993: 78) define in n ovation as “ n o n rou tin e, sign­
ificant, and d isco n tin u o u s o rgan ization al change that em bodies a new idea that is
not consistent w ith the cu rrent concept o f the organization's business ” T h is a p ­
proach defines an in n o vative o rgan ization as one that is intelligent and creative
(G lyn n 1996; W o od m an et a l 1993), capable o f learning effectively (Senge 1990;
A gyris and S ch o n 1978), and creating n ew know ledge (N on aka 1994; N o n ak a and
Takeuchi 1995). C o h en and Levin th al (1990) argue that in n ovative ou tpu ts depend
on the p rio r accu m u latio n o f know ledge that enables in n o vato rs to assim ilate and
exploit new know ledge. F ro m this perspective, u nderstandin g the role o f cogn ition
and o rgan izatio n al learn in g in fostering o r in h ib itin g in n o vation becom es crucially
im portan t.
The co gn itively orien ted literatu re in organization and m anagem ent research is
rooted in cogn itive p sych o lo g y and analyzes the vario u s interven in g m ental p ro ­
cesses that m ed iate responses to the environ m en t (see H od gk in so n 2003). T h e term s
“ co gn ition ” o r “ co g n itive” refer to the idea that in d ivid u als develop m ental m odels,
b elief system s, an d know ledge structures that they use to perceive, construct, and
m ake sense o f their w o rld s and to m ake decisions ab ou t w hat actions to take (W eick
1979,1995; W alsh 1995). In d ivid u als are lim ited in their ab ility to process the co m p lex
variety o f stim u li co n tain ed in the external en viron m en t (Sim o n s “ b o u n d ed ratio n ­
ality” p ro b lem ), an d hence they develop “ m ental representations” to filter, interpret,
and recon stru ct in co m in g in fo rm atio n w h ich, u n d er certain circum stances m ay
fo rm the basis o f creative ideas and n ew insights, but m ay also lead to biases and
inertia. T h e p sych o lo g ical literatu re has focused p red o m in an tly on the in fo rm atio n
processin g consequences o f m ental m odels. O rgan ization and m an agem en t re­
searchers have exten d ed the analysis to the g ro u p and o rgan ization al levels. T h eir
analysis suggests that o rgan izatio n s develop collective m ental m odels an d in terpret­
ative schem es w h ich affect m an agerial d ecision m ak in g and organ izatio n al action.
Organizational cognition differs from individual cognition because it encompasses a
social d im en sio n . T h u s m u ch o f the research has focused on the so cio -co gn itive
connectedness, an d seeks to accou n t fo r the social processes in the fo rm atio n o f
collective c o g n itio n an d kn o w led ge structures.
124 ALICE LA M

The idea that organizations can think and act collectively, and serve as a repository
o f organized knowledge has stim ulated m uch research o n organizational learning
and knowledge creation. This w ork has sought to u n d erstan d h ow social inter­
action and group dynam ics w ithin organizations sh ape collective intelligence,
learning, and knowledge generation, and yields im p o rtan t insights into the micro­
dynam ics underpinning the innovative capab ility o f organizations. It has also
exam ined h ow shared m ental m odels o r interpretive schem es affect organizational
adaptiveness. On the positive side, som e argue that sh ared interpretative schemes
facilitate an organization's capacity to process and interpret inform ation in a
purposeful manner, prom ote organizational learning and collective problem solv­
ing, and thus enhance its adaptive potential (Fiol 1993; B ro w n and D uguid 1991),
Other studies suggest that organizational interpretative schem es can create “ blind
spots" in organizational decision m aking and b lo ck organ izatio n al change (Shri-
vasta and Schneider 1984; Shrivastava et al. 1987). T h e p a ra d o x seem s to be that
organizational cognition can be at once enabling and crip p lin g, like tw o sides o f the
same coin.
View ing organizational innovation fro m the cognitive perspective shifts our
analysis from organizational structures and system s to the processes o f organiza­
tional learning and knowledge creation. The analysis b elo w suggests that organiza­
tions with different structural form s v ary in their patterns o f learning and knowledge
creation, giving rise to different types o f innovative capabilities. Organizational
boundaries and the social context o f learning influence an organization s cognitive
vision and its capacity for radical change and innovation.

5*3.2 Organizational Learning and Knowledge Creation:

Shared Context and Collective Learning
Innovation can be understood as a process o f learning and know ledge creation
through which new problem s are defined and new know ledge is developed to solve
them. Central to theories o f organizational learning and know ledge creation is the
question o f how organizations translate in divid u al insights and knowledge into
collective knowledge and organizational capability. W hile som e researchers argue
that learning is essentially an individual activity (Sim on 1991; G ran t 1996), most
theories o f organizational learning stress the im portance o f collective knowledge as a
source o f organizational cap ab ility Collective know ledge is the accum ulated know­
ledge o f the organization stored in its rules, procedures, routines and shared norms
which guide the problem -solving activities and patterns o f interaction am ong its
mem bers. Collective knowledge resembles the “ m e m o ry " or “ collective m in d " o f the
organization (Walsh and Ungson 1991). It can either be a “ sto ck " o f knowledge
stored as hard data or represent knowledge in a state o f “ flo w " em erging from
 ORGANIZATIONAL INNOVATION 125

interaction. C o llective know ledge exists between rather than w ithin individ uals. It
can be m ore or less than the sum o f the in d ivid u als' know ledge, d epending on the
m echanism s that translate in d ivid u al into collective know ledge (G lyn n 1996). Both
in divid u als and o rgan ization s are learning entities. A ll learning activities, however,
take place in a social context, and it is the nature and bou n d aries o f the context that
m ake a difference to learning outcom es.
M u ch o f the literature o n o rgan ization al learning p oin ts to the im po rtan ce o f
social interaction, context, and shared cognitive schem es for learning and kn ow -
ledge creation (N o n ak a 1994; A g y ris and Schon 1978; Lave and W enger 1991; Brow n
and D u gu id 1991, 1998). T h is bu ilds on P o lan yi’s (1966) idea that a large part o f
hum an kn ow ledge is subjective and tacit, and cannot be easily codified and tran s­
m itted in dep en d en t o f the kn o w in g subject. H ence its transfer requires social
interaction an d the d evelopm en t o f shared u nd erstand in g and co m m o n in terpret­
ative schem es.
N o n ak a s th eo ry o f organ izatio n al know ledge creation is rooted in the idea that
shared co gn itio n and collective learning constitute the fo u n d atio n o f organizational
know ledge creation (N o n ak a 1994; N o n ak a and Takeuchi 1995). At the heart o f the
th eo ry is the idea that tacit k n o w in g constitutes the origin o f all h u m an know ledge,
and o rgan izatio n al know ledge creation is a process o f m ob ilizin g in d ivid u al tacit
know ledge an d fo sterin g its in teraction w ith the explicit know ledge base o f the firm .
N on aka argues that know ledge needs a context to be created. H e uses the Japanese
w ord “ b a ” w h ich literally m eans “ place,” to describe such a context. “ B a” pro vid es a
shared social and m ental space fo r the interpretation o f in fo rm atio n , interaction,
and em erging relation sh ips that serves as a fo u n d atio n fo r know ledge creation.
P articipatin g in a “ ba” m eans tran scen ding one's lim ited cognitive perspective o r
social b o u n d a ry to engage in a d yn am ic process o f know ledge sh aring and creation.
In a sim ilar vein , the n o tio n o f “ c o m m u n ity o f p ractice” developed in the w o rk o f
Lave and W enger (19 9 1), W enger (1998), and B ro w n and D u gu id (1991; і 99&)5
suggests that o rgan izatio n al m em bers construct their shared identities and perspec­
tives th ro u gh “ practice,” that is shared w o rk experiences. Practice pro vid es a social
activity in w h ich sh ared perspectives and cognitive repertoires develop to facilitate
know ledge sh arin g an d transfer. H ence, the w o rk g ro u p p ro vid es an im p o rtan t site
w here intense learn in g an d know ledge creation m ay develop. T h e grou p , placed at
the in tersection o f h o rizo n tal and vertical flow s o f know ledge w ith in the o rgan iza­
tion, serves as a b rid ge betw een the in d ivid u al and organ izatio n in the know ledge
creation process. N o n ak a ’s th e o ry stresses the critical role o f the sem i-au to n o m o u s
project team s in kn ow led ge creation. M u ch o f the recent literature o n new and
in n o vative fo rm s o f o rg an izatio n also focuses on the use o f decentralized, group-
based structure as a key organizing principle.
Many organizational and management researchers regard the firm as a critical
social context where collective learning and knowledge creation takes place. Nonaka
and--Takeuchi- (1995) talk ab o u t the “ kn ow led ge creating c o m p a n y ” A rg y ris and
126 ALICE LA M

Schon (1978 ) suggest that an organization is, at its ro o t, a cogn itive enterprise that
learns and develops knowledge. "O rgan ization al k n o w led g e" essentially refers to the
shared cognitive schem es and distributed co m m o n u n d erstan d in g w ithin the i m
that facilitate knowledge sharing and transfer. It is sim ila r to N elson and Winter3!
{1982) concept o f "organizational routines * a kin d o f collective know ledge rooted in
shared norm s and beliefs that aids jo in t prob lem so lvin g an d capable o f supporting
com plex patterns o f action in the absence o f w ritten rules. T h e notion o f "cote
com petence" (Prahalad and H am el 1990) im plies that the learn in g and knowledge
creation activities o f firm s tend to be cum ulative and path -d epen dent. Firm s tend to
persist in what they do because learning and know ledge are em bedded in social
relationships, shared cognition, and existing w ays o f d o in g things (Kogut and
Zander 1992). Several authors have analyzed h o w collective learning in technology
depends on firm s' cum ulative com petences and evolves alo n g specific trajectories
(Dosi 1988; Pavitt 1991J. Thus, the shared context and social iden tity associated with
strong group-level learning and know ledge accu m u latio n processes m ay constrain
the evolution o f collective knowledge. Firm s m ay find it d ifficu lt to unlearn past
practices and explore alternative ways o f d oin g things. Levinthal and M arch (1993)
argue that organizations often suffer from "le a rn in g m yo p ia," and have a tendency
to sustain their current focus and accentuate their distinctive com petence, what they
call falling into a "com petency trap." The em pirical research by Leonardo-Barton
(1992) illustrates how firm s’ "core capabilities" can turn into "co re rigidities" in new
product developm ent.
A n inherent difficulty in organizational learning is the need to maintain an
external b oun dary and identity while at the sam e tim e keeping the boundary
sufficiently open to allow the flow o f new know ledge and ideas from outside.
M arch (1991) points out that a fundam ental tension in organizational learning is
balancing the com peting goals o f "th e exploitation o f old certainties” and the
exploration o f new p ossib ilities” W hereas know ledge creation is often a product
o f an organization s capability to recom bine existing know ledge and generate new
applications from its existing knowledge base, radically new learning tends to arise
from contacts with those outside the organization w ho are in a better position to
challenge existing perspectives and paradigm s. E m p irical research has suggested that
sources o f innovation often lie outside an organization ( vo n H ip pel 1988; Lundvall
1992). External business alliances and netw ork relationships, as w ell as using new
personnel to graft new knowledge onto the existing learning system s, are important
m echanism s for organizational learning and know ledge renew al in an environment
characterized by rapid technological developm ent and d isru p tive changes. The
dynam ic capability” perspective argues that the lon g-term com petitive perform-
ance o f the firm lies in its ability to build and develop firm -sp ecific capability and,
sim ultaneously, to renew and reconfigure its com petences in response to an environ­
m ent m arked by creative destruction” (Teece and Pisano 1994}. T h u s, a fundamen­
tal organizational challenge in innovation is not sim p ly to m aintain a static balance
 ORGANIZATIONAL INNOVATION 12/

between exp lo itatio n and exp loration , o r stability and change, but a co n tin u ou s
need to balance and co ord in ate the tw o d yn am ically th rou gh ou t the organization.

5*3*3 Two Alternative Models o f Learning and Innovative

Organizations: “ J-form ” vs. “Adhocracy”
W hile all o rgan izatio n s can learn and create know ledge, their learning patterns and
innovative capabilities vary. D u rin g the past decade, a large literature has discussed
new o rgan izatio n al m odels and concepts designed to su p p o rt organizational learn ­
ing and in n o vatio n . These m odels include “ high perfo rm an ce w o rk system s” o r
“ lean p ro d u ctio n ” (W om ack et al. 1990), pioneered b y Japanese firm s in the a u to ­
m obile in d u stry; an d the “ N -fo rm co rp o ratio n ” (H ed lu n d 1994) and “ hypertext
organization” (N o n ak a an d Takeuchi 1995). M o re recent concepts such as “ cellular
fo rm s” (M iles et a l 1997), “ m o d u lar fo rm s” (G alu n ic and E isen hardt 2001) and
“ project-based n etw o rk s” (D eF illip p i 2002) reflect the grow th o f flexible and ad a p ­
tive fo rm s o f o rgan izatio n w ith a strategic focus on entrepreneurship and radical
in n ovation in kn o w led ge-in ten sive sectors o f the econom y. These studies highlight
the different w ays in w h ich firm s seek to create learning organizations capable o f
contin uous p ro b lem so lvin g and in n o vation . V ery few studies explain the nature o f
the learn in g processes u n d erp in n in g these structu ral form s, the types o f in n ovative
com petences generated and the w id er institu tion al context w ith in w h ich this o rg a n ­
izational learn in g is em bedded.
A closer exam in atio n o f the literature on new fo rm s suggests that the vario u s
m odels o f in n o vative o rgan izatio n s can be b ro ad ly classified into tw o p o lar ideal
types, nam ely, the “ J-fo rm ” and “ ad h o cracy”. The fo rm er refers to an o rgan izatio n
w hich is g o o d at cu m u lative learn in g and derives its in n o vative capabilities fro m the
d evelopm ent o f o rg an izatio n -sp ecific collective com petences and p ro b lem -so lvin g
routines. T h e term J-fo rm is used because its archetypal features are best illustrated
b y the “ Jap an ese ty p e ” o f organ izatio n s, such as A o k i’s (1988) m o d el o f the “ J-firm ”,
and N o n ak a an d T akeu ch i’s (1995) “ k n ow ledge-creatin g com panies.” A d h o cracy
(M intzberg, 1979), b y con trast, tends to rely m o re u p o n in d ivid u al specialist exp ert­
ise organ ized in flexib le m arket-b ased project team s capable o f speedy responses to
changes in k n o w led ge and skills, an d integratin g n ew kinds o f expertise to generate
radical n ew p ro d u cts and processes. M in tzb erg’s term is used here to capture the
d yn am ic, en trep ren eu rial, an d adap tive character o f the kin d o f organ ization
typified b y Silicon Valley type companies (Bahrami and Evans 2000). Both the
“ J-fo rm ” a n d “ a d h o c ra c y ” are learn in g o rgan ization s w ith strong innovative
capabilities, b u t th e y d iffe r m ark ed ly in their stru ctu ral fo rm s, patterns o f learning,
and the ty p e of innovative competences generated.
128 ALICE LAM

The J-form organization relics on know ledge that is em bed ded in its operating
routines, team relationships, and shared culture. L earn in g and know ledge creation
within the J-fo rm takes place w ithin an “ o rgan izatio n al co m m u n ity” that in­
corporates sh op-floor skills in problem solving, an d inten sive interaction and
knowledge sharing across different fu nctional units. T h e existence o f stable organ­
izational careers rooted in an internal lab or m arket p ro vid e an incentive for
organizational m em bers to com m it to organizational goals and to develop firm-
specific problem -solving knowledge for co n tin u ou s p ro d u ct and process im­
provem ent. N ew knowledge is generated th ro u gh the fu sion, synthesis, and
com bination o f the existing knowledge base. The J-fo rm tends to develop a strong
orientation towards pursuing an increm ental in n o vatio n strategy and to do well in
relatively m ature technological fields characterized b y rich possibilities o f combin­
ations and increm ental im provem ents o f existing co m p o n en ts and products (e.g.
m achine-based industries, electronics com ponents, and autom obiles). But the
J-form 's focus on nurturing organizationally em bedded, tacit knowledge and its
em phasis on continuous im provem ent in such know ledge can inhibit learning
radically new knowledge from external sources. T h e d isap p o in tin g performance of
Japanese firm s in such fields as software and b io tech n o lo g y d u rin g the 1990s may
constitute evidence o f the difficulties faced b y “ J-fo rm firm s'" in entering and
innovating in rapidly developing new technological fields (Lam 2002; Whitley
2003; see also Box 5.2).
The adhocracy is an organic and adaptive form o f organ izatio n that is able to fuse
professional experts w ith varied skills and know ledge into adh oc project teams for
solving com plex and often highly uncertain problem s. L earn in g and knowledge
creation in an adhocracy occurs w ithin professional team s that often are composed
o f em ployees from different organizatons. C areers are u su ally structured around a
series o f discrete projects rather than advancing w ithin an in tra-firm hierarchy The
resulting project-based career system is rooted in a relatively flu id occupational
labor m arket which perm its the rapid reconfiguration o f h u m an resources to align
with shifting m arket requirem ents and technological changes. T h e adhocracy has a
m uch m ore perm eable organizational b o u n d ary that allow s the insertion o f new
ideas and knowledge from outside. This occurs th rou gh the recruitm ent o f new staff,
and the open professional networks o f the organizational m em bers that span
organizational boundaries. The adhocracy derives its com petitive strength from its
ability to reconfigure the knowledge base rapidly to deal w ith high levels o f technical
uncertainty, and to create new knowledge to produ ce n ovel in n ovation s in emerging
new industries. It is a very adaptive form o f o rgan ization capable o f dynamic
learning and radical innovation. H owever, the flu id structure and speed o f change
m ay create problem s in knowledge accum ulation, since the organization's compe­
tence is em bodied in its m em bers' professional expertise and m arket-based know­
how which are potentially transferable. The adh ocracy is subject to knowledge loss
when individuals leave the organization. Starbuck (1992: 725), for instance, talks
 ORGANIZATIONAL INNOVATION 129

Box 5.2 Japan: an exam ple o f organizational community model o f learning

The Japanese economy is characterized by a high level o f cooperation and organiza­

tional integration. This occurs through extensive long-term collaboration between
firms in business groups and networks. Additionally, integration within large firms is
particularly strong. Japanese social institutions and employment practices foster the
close involvement o f shop-floor workers in the development o f organizational cap­
ability. The successful state education system and large, company-driven networks
equip the majority o f workers with a high level o f skills that employers respect and so
can rely on them to contribute usefully to innovation activities. The internal labor
market system is characterized not only by long-term attachment but also by well-
organized training and job rotation schemes. These practices promote continuous
skills formation through learning-by-doing and systematic career progression (Lam
1996; 1997). Hence, a strong organizational capacity to accumulate knowledge and
learn incrementally. Over the past three decades, Japanese firms have gained inter­
national competitive advantage in those industries such as transport equipment, office
machines, consumer electronics, electronic components for computing equipment,
and telecommunication hardware. The strength o f Japan in these sectors stems from
the capability o f firms to develop highly flexible production systems through the close
integration o f shop-floor skills and experience, the tight linkages between R&D,
production, and marketing, and a unique innovation strategy based on continual
modification and upgrading o f existing components and products (Womack et al.
1991). Conversely, organization-specific and path-dependent learning have con­
strained Japan's success in a number o f leading-edge technological fields. Japan finds
it harder to excel in sectors which do not exclusively rely on incremental upgrading o f
system components (e.g. aerospace; supercomputers) and those in which fast-paced
radical innovations are crucial for success (e.g. pharmaceuticals and biotechnology).
The human-network-based interaction and internal tacit knowledge transfer appear to
be less effective in coordinating systems involving complex interactions among com­
ponents. The organizational community model o f learning limits the development o f
highly specialized scientific expertise, and makes it difficult to adopt radically new skills
and knowledge needed for radical learning in emerging new technological fields.

about the “ porous boundaries” of this type of organizations and points out that they
often find it hard to keep unique expertise exclusive.
The long-term survival of this loose, permeable organizational form requires the
support of a stable social infrastructure rooted in a wider occupational co m m u n ity
or localized firm networks. The example of high-technology firms in Silicon Valley
highlights the importance for the “ adhocracy” of supportive local labor markets and
other external institutions typically included in analyses of national, sectoral and
regional innovation systems (Saxenian 1996» Bahrami and Evans 2000; Angels 2000;
see also Ch. 7 in this volume by Edquist, Ch. 11 by Asheim and Gertler, and C k 14 b y
Malerba, as well as Box 5.3).
130 ALICE LAM

Box 5.3 Silicon Valley; an example of professional team model of learrttW|$|?

. . . ■ • ■ .................................................. ■ ' ' 1• 1 ' " ' " 1I 1 ~ '"r "" '■ ■ V , 4 т І '.~~

Silicon Valley has been an enormously successful and dynamic region characterisedЩ.
rapid innovation and commercalization in the fast growing technological fidd$/ j f c f
core industries o f the region include microelectronics, semiconductors, computer
networking, both hardware and software, and more recently biotechnology Firms'
operating in these industries undergo frequent reconfiguration and realignment in
order to survive in a constantly changing environment marked by incessant innov­
ation. The availability o f a large pool o f professional experts with known reputations in
particular fields enables firms to quickly reconstitute their knowledge and sk i! base m
the course o f their innovative endeavours, The rapid creation o f new start-up firms
focusing on novel innovative projects, and the ease with which project-based firms are
able to assemble and reassemble their teams o f highly skilled scientists and engineers Ш
engage in new innovative activities are central to the technological and orpnkatioiM
dynamism o f the region. The high rate o f labor mobility and extensive hiring and firing
creates a permissive environment for entrepreneurial start-ups and flexible recon­
figuration o f project teams and knowledge sources. Labor mobility within the context
o f a region plays a critical role in the generation o f professional networks and facilitates
the rapid transmission o f evolving new knowledge, a large part o f which may be tacit
Such a regionally based occupational labor market provides a stable social context and
shared industrial culture needed to ensure the efficient transfer o f tacit knowledge m an
interfirm career framework. The shared context and industry-specific values within the
regional community ensure that tacit knowledge will not be wasted when one changes
employers, and this gives the individual a positive incentive to engage in tacit “ know­
how” learning (Deffillipi and Arthur 1996). A regionally based labor market and
networks o f firms create a stable social structure to sustain collective learning and
knowledge creation within and across firm boundaries. The creation o f a wider social
learning system amplifies the learning and innovative capability o f the individual firms
locating within the system. It provides an anchor o f stability for fostering and sustain­
ing the innovative capability o f the adhocracy.

A lthough firm s in the high-technology sectors are u n d er intense pressure to learn

faster and organize m ore flexibly, evidence thus far suggests that com plete adhoc­
racies rem ain rare, Adhocracies are usually confined to organizational subunits
engaged in creative w ork (e.g. “ skunk w o rk ” adhocracies) (Q uinn 1992: 294-5), or
know ledge-intensive professional service fields (e.g. law, m anagem ent consultan­
cies, software engineering design) where the size o f the firm is generally relatively
small, enabling the whole organization to fu nction as an interdependent network of
project teams (D eFillippi 2002). Attem pts b y large co rp oratio n s to adopt the
adhocracy m ode have proved to be difficult to sustain in the lo n g run. An illustration
is the case o f O ticon, the Danish m anufacturer o f hearing aids, w hich adopted a
radical form of project-based organization (described as the “ spaghetti organiza­
tion ) to stimulate entrepreneurship and innovation but o n ly to find itself giving
 ORGANIZATIONAL INNOVATION 131

w ay to a m o re trad itio n al m a trix o rgan ization a decade later (Foss 2003; see Box 5.4).
Elsewhere, the m ost successful exam ples o f adhocracies are fou nd in region ally based
industrial co m m u n ities, as in the case o f Silicon Valley, and other h igh -tech n olo gy
clusters. There, the agglom eratio n o l firm s creates a stable social context and shared

Box 5,4 Oticon: the rise and decline of the "spaghetti organization"

Oticon, a Danish electronics producer, is one o f the world market leaders in hearing
aids. The company became world famous for radical organizational transformation in
the early 1990s, and has been treated as an outstanding example o f the innovative
benefits that a radical project-based organization may generate (Verona and Ravasi
2003). The “ spaghetti organization,” as it has come to be known, refers to a flat, loosely
coupled, project-based organization characterized by ambiguous job boundaries and
extensive delegation o f task and project responsibilities to autonomous teams. The
adoption o f the radical structure in 1990 represented a dramatic break from the
traditional hierarchical, functional-based organization that the company had relied
upon in the past.
The background to the implementation o f the spaghetti organization was the loss o f
competitive advantage that Oticon increasingly realized during the 1980s. Although
for decades the company had played a leading role in the hearing aids industry, at the
end o f the 1980s its products largely depended on a mature and declining technology.
The advent o f digital technology had gradually led to a shift in the technological
paradigm during the 1980s and Oticon was losing ground to its major competitors.
In 1990, the company underwent extensive restructuring in response to the crisis. The
spaghetti organization was introduced, aiming at developing a more creative and
entrepreneurial organization. The radical reorganization had immediate and strong
performance effects, resulting in a series o f remarkable innovations during the 1990s.
Despite this success, the spaghetti organization was partially abandoned from about
1996 and was gradually superseded by a more stable, traditional matrix organization.
The study by Foss ( 2003) suggests that the Oticon spaghetti organization had
encountered severe problems o f coordination and knowledge sharing between projects
because o f the fluid and adhocratic nature o f project assignments, and difficulties in
ensuring employee commitments to projects. M ore notably, Foss argues that the
spaghetti organization, as an “ internal hybrid” (Le. the infusion o f elements o f market
autonomy and flexibility into a hierarchy), was inherently unstable partly because
o f the motivational problems caused by “ selective intervention” Attempts by man­
agement selectively to intervene in project selection and coordination became increas­
ingly at odds with the official rhetoric that stresses self-organization. The mounting
frustration among employees eventually led to the retreat from the radical spaghetti
or^anBation.
Mthough the Oticon experiment is widely considered as a success story oforganiza-'
':>:':fionali'iffibvation, the partial retreat from the spaghetti organization iflustrates ffie
: \ihitei#:nt>diffic;ulties'. o f sustaining complete adhocracy.
.v
132 ALICE LAM

cognitive fram ew ork to sustain collective learning and reduce uncertainty associated
with swift form ation o f project team s and o rgan izatio n al changes* A n important
item for future research is a clear identification o f the p o p u latio n o f adhocracies in
different industries and regions o f the global “ kn ow ledge-b ased econom y/5 Current
w ork on this organizational type consists largely o f case studies and anecdotes,

5*3.4 The Social Embeddedness of Organizations and

their Innovative Capabilities
A lthough com petitive pressures are felt b y nearly all o rgan ization s in the advanced
econom ies, the em ergence and structure o f new organ ization al fo rm s are affected by
their particular institutional contexts* A large literature contrasts the patterns of
innovation and technological change in different co u n tries and attributes these
differences to national institutional fram ew orks and the w ays in which they
shape organizational form s and innovative com petences (W hitley 2000, 2003;
H ollingsw orth 2000). The “ varieties o f capitalism ” fram ew ork, for example,
makes a stylized contrast between coordinated (C M E ) and liberal m arket economies
(LM E). It highlights how differences in lab or m arket organization* training systems,
and societal norm s and values governing business and econom ic relationships
encourage firm s to organize and coordinate their skills and know ledge resources
differently to pursue distinctive innovation strategies (Soskice 1999; Hall and
Soskice 2001)*
M uch o f the w ork adopting the “ varieties o f cap italism ” perspective argues that
“ coordinated m arket econom ies” such as Jap an and G erm an y have developed insti­
tutions that encourage long-term em ploym ent and business relationships, facilitat­
ing the developm ent o f distinctive organizational com petences conducive to
continuous but increm ental innovation. T h e J-fo rm o rgan ization is facilitated by
this type o f institutional context. Conversely, “ liberal m arket econom ies” like the US
and U K are better able to foster adhocracies in rap id ly em erging new industries
through radical innovation. The m ore perm issive institu tion al environm ent associ­
ated with the US and U K facilitates high lab or m o b ility between firms, and
reconfiguration o f new knowledge and skills w ithin flexible fo rm s o f organization
to support risky entrepreneurial activities. In ad d ition to labor markets, other
institutional features such as education system s and finan cial m arkets also shape
the developm ent o f skills and innovative com petences o f firm s (Lam 2000; Casper
2000, see also О Sullivan, this volum e). The linkages am o n g institutions, organiza­
tions, and innovation are m ore com plex than the sim p lified stylized contrast
between J-fo rm and adhocracy suggests. W hat the po lar-typ e contrast suggests is
that the ability o f firm s to develop different patterns o f learning and innovative
 ORGANIZATIONAL INNOVATION 133

com petences is contingent u p on the w id er social context, and that institutional

fram ew orks affect h o w firm s develop and organize their innovative activities in
different societies. Societal institution s create constraints on and possibilities for
firm s to develop d ifferent types o f organizations and innovative com petences, giving
rise to d istin ctive n ation al in n ovative trajectories.

5.4 O r g a n iz a t io n a l C h a n g e
a n d In n o v a t io n

O rganizational theories have lo n g considered the w ays in w hich organizations

evolve and adap t to their environ m en ts, in clu d in g the influence o f technological
change on the evo lu tio n o f organ ization s (see Tushm an and N elson 1990). A core
debate concerns w h eth er o rgan izatio n s can change and adapt to m ajo r d isco n tin u ­
ous tech nological change an d en viron m en tal shifts, o r w hether radical change in
organizational fo rm s occu rs p rin cip ally at the p o p u latio n level th rou gh the process
o f selection (L ew in and V olberda 1999). T h is literature includes at least three broad
view s on the n atu re o f o rgan izatio n al adap tation an d change. O rganizational eco l­
ogy and in stitu tion al theories, as w ell as evo lu tio n ary theories o f the firm , em phasize
the po w erfu l forces o f o rgan izatio n al in ertia and argu e that organizations respon d
only slow ly and in crem en tally to en viron m en tal changes. T h is strand o f w o rk
focuses o n the w ay en viron m en ts select organizations, and h o w this selection
process creates change in organ izatio n al form s. A second view, the pu nctuated
equilib rium m o d el, p ro p o ses that ogan izations evolve th ro u gh lo n g p eriods o f
increm ental an d e v o lu tio n a ry change pu n ctu ated b y d isco n tin u o u s o r revo lu tio n ­
ary change. It sees o rgan izatio n al evo lu tio n as closely linked to the cyclical pattern o f
technological change. T h e p u n ctu ated m odel regards organ ization al tran sfo rm ation
as a d isco n tin u o u s event o ccu rrin g o ver a sh ort p erio d o f tim e. T h e third perspec­
tive, w h ich m igh t be d escribed as strategic adaptation, argues that organ ization s are
not alw ays passive recipien ts o f en viron m en tal forces b u t also have the p o w er to
influence an d sh ape the en viron m en t. T h e strategic ad ap tation perspective stresses
the role o f m an ag erial action an d o rgan izatio n al learning, and the im p o rtan ce o f
continuous change and adaptation in coping with environmental turbulence and
uncertainty,
T h e fo llo w in g sections exam in e their m ain argu m en ts an d relevance to o u r
u n d erstan d in g o f the relatio n sh ip s between organizational change and innovation.
134 ALICE LAM

5.4.1 Incremental/Evolutionary View

of Organizational Change
Organization population ecologists (e.g. Hannan and Freeman 1977> 1984) argue
that individual organizations seldom succeed in m ak in g rad ical changes in strategy
and structure in the face o f environm ental turbulence because they are subject to
strong inertial forces. Such forces are inherent in the established structures o f the
organization which represent relatively fixed repertoires o f h igh ly reproducible
routines. W hile giving organizations reliability and stability, these routines also
make them resistant to change. A s a result, o rgan ization s respon d relatively slowly
to threats and opportunities in the environm ent. O rgan ization al ecology theories
posit that adaptation o f organizational structures w ith in an in d u stry occurs princi­
pally at the population le v e l w ith new o rganizations rep lacin g the old ones that fail
to adapt.
The institutional perspective on organizations also em phasizes the stability and
persistence o f organizational form s in a given p o p u latio n o r field o f organizations
(D iM aggio and Powell 1983; Zucker 1987). A m a jo r source o f resistance to change
arises from the norm ative em beddedness o f an o rgan ization w ith in its institutional
context. O rganizations are socially defined and operate w ith in a web o f values,
norm s, rules, and beliefs and taken-for-granted assu m p tio n s that they represent
values, interests, and cognitive schem as o f organizational and institutional actors
which are hard to change (H inings et al. 1996). In this view , organizational change
consists largely o f constant reproduction and rein forcin g o f existing modes of
thought and organization (G reenw ood and H in in gs 1996). In other words, organiza­
tional change is usually convergent change that occurs w ithin the parameter of an
existing archetype, rather than revolu tion ary change w h ich involves m oving from
one archetype to another.3
Evolutionary theories o f the firm (N elson and W inter 1982) also argue that
organizations are subject to inertial forces. O rganizations accum ulate know-how
and tacit knowledge in the course o f their d evelopm ent, and the resulting organiza­
tional routines and skills becom e core com petences and are difficult to change.
E volutionary theories regard organizational change as a p ro d u ct o f the search for
new practices in the neighborhood of an organization s existing practices, that is,
local search, and thus organizational routines and skills change only slowly and
incrementally.
In the face o f environm ental change, new entrants w ith in the industry may
displace the established organizations that cannot adapt fast enough; new organiza­
tional form s thus tend to evolve and develop fro m the entrepreneurial activities of
new firm s. This view point is consistent w ith the w idespread argum ent in the
litei ature on technological innovation that it is u sually new firm s which pioneer
novel form s o f organization to take full advantage o f radical changes in technology
(Schum peter 1950; Aldrich and M ueller 1982). H ow ever, the relative importance
 ORGANIZATIONAL INNOVATION 135

o f new entrants versus established organizations in d eveloping new fo rm s o f organ-

izing is p artly shaped b y the scale and pace o f environ m en tal change* Som e evidence
suggests that the effects o f tech nological change on organizational evolution depend
on w hether the new tech n o logy destroys or enhances the com petences o f existing
organizations (Tushm an and A n d erson 1986; H en derson and C lark 1990). The
general o b servation is that new entrants play a m uch m ore significant role in
organizational evo lu tio n in the face o f “ co m peten ce-d estroyin g55 technological
innovations; w hile established o rganizations are in a better p osition to initiate
changes to adap t to “ co m p eten ce-en h an cin g” technological changes.
The ab ility o f an organ izatio n to adapt to technological change is thus influenced
by the speed at w h ich n ew com petences and skills can be developed to m atch the
dem ands o f the n ew technologies. T h is is anoth er reason to expect the institutional
context to p lay an im p o rtan t role in shaping the dynam ics o f organizational change,
for reasons noted above. N ew firm s have played a m u ch m ore p ro m in en t role in
capitalizing o n the new o p p o rtu n ities opened b y radical technological changes in the
United States than in oth er indu strial econom ies because o f the flexib ility o f
profession al la b o r m arkets and venture capital m arkets. In co ord in ated m arket
econom ies such as Jap an o r G erm an y, new firm s are not created as q u ickly because
o f the in flexib ility o f the lab o r m arket and relative absence o f ventu re capital. A s a
result, established o rgan izatio n s m a y have m ore tim e to create new organizational
structures and com petences to adap t to tech nological changes. T h e relative im p o rt­
ance o f selection versu s ad ap tation as a m echan ism u n d erlyin g the creation o f new
organizational fo rm s thus m a y v a ry betw een different contexts. E co lo g y and evo lu ­
tio n ary theories o f o rgan izatio n al change have tended not to take these contextual
factors in to account.

5.4.2 Punctuated Equilibrium and Discontinuous

Organizational Transformation
In contrast, the punctuated equilibrium model proposes that organizations are
capable o f initiating revolutionary structural change during periods o f environ­
mental turbulence. It depicts organizations as evolving through relatively long
periods o f stability (equilibrium periods) in their basic patterns o f activity that are
punctuated by relatively short bursts o f fundamental change (revolutionary
periods) (Gersick 1991; Romanelli and Tushman 1994). It argues that organizations
will typically accomplish fundamental transformations in short, discontinuous
bursts o f change involving most or all key domains o f organizational activity.
These include changes in strategy, structure, power distribution, and control
systems. Punctuated equilibrium theorists argue that the common state o f
13 6 ALICE LAM

organizations is one o f stability and inertia, and as a result, these “ revolutionary

periods" provide rare opportunities for organ ization s to break the grip o f structural
and cultural inertia* In this view, organizations are m ost likely to introduce radical
changes in tim es o f perform ance crisis o r w hen they are con fron ted with disruptive
environm ental conditions such as radical com petence d estroyin g new technologies
(Anderson and Irish m an 1990). A n um ber o f em pirical studies based on company
histories (e.g. Tushm an, N ew m an, and R om anelli 1986; R om an elli and Tushman
1994) show that in m any organizations fu ndam ental o rgan ization al transformations
occur according to the patterns predicted b y the pu n ctu ated m odel. Other studies
(e.g. M iller and Friesen 1982; Virany, Tushm an, and R o m an elli 1992) show that
organizations that were able to drastically tran sfo rm them selves perform better
than those that changed increm entally. H ow ever, m ost o f the em pirical evidence
supporting the radical transform ative m ode o f o rgan izatio n al change was based on
retrospective archival studies o f su rvivin g com panies. T h is approach does not
perm it analysis o f the dynam ics o f the change process, and fails to account for
unsuccessful transitions.
The punctuated m odel also suggests that the u n d erlyin g dynam ics o f techno­
logical change influence patterns o f organizational evolu tion . T his argument builds
on the technology cycle m odel developed by A n d erson and Tushm an (1990) which
proposes that technological progress is characterized by relatively long periods of
increm ental, com petence-enhancing in n ovation d evoted to elaboration and im­
provem ent in dom inant design. These periods o f increasing consolidation and
organizational alignm ent are punctuated by radical, com petence-destroying tech­
nological discontinuities which pose fu ndam ental challenges and strategic oppor­
tunities for organizations. The im plication o f the tech n o logy cycle concept is that the
com petitive environm ent repeatedly changes over tim e, and successful organiza­
tions accordingly have to initiate periodic d iscon tin u ou s o r revolu tion ary change to
accom m odate changing environm ental conditions. A fu ndam ental challenge facing
organizations is to develop diverse com petences and capabilities to shape and deal
with the technology cycle, Tushm an and O 'R eilly (1996; 1999) argue that firms
operating in the turbulent technological environ m en t need to becom e “ ambidex­
trous , that is, capable o f sim ultaneously pu rsu in g both increm ental and discontinu­
ous technological changes.4
The punctuated m odel provides im portant insights into patterns of organiza­
tional evolution and their relationship to the u nd erlying d yn am ics o f technological
change, but it is largely descriptive. T his m odel assum es that new organizational
form s would emerge during periods o f radical, d isco n tin u o u s change; but fails to
address the crucial question of how organizational actors create new form s during
the revolutionary period. The m odel also does not address the long-term prospects
for survival of the new oiganizational form s that em erge d u rin g the revolutionary
period.
 ORGANIZATIONAL INNOVATION 137

5.4.3 Strategic Adaptation and Continuous Change

Theories o f strategic organ izatio n al adaptation and change focus on the role o f
m anagerial action and strategic choice in sh apin g organizational change (C hild 1972;
1997; B u rglem an 1991). T h ey vie w the evolu tion o f organizations as a prod u ct o f
actors" decisions and learning, rather than the ou tcom e o f a passive environ m en tal
selection process. O rganizational agents are seen as en joyin g a kin d o f “ boun ded
autonomy."" A cco rd in g to C h ild (1997: 60), organizational action is b ou n d ed b y the
cognitive, m aterial, and relational structures internal and external to the o rgan iza­
tion, but at the sam e tim e it im pacts u p on those structures. O rganizational actors,
through their action s and “ enactment"" (W eick 1979), are capable o f redefin in g and
m o d ifyin g structures in w ays that w ill open up new possibilities for future action. As
such, the strategic choice perspective projects the po ssib ility o f creativity and
innovative change w ith in the organization.
M an y strategic ad ap tation theorists view organizational change as a co n tin u ou s
process en com p assin g the p arad o xical forces o f co n tin u ity and change, rather than
an abrupt, d isco n tin u o u s, episod ic event described b y the pu nctuated equ ilib riu m
m odel. C o n tin u ity m ain tain s a sense o f id en tity fo r organizational learning (W eick
1996; K o d am a 2003), and pro vid es political legitim acy and increase the acceptability
o f change am o n g those w h o have to live w ith it (C h ild and Sm ith 1987). B u rg le m a n s
(1983; 1991) stu d y o f the Intel C o rp o ra tio n illustrates h o w the co m p an y successfully
evolved fro m a m e m o ry to a m icro p ro cesso r co m p an y b y co m b in in g the tw in
elem ents o f c o n tin u ity and change for strategic renewal. B u rglem an argues that
consistently successful o rgan ization s use a co m b in atio n o f “ induced"" and “ au to n o ­
mous"" processes in strategy m ak in g to b rin g abou t o rgan ization al renewal.
A ccord ing to the author, the in d u ced process develops initiatives that are w ith in
the scope o f the organizations" cu rrent strategy and b u ild on existing organizational
learning (i.e. co n tin u ity). In contrast, the au to n o m o u s process concerns initiatives
that em erge o u tsid e o f the organ izatio n and p ro vid e the o p p o rtu n ities fo r new
organizational learn in g (i.e. change). These tw in processes are considered vital fo r
successful o rgan izatio n al tran sfo rm atio n . In a sim ilar vein, B ro w n an d Eisen hardt
(1997) note that co n tin u o u s o rgan izatio n al change for rap id p ro d u ct in n o vatio n is
b eco m in g a cru cial cap ab ility fo r firm s o peratin g in h igh -velo city indu stries w ith
short p ro d u ct cycles. B ased o n detailed case studies o f m u lti-p ro d u ct in n o vation s in
six firm s in the co m p u ter indu stry, the au thors conclude that co n tin u o u s change
and p ro d u ct in n o va tio n s are su p p o rted b y organ izatio n al structures that can be
described as “ sem i-structures,"" a co m b in atio n o f “ m echanistic"" an d “ organic""
features, that b alan ce o rd er an d chaos. M o re notably, the au thors id en tify “ links in
time"" that force sim u ltan eo u s attention an d linkages am o n g past, present an d fu ture
projects as essential to change processes. The key argument is that links in time create
the d irection , co n tin u ity, an d tem po o f change to su p p o rt fast pace ad ap tation in an
.uncertain an d vo latile en viron m en t.
П8 ALICE LAM

M ost strategic adaptation theories assum e that organ izatio n al adaptation can
occur through increm ental and frequent shifts, and that n ew organizational forms
and discontinuous transform ation can be b ro u g h t ab ou t b y such processes. This
strand o f research highlights the im portance o f firm -level adap tation and internal
organizational processes in the creation o f new organ izatio n al form s. Once again,
however, m ost studies o f strategic adaptation present reti ospective studies o f suc­
cessful organizational adaptation. T h ey tend to focus o n organizational restructur­
ing and transform ation w ithin prevailing o rgan ization al fo rm s and are not
specifically concerned with the creation o f new organ izatio n al form s (Lewin and
Volberda 1999). We rem ain in need o f a th eo ry to account for h o w and under what
conditions m anagerial action and organizational learn in g is connected to the emer­
gence o f new organizational form s.

5.5 C o n c l u s io n

The relationship between organization and in n o vation is com plex, dynamic, and
m u ltilevel The existing literature is volu m in o u s and diverse. T h is chapter has
sought to understand the nature o f the relationship fro m three different but inter­
dependent perspectives: (a) the relationship betw een organizational structural
form s and innovativeness; (b) innovation as a process o f organizational learning
and knowledge creation; and (c) organizational capacity for change and adaptation.
A lthough there are potentially im portant overlaps and interconnections between
these different aspects o f the relationships, the different strands o f research have
rem ained separate and there is no single coherent conceptual fram ew ork for under­
standing the phenom enon of “ organizational innovation.” T h is is partly due to the
great conceptual am biguity and confusion su rro u n d in g the term “ organizational
innovation. O ur review o f the existing literature reveals no consensus definition of
the term organizational innovation ” D ifferent researchers have used the term to
describe different aspects o f the relationships betw een o rgan ization and innovation.
Indeed, the concept has been used in a rather loose and slipp ery m anner in many
writings and som e authors are coy about stating definitions. Perhaps this conceptual
indeterm inacy reflects the fact that “ organizational in n o va tio n ” embraces a very
wide range o f phenom ena. M uch work rem ains to be done i f we are to understand
how the different dim ensions fit together.
I his large literature has advanced our understandin g o f the effects o f organiza­
tional structure on the ability of organizations to learn, create knowledge, and
generate technological innovation. We know relatively less, however, about how
internal organizational dynam ics and actor learning interact w ith technological and
 ORGANIZATIONAL INNOVATION 139

environm ental forces to shape o rgan ization al evolution. It rem ains unclear how and
under w hat co n d itio n s organ ization s shift fro m one structu ral archetype to another,
and the role o f tech nological in n o vation in d rivin g the process o f organizational
change is also obscure. Progress in these areas will require greater efforts to b rid ge the
different levels o f analysis and m u ltid iscip lin ary research to add insight and depth
beyond one n arro w perspective.
At present, research on organizational change and adaptation is fragmented: the
different levels o f analysis are disconnected and often rooted in different theoretical
paradigm s that use different research m ethods. T h us, ecology and evo lu tio n ary
theorists have sough t to u nd erstand the dynam ic relationship between in n o vation
and o rgan ization al evo lu tio n at the p o p u latio n o r in d u stry levels using retrospective
historical data, w hile o rgan izatio n al and m anagem en t researchers tend to exam ine
the process o f ad ap tation at the level o f in d ivid u al organizations, m o stly based on
cross-sectional case studies. The fo rm er is rooted in a structuralist determ inistic
paradigm w hereas the latter takes into account actor choice and intentionality. T h e
d isconnection betw een these tw o different levels o f analysis has m eant that w e
continue to treat selection and ad ap tation as tw o separate processes in o rgan iza­
tional evolution , w hereas in reality new fo rm s o f o rgan ization em erge from the
dynam ic in teractio n betw een the tw o processes (Lew in and V olberda 1999), The
biggest challenge fo r researchers is to bridge the w id e g u lf betw een ecology/evolu-
tionary theories (dealing with organizational evolution and external forces o f
change) and strategic ch oice and learning theories (focu sin g on actor choice,
interpretation, and gro u p d yn am ics w ith in organizations). A useful avenue for
future research w o u ld con sider h o w o rgan ization al choice and evo lu tio n ary p ro ­
cesses interact to facilitate o rgan izatio n al change and inn ovation . T h is w ill require
lon gitud in al research on organ izatio n al ad ap tation in “ real tim e,” as distinct fro m
retrospective h isto rical case studies (Lew in et al. 1999).
A n oth er facto r that inh ibits m ajo r theoretical progress in the field is the failure o f
researchers in the fields o f in n o vatio n and o rgan ization al studies to w o rk m ore
closely together. A lth o u g h in n o vatio n scholars have lo n g recognized the im portan ce
o f the o rgan izatio n al d im en sio n o f in n o vation , m an y in n o vatio n studies continue to
be d om in ated b y an eco n o m ic app roach that allow s little ro o m for the analysis o f
creative change and in n o va tio n w ith in the o rgan ization itself. B y contrast, research­
ers in the field o f o rgan izatio n al studies w h o have developed a rich literature on
organizational co gn itio n , learn in g, an d creativity rarely relate their w o rk exp licitly
to in n o vation , A s a result, this stream o f w o rk w h ich offers great poten tial fo r
un d erstan d in g the m icro -d y n a m ics o f o rgan ization al change and in n o vation
rem ains o u tsid e the m a in arena o f in n o va tio n studies. T h e b u lk o f the existing
research on the relatio n sh ip betw een o rgan ization and in n o vatio n con tin ues to
focus o n h o w tech n o lo g y an d m arket forces shape o rgan ization al ou tcom es and
treat o rgan izatio n s primarily as a vehicle or facilitator o f innovation, rather than as
innovation itself. For example, we tend to assume that technological innovation
140 ALIC E LAM

triggers o rg a n iz a tio n a l change because it shifts the com petitive environment and
forces organizations to adapt to the new set o f dem ands* T h is deterministic view
neglects the possibility that differences in organ izatio n al interpretations of, and
responses to, external stim uli can affect the ou tcom es o f organizational change. The
literature in organizational cognition argues that the en viron m en t is equivocal and
changes in the environm ent creates am b ig u ity and u n certain ty w hich prompts the
organization to em bark on a cycle o f environ m en tal scann in g, interpretation, and
learning (Daft and W eick 1984; G reve and Taylor 2000). T h e scanning and search
process m ay lead to new interpretative schem ata and o rgan ization al action which
could be an im portant source o f innovative organ izatio n al change. Treating the
organization as an interpretation and learning system directs o u r attention to the
im portant role o f internal organizational d ynam ics, actor cogn ition , and behavior in
shaping the external environm ent and outcom es o f o rgan izatio n al change.
Another prom ising direction for future research recognizes that organizational
innovation m ay be a necessary precon dition fo r tech n ological innovation, rather
than treating this process u niform ly as a response to extern al forces, and focuses on
the processes o f internal organizational reform and tran sfo rm ation that are neces­
sary to create such preconditions. This requires that scholars take greater account of
the role o f endogenous organizational forces such as cap acity for learning, values,
interests, and pow er in shaping organizational evo lu tio n and technological change,
This is an area where organization and m anagem ent researchers could make a
significant contribution by placing a greater em phasis o n rigo rou s empirical re­
search and theory building.

N otes

1. The term organizational innovation” is ambiguous. Some authors use it to refer to the
broad meaning of innovation or innovative behaviour in organizations” (Slappendel
1996; Sorensen and Stuart 2000), or “ organizational adoption of innovations” (Kimberley
and Evanisko 1981; Damanpour and Evan 1984; Damanpour 1996). Within these broad
meanings, the dependent variable “ innovation” is defined to encompass a range of types,
including new products or process technologies, new organizational arrangements or
administrative systems. Ihe main aim of these studies has been to identify a range of
individual, organizational, and environmental variables that affect an organization's
propensity to adopt an innovation. Others (e.g. Pettigrew and Fenton 2000) use the
term in a more restrictive way simply to refer to innovation in organizational arrange­
ments. Here the dependent variable is new organizational practices or organizational
forms. Innovation may refer to the widespread adoption by organizational population of
an organizational innovation, or merely some novel combination of organizational
piocesses or structures not previously associated. There is a tendency for authors in this
camp to equate organizational innovation to organizational change or development,
 ORGANIZATIONAL INNOVATION 141

assuming that change in itself is necessarily innovative, without making an explicit link
between organizational change and technological innovation,
2. For a detailed analysis o f the interaction between institutions and organizations in
innovation systems, see Edquist and Johnson (1992) and Hollingsworth (2000).
3. Institutional theorists accept that radical, innovative change would be possible in newly
emerging sectors (e.g. biotechology) where the organizational fields are “ informed” and
there is no stipulated template for organizing (Greenwood and Hinings 1996).
4. According to Tushman and O'Reilly (1996; 1999), ambidextrous organizations are ones
that can sustain their competitive advantage by operating in multiple modes simultan­
eously— managing for short-term efficiency by emphasizing stability and control, and for
long-term innovation by taking risks. Organizations that operate in this way develop
multiple, internally inconsistent architectures, competences, and cultures, with built-in
capabilities for efficiency, consistency, and reliability on the one hand, and experimen­
tation and improvisation on the other. During periods o f incremental change, organiza­
tions require units with relatively formalized roles, responsibilities, functional structures,
and efficiency-oriented cultures that emphasize teamwork and continuous improvement.
By contrast, during periods o f ferment— times that can generate architectural and
discontinuous innovation— organizations require entrepreneurial “ skunkworks” types
o f units. These units are relatively small, have loose decentralized product structures,
experimental cultures, loose work processes, strong entrepreneurial and technical com­
petences. Examples o f companies that have successfully developed ambidextrous organ­
izations include Hewlett-Packard, Johnson and Johnson, and ABB (Asea Brown Boveri),
as well as such large Japanese companies as Canon and Honda.

R eferen c es

A l d r ic h , H. E., and M u e l l e r , S. (1982), “ The Evolution o f Organizational Forms: Technol­

ogy, Coordination and Control ” in В. M. Staw and L. L. Cummings (eds.), Research in
Organizational Behaviour, Greenwich, Conn.: JAI Press, 4:33-87.
A m a b il e , T M. (1988), “A Model o f Creativity and Innovation in Organizations,” in N. M.
Staw and L. L. Cummings (eds.). Research in Organizational Behaviour, Greenwich, Conn.:
JAI Press, 10:123-6 7.
A n d e r s o n , P., and T u s h m a n , M. L. (1990), “ Technological Discontinuities and Dominant
Designs: a Cyclical Model o f Technological Change ” Administrative Science Quarterly
35(4): 604-33.
A n g e l s , D. R (2000), “ High-Technology Agglomeration and the Labour Market: The Case o f
Silicon Valley,” in K. Martin (ed.), Understanding Silicon Valley: The Anatomy of an
Entrepreneurial Region, Stanford: Stanford University Press, 125-89.
A o ki , M. (1988), Information, Incentives and Bargaining in the Japanese Economy Cambridge:
Cambridge University Press.
*A rgyris , C , and Schon , D. (1978), Organizational Learning: A Theory of Action Perspective,
Reading, Mass,: Addison-Wesley.

* Asterisked items are suggestions for further reading.

14 2 A L IC E LAM

B a h r a m i, H., and E v a n s , S. (2000), “Flexible Recycling and High-Technology Entrepre­

neurship,” in K. Martin (ed.), Understanding Silicon Valley: The Anatomy o f an Entrepre­
neurial Region, Stanford: Stanford University Press, 166-89.
B a l d r i d g e , J. V., and B u r n h a m , R. A. (1975). “Organizational Innovation: Individual, Organ­
izational, and Environmental Impacts,” Administrative Science Quarterly 20(2): 165-76.
B l a u , P. M. (1970), “A Formal Theory of Differentiation in Organizations,” American
Sociological Review 35(2): 201-18*
*B rown , J. S.> and D u gu id , P. (1991), “Organizational Learning and Communities of
Practice: Towards a Unified View o f Working, Learning and Innovation” Organization
Science 2(1): 40-57.
____ ____ (1998), “Organizing Knowledge” California Management Review 40(3): 90-111,
B rown , S. L., and E isen h ard t , K. M. (1997). "The Art o f Continuous Change: Complexity
Theory and Time-Paced Evolution in Relentlessly Shifting Organizations,” Administrative
Science Quarterly 42(1): 1-34.
B urglem an , R, A. (1983), “A Model o f the Interaction o f Strategic Behaviour, Corporate
Context, and the Concept of Strategy,” Academy of Management Review 8(1): 61-70.
------ (1991), “Intraorganizational Ecology of Strategy Making and Organizational Adapta­
tion: Theory and Research,” Organization Science, 2(3): 239-62.
*B urns , T., and S ta lk er , G. M. (1961), The Management of Innovation, London: Tavistock.
C a sp e r , S. (2000), “ Institutional Adaptiveness, Technology Policy and the Diffusion of
New Business Models: The Case of German Biotechnology,” Organization Studies 21:
887-914.
C h andler , A. D. (1962), Strategy and Structure: Chapters in the History of the American
Industrial Enterprise, Cambridge, Mass.: MIT Press.
C h ild , J. (1972), “ Organizational Structure, Environment and Performance —-the Role of
Strategic Choice,” Sociology 6(1): 1-22.
------ (i997)i “Strategic Choice in the Analysis o f Action, Structure, Organizations and
Environment: Retrospect and Prospect,” Organization Studies 18(1): 43-76.
------and S m ith , C, (1987), “ The Context and Process o f Organizational Transformation—
Cadbury Limited in its Sector,” Journal of Management Studies 24: 565—93.
C ohen , W. M., and L evin th al , D. A. (1990), “Absorptive Capacity: A New Perspective on
Learning and Innovation,” Administrative Science Quarterly 35:123—38.
D a ft , R. L. (1978), ‘ A Dual-Core Model of Organizational Innovation,” Academy of Man­
agement Review 21:193-210.
and L ew in , A. (1993), Where Are the Theories for New Organizational Forms? An
Editorial Essay,” Organization Science 4(4): i-vi.
and W e ic k , K. E. (1984), “ Toward a Model o f Organizations as Interpretation
Systems,” The Academy of Management Review 9(2): 284-95.
D am anpour , F. (1996), Organizational Complexity and Innovation: Developing and
Testing Multiple Contingency Models,” Management Science 42(5): 693—716.
and E v a n , W. M. (1984), Organizational Innovation and Performance: The Problem of
Organizational Lag, Administrative Science Quarterly 29: 392—402.
D e F il l ip i , R. (2002), Organization Models for Collaboration in the New Economy,”
Human Resource Planning 25(4): 7-19.
and A rth ur , M. B. (1996), “ Boundarlyess Contexts and Careers: A Competency-Based
Perspective, in M. B. Arthur and D. M. Rousseau (eds.). The Boundaryless Career: A New
 ORGANIZATIONAL INNOVATION 143

Employment Principle for a New Organizational Era, New York; Oxford University Press,
116-31.
D iM ag g io , P J., and P ow ell , W. W. (1983), “ The Iron Cage Revisited: Institutional Iso­
morphism and Collective Rationality in Organizational Fields” American Sociological
Review 48; 147-60,
Dosi, G. (1988), Sources, Procedures, and Microeconomic Effects o f Innovation,” journal of
Economic Literature 2 6 :1120—71.
E dquist , C., and J ohnson , B. (1997), “ Institutions and Organizations in Systems o f Innov­
ation ” in C. Edquist (ed.), Systems of Innovation: TechnologiesyInstitutions and Organiza-
tionSy London: Pinter, 41-63.
F iol , С. M. (1993), “ Consensus, Diversity, and Learning in Organizations,” Organization
Science 5: 403-20,
Foss, N. J. (2003), “ Selective Intervention and Internal Hybrids: Interpreting and Learning
from the Rise and Decline o f the Oticon Spaghetti Organization,” Organization Science
14(3): 331- 4 9 .
G a lu n ic , D. C., and E ise n h a r d t , K. M. (2001), “Architectural Innovation and Modular
Corporate Forms,” Academy of Management Journal 44(6): 1229-49.
G e r sic k , C. J. G. (1991), “ Revolutionary Change Theories: A Multilevel Exploration o f the
Punctuated Paradigm,” Academy of Management Review 16(1): 10-36.
G l y n n , M. A. (1996), “ Innovative Genius: A Framework for Relating Individual and Organ­
izational Intelligence to Innovation,” Academy of Management Review 21(4): 1081-111.
G ra n t , R. M. (1996), “ Toward a Knowledge-Based Theory o f the Firm,” Strategic Manage­
ment Journal 17:109 -22.
^G reenw ood , R., and H in in g s , C R. (1996), “ Understanding Radical Organizational
Change: Bringing Together the Old and New Institutionalism,” Academy of Management
Review 21(4): 1022-54.
G r e v e , H. R., and T a y l o r , A. (2000), “ Innovations as Catalysts for Organizational Change:
Shifts in Organizational Cognition and Change,” Administrative Science Quarterly 45:
54-80.
H a l l , P., and S o s k i c e , D. (eds.) (2001), Varieties of Capitalism: The Institutional Foundations
of Comparative AdvantageyOxford: Oxford University Press.
H a n n a n , M. X , and F r e e m a n , J. H. (1977). “ The Population Ecology o f Organizations,”
American Journal of Sociology 82(5): 929-63,
*------ — (1984), “ Structural Inertia and Organizational Change,” American Sociological
Review 49(2): 149-64.
H e d l u n d , G. (1994), “A Model o f Knowledge Management and The N-Form Corporation,”
Strategic Management Journal 15: 73-90.
H en derso n , R. M., and C l a r k , R. B. (1990), “Architectural Innovation: The Reconfigura­
tion o f Existing Product Technologies and the Failure o f Established Firms,” Adminis­
trative Science Quarterly 29: 26-42.
H in in g s , C R., T h i b a u l t , L., S l a c k , X , and K ik u l is , L. M. (1996), “ Values and Organiza­
tional S t r u c t u r e Human Relations 49(7): 885-916.
H o d g k i n s o n , G. P. (2003), “ The Interface o f Cognitive and Industrial, Work and Organiza­
tional Psychology” Journal o f Occupational and Organizational Psychology 76(1): 1—24.
H o l l i n g s w o r t h , J. R, (2000), “ Doing Institutional Analysis: Implications for the Study o f
Innovations,” Review o f International Political Economy 7(4): 595—644.
14 4 ALICE LAM

Kanter, R. M. (1983), The Change Masters, New York: Simon & Schuster.
K i m b e r l y . J. R„ and E v a n i s k o , M. J. (1981), “Organizational Innovation: The Influence
of Individual, Organizational, and Contextual Factors on Hospital Adoption of
Technological and Administrative Innovations,” The Academy o f Management Journal
24(4): 689-713.
Kodama, M. (2003), “Strategic Innovation in Traditional Big Business: Case Studies ofTtoo
Japanese Companies,” Organization Studies 24(2): 235-68.
K o g u t , B „ and Z a n d e r , U. (1992), “Knowledge of the Firm, Combinative Capabilities, and
the Replication of Technology,” Organization Science 3(3): 383-97.
L a m , A. (1996), “Engineers, Management and Work Organization: a Comparative Analysis of
Engineers’ Work Roles in British and Japanese Electronics Firms,” Journal of Management
Studies 33(2): 183-212.
------(1997)1 “ Embedded Firms, Embedded Knowledge: Problems o f Collaboration and
Knowledge Transfer in Global Cooperative Ventures,” Organization Studies 18(6): 973-96.
*------ (2000), “ Tacit Knowledge, Organizational Learning, Societal Institutions: an Inte­
grated Framework,” Organization Studies 21(3): 487-513.
------ (2002), “Alternative Societal Models o f Learning and Innovation in the Knowledge
Economy,” International Social Science Journal 17(1): 67-82.
L a v e , J., and W e n g e r , E. (1991), Situated Learning: Legitimate Peripheral Participation. New
York: Cambridge University Press.
Law ren ce, P. R., and L o r s c h , J. W. (1967), “Differentiation and Integration in Complex
Organizations,” Administrative Science Quarterly 12:1-47.
L a z o n i c k , W., and W e s t , J. (1998), “Organizational Integration and Competitive Advan­
tage,” in G. Dosi et al. (eds.). Technology, Organization, and Competitiveness, Oxford:
Oxford University Press.
L e o n a r d - B a r t o n , D. (1992), “Core Capabilities and Core Rigidities: A Paradox in Man­
aging New Product Development,” Strategic Management Journal 13:111-25.
L e v i n t h a l , D. A., and M a r c h , J. G. (1993), “The Myopia of Learning,” Strategic Manage­
ment Journal 14: 95-112.
* L e w i n , A. Y., and V o l b e r d a , H. W. (1999), “Prolegomena on Coevolution: a Framework
for Research on Strategy and New Organizational Forms,” Organization Science 10(5):
519- 3 4 *
L e w i n , A. Y, Long, С. P, and C a r r o l l , T. N. (1999), “The Co-evolution of New Organiza­
tional Forms,” Organization Science 10: 535— 50.
L u n d v a l l , B.-A. (ed.) (1992), National Systems of Innovation: Towards a Theory ofInnovation
and Interactive Learning, London: Pinter.
M a r c h , J. G. (1991), Exploration and Exploitation in Organizational Learning,” Organiza­
tion Science 2: 71— 87.
M ezia s , S. J., and G lynn , M. A. (1993), “The Three Faces of Corporate Renewal: Institution,
Revolution, and Evolution,” Strategic Management Journal 14: 77-101.
M i l e s , R. E„ S n o w , C. C„ M a t h e w s , J. A., M i l e s , G„ and C o l e m a n , H. J. Jr. (1997).
“Organizing in the Knowledge Age: Anticipating the Cellular Form,” Academy of Manage-
ment Executive 11(4): 7-20.
D„ and F r i e s e n , P. H. (1982), “ Structural Change and Performance: Quantum
M il l e r ,
versus Piecemeal-Incremental Approaches,” Academy of Management Journal 25(4):

M intzberg, H. (1979), The Structuring of Organization, Englewood Cliffs, NJ: Prentice Hall.
 ORGANIZATIONAL INNOVATION Ц5

N elso n , R. R., and W in t e r , S. G. (1982), An Evolutionary Theory of Economic Change,

Cambridge, Mass.: Belknap Press.
N on a k a , I. (1994), “A Dynamic Theory o f Organizational Knowledge Creation/' Organiza­
tion Science 5:14-37.
*------ &nd T a k e u c h i , H, (1995), The Knowledge Creating Company; New York: Oxford
University Press.
Pa v it t , K. (1991)1 “ Key Characteristics o f the Large Innovating Firm/' British Journal of
Management 2: 41-50.
P errow , C. (1970)1 Organizational Analysis, London: Tavistock.
P et t ig r ew , A, M., and F en to n , E. M. (eds.) (2000), The Innovating Organization, London:
Sage Publications.
Po la n yi , M. (1966), The Tacit Dimension, New York: Anchor Day Books.
P ow ell , W. W., and DiMaggio, P. J. (eds.) (1991), The New Institutionalism in Organizational
Analysis, Chicago: University o f Chicago Press.
P ra h a la d , С K., and H a m el , G. (1990), “ The Core Competence o f the Corporation/'
Harvard Business Review (May/June): 79-91.
P ugh , D. S,, H ic k so n , D. J., and H in in g s , C. R. (1969), “ The Context o f Organization
Structures/' Administrative Science Quarterly 14: 47-61.
Q u in n , J. B. (1992), Intelligent Enterprise: A Knowledge and Service Based Paradigm for
IndustryуNew York: The Free Press.
R o m an elli , E., and T u sh m a n , M. L. (1994), “ Organizational Transformation as Punctuated
Equilibrium: An Empirical Test/' Academy of Management Journal 37(5): 1141-66.
Sa x e n ia n , A. (1996), “ Beyond Boundaries: Open Labour Markets and Learning in the
Silicon Valley," in M. B. Arthur and D. M. Rousseau (eds.), The Boundaryless Career: A
New Employment Principle for a New Organizational Era, New York: Oxford University
Press, 23-39.
Sc h u m p e t e r , J. (1950), “ The Process o f Creative Destruction," in J. Schumpeter (ed.).
Capitalism, Socialism and Democracy; 3rd edn., London: Alien and Unwin.
S en g e , R (1990), The Fifth Discipline: the Art and Practice of the Learning Organization, New
York: Doubleday.
S h r iv a st a v a , R, and S c h n e id e r , S. (1984), “ Organizational Frame o f Reference," Human
Relations 37(10): 795-809.
------ M itr o ef , I., and A l v esso n , M. (1987), “ Nonrationality in Organizational Actions,"
International Studies of Management and Organization 17: 90-109.
S im on , Н. A. (1991), “ Bounded Rationality and Organizational Learning," Organization
Science 2:125-34.
S la p p e n d el , C. (1996), “ Perspective on Innovation in Organizations/' Organization Studies
17(1): 107-29.
S o ren sen , J. B., and S t u a r t , T. E. (2000), “Age, Obsolescence, and Organizational Innov­
ation," Administrative Science Quarterly 45(1): 81-112.
Soskice, D. (1999), “ Divergent Production Regimes: Coordinated and Uncoordinated
Market Economies in the 1980s and 1990s," in H. Kitschelt, P. Lange, G. Marks, and
J. Stephens (eds.). Continuity and Change in Contemporary Capitalism, Cambridge: Cam ­
bridge University Press, 101-34.
St a r b u c k , W. Н» (1992), “Learning by Knowledge-Intensive Firms," Journal of Management
Studies 29(6): 713-40.
146 A L I C E LAM

T e e c e , D.J. (1998), “Design Issues for Innovative Firms: Bureaucracy, Incentives and Indus­
trial'Structure,” in A. D. Chandler Jr., P. Hagstrom, and O. Solvell (eds.). The Dynamic
Firm, Oxford: Oxford University Press, 134-65.
_____ and P i s a n o , G. (1994), “The Dynamic Capabilities of Firms: an Introduction,” Indus­

trial and Corporate Change 3(3): 537— 56-

T id d , J., B e s s a n t , J., and Р а ѵ іт т , K. (1997). Managing Innovation, Chichester: John Wiley
& Sons.
* T u s h m a n , M . L , and A n d e r s o n , P. (1986), "Technological D iscontinuities and Organiza­
tional Environm ents,” Administrative Science Quarterly 31(3)* 439—65.
____ and N elson , R. R. (1990), "Introduction: Technology, O rganizations and Innovation”
Administrative Science Quarterly 35(1): 1_8.
------ and O 'R e il l y , C. A. I ll (1996), "A m bidextrous O rganizations: M anaging Evolutionary
and Revolutionary Change ,” California Management Review 38(4): 8-30.
------------ (1999), “ Building Ambidextrous Organizations: Forming Your Own 'Skunk
W orks'” Health Forum Journal 42(2): 20 -3,
------ N ew m an , W. H., and R o m anelli , E. (1986), "C on vergen ce and Upheaval: Managing
the Unsteady Pace o f O rganizational Evolution,” California Management Review 29(1):
29-44.
V an de Y en , A ., P olley , D., G aru d , S., and V e n k a t a r a m a n , S. (1999), The Innovation
Journey New York: O xford University Press.
V erona , G ., and R a v a si , D. (2003), "U n b u n d lin g D ynam ic Capabilities: an Exploratory
Study o f Continuous Product Innovation,” Industrial and Corporate Change 12(3):
577-606.
V ir a n y , B., T ushm an , M . L , and R o m anelli , E. (1992), "E xecu tive Succession and Organ­
izational Outcom es in Turbulent Environm ents: A n O rganization Learning Approach,”
Organization Science 3: 72-91.
V on H ip p el , E. (1988), The Sources of Innovation N ew York: O xfo rd U niversity Press.
W a lsh , J. P. (1995), "M anagerial and O rganizational C ogn ition : N otes From a Trip Down
M em ory Lane,” Organization Science 6(3): 280-321.
------ and U ngson , G. R. (1991), "O rganizational M em ory,” Academy of Management Review
16: 57-91.
W e b e r , M. (1947), The Theory of Social and Economic Organization Glencoe, 111.: The Free
Press.
W e ic k , K. E. (1979), 7 he Social Psychology of Organizings 2nd edn., R eading, Mass.: Addison-
Wesley.
------ (1995), Sensemaking in Organizations, Thousand Oaks, Calif.: Sage.
(1996), The Role of Renewal in Organizational Learning,” International Journal of
Technology Management 11(7—8): 738-46.
W en ger , E. (1998), Communities of Practice: Learnings Meanings and Identity, New York:
Cambridge University Press.
W h it le y , R. (2000), The Institutional Structuring o f Inn ovation Strategies: Business
Systems, Firm Types and Patterns o f Technical Change in D ifferent M arket Economies,”
Organization Studies 21(5): 855-86.
(2003), The Institutional Structuring o f O rganizational Capabilities: the Role of
A uthority Sharing and O rganizational Careers,” Organization Studies 24(5): 667-95*
Wo lfe , B. (1994)) Organizational Innovation: Review, C ritique and Suggested Research
Directions,” Journal of Management Studies 31: 405-31.
 ORGANIZATIONAL INNOVATION 147

Wom ack , J. P., J o n e s , D. T „ and R o o s , D. (1990), The Machine that Changed the World, N ew
York: Rawson Associates,
Woodman , R. W., S a w y e r , J. E., an d G r i f f i n , R, W. (1993), "'Toward a T h eo ry o f O rganiza­
tional Creativity/5 Academy of Management Review 18(2): 293-321,
Woodward , J. (1965), Industrial Organization, Theory and Practice, London: O xford U n i­
versity Press.
Z u cker , L. G, (19B7), “ In stitution al T heories o f O rganizations,55 Annual Review of Sociology
13 :443-64.
 C H A P T E R 6

MEASURING
INNOVATION

KEITH SMITH

6.1 In t r o d u c t io n 1

It is som etim es suggested that in n ovation is in h eren tly im p o ssib le to quantify and
о m easure. This chapter argues that w hile this is tru e fo r so m e aspects o f innovation,
its overall characteristics do not preclude m easu rem en t o f key dimensions of
processes and outputs. A n im portan t d evelopm ent h as been the emergence o f new
ti , rS ° mi*ovat*on in P uts and outputs, in clu d in g econ o m y-w id e measures
e som e egree o f international com parability. F o llo w in g sections discuss
: .C *ss^ies *n construction and use o f science, technology, and
Г г ,™ , H- У m dlcators' the" tu ™ (briefly) to the strengths and weaknesses of
t- fn r • 1Ca ° Г k Part ctdarty and patents. Final sections cover recent initia-
innovation ^ ° n * 6 conceP tua^ zat *on, collection, and analysis o f direct measures of

A r m ™ l thr “traditi° nal” indicat° r s are emphasized here because, as Kenneth

ow remarked many years ago, “too much energy has gone into squeezing the last
П » ^ f" d № ™ P” p o L г о Г ѵ Г Г І ^ of
data is collected” ( А т о Т іХ У У 'с Г т h СпІ'СЬ<!3: " far ,0° li,,le fresh economici
have responded to this kind of challen ‘ ‘^ Г ^ * 4 ) ' Innovatlon data P roducerS
new survey-based
У
indicators e , • %n ^ m° St tmPortant development has been
cl indicators, especially the Community Innovation Survey (CIS),
 MEASURING INNOVATION

which has been carried o u t three tim es in all EU M em ber States. The basic form at o f
CIS has diffused to m an y o th er cou n tries (in clu d in g C anad a, A ustralia, H ungary,
Brazil, A rgentina, and C h in a ). H as this effort been justified? In answ ering this
much depends on the q u ality o f analysis these surveys m ake possible, so the final
section discusses the rap id ly grow in g research and pu blication efforts deriving
from CIS.

6.2 T h e C o n c e p t u a l B a c k g r o u n d :
M ea su r em en t Issues

Measurement im plies co m m en su rab ility : that there is at least som e level on w hich
entities are qu alitatively sim ilar, so that co m p arison s can be m ade in quantitative
terms.
An im m ediate p ro b lem is that in n o vatio n is, b y definition, novelty. It is the
creation o f so m eth in g q u alitatively new, v ia processes o f learning and know ledge
building. It involves ch an g in g com petences and capabilities, and p ro d u cin g q u ali­
tatively new p erfo rm an ce o u tco m es. T h is m ay lead to new prod u ct characteristics
that are in trin sically m easu rab le in so m e w a y — new lift/drag aspects o f an aircraft
wing, for exam ple, o r im p ro ve d fu el efficien cy o f an engine. H ow ever, such technical
measurement co m p ariso n s are o n ly rarely m ean in gfu l across products. M ore
generally, in n o vatio n in volves m u ltid im en sio n al n ovelty in aspects o f learning
or knowledge o rg an izatio n that are d ifficu lt to m easure or intrinsically n on-
measurable. K ey p ro b lem s in in n o vatio n in dicators therefore concern the und er­
lying conceptualization o f the object being m easured, the m eaning o f the
measurement concept, an d the general feasib ility o f d ifferent types o f m easurem ent.
Problems o f c o m m en su ra b ility are n o t necessarily insoluble, b u t a m ain point
arising from recent w o rk is the need fo r care in distin gu ish in g betw een w hat can
and what cannot be m easu red in in n o vatio n .
Quite apart from the problem o f whether novelty can be m easured, a fundamental
definitional issue is w h at w e actu ally m ean b y “ new” (see Ch. i b y Fagerberg in this
volume). D oes an in n o v a tio n h ave to co n tain a basic n ew p rin cip le that has never
been used in the w o rld b efo re, o r d oes it o n ly need to be new to a firm ? D oes an
innovation have to incorporate a radically novel idea, or only an incremental
change? In general, what kinds o f novelty count as an innovation? These issues o f
commensurability and novelty are basic p ro b lem s for all S & T indicators R & D in
particular—-but have been most exp licitly addressed in the development of direct
innovation indicators.
150 KEITH SMITH

6 ч T h e o r i e s of I n n o v a t i o n a n d t h e i r
U se , n I n d i c a t o r D e v e l o p m e n t

A lthough statistics are often treated as though their m eanings are transparent, they
always rest on som e kind o f (usually im plicit) conceptual fou n d ation s. The system of
national accounts, for exam ple, derives from K eynesian m acroecon om ic concepts
that seek to identify com ponents o f aggregate dem and . R & D data has a complex
background in the scientification o f in n o vation — the n o tio n that acts o f research
and discovery underpin innovation (Laestadius 2003). These conceptual founda­
tions are rarely considered when indicators are used. Such issues are complicated by
the fact that som e key S8cT indicators are b y-p rod u cts o f other processes— legal
procedures (as with patents), or academ ic institution s (as w ith bibliom etrics, which
rest on publishing conventions).
W hat kinds o f ideas have form ed the conceptual fo u n d atio n s o f innovation
indicators? A n im portant figure here has been N ath an R osenberg, whose work
quite explicitly affected the O E C D s Innovation M anual (O E C D 1992,1997). (This
m anual is usually called the Oslo M anual because m u ch o f the drafting and expert
meetings on it occurred there.) First, R osenberg challenged the notion o f research-
based discovery as a prelim inary phase o f in n ovation . Second, he challenged the
idea o f separability between innovation and d iffu sio n processes, pointing out
that m ost diffusion processes involve long and cu m u lative program s o f post­
com m ercialization im provem ents (see R osenberg 1976 and 1982). Perhaps his
best-know n contribution, w ith Steven Kline, has been the so-called chain-link
m odel o f innovation, which stresses three basic aspects o f in n ovation (Kline and
Rosenberg 1986):

* innovation is not a sequential (linear) process but one in vo lvin g m any interactions
and feedbacks in knowledge creation
* innovation is a learning process involving m ultiple inputs
* innovation does not depend on invention processes (in the sense o f discovery of
new principles), and such processes (in volving fo rm al R8cD) tend to be under­
taken as problem -solving w ithin an on goin g in n o vatio n process rather than an
initiating factor

The w ork o f Rosenberg alone, and o f R osenberg and Kline, has at least two
im portant im plications for indicator developm ent. The first is that novelty implies
not just the creation o f com pletely new products o r processes, but relatively small-
scale changes in product perform ance which m ay— over a long p eriod — have major
technological and econom ic im plications. A m eaningful in n ovation indicator
should therefore be able to pick up such change. T h e second is the im portance of
n o n -R & D inputs to innovation— design activities, engineering developm ents and
experim entation, training, exploration o f m arkets for new products, etc. So there is a
 MEASURING INNOVATION 151

need for input in d icators that reflect this inpu t variety and its diverse d istributions
across activities.
The C IS effort has in general been in form ed b y ideas from recent innovation
research. O ne in p articu lar sh ou ld be m entioned, especially because it has had a
strong im pact on research u sing the new data. This is the idea that innovation relies
on collaboration an d interactive learning, in volvin g other enterprises, organ iza­
tions, and the science and tech n o logy infrastructure. D ata gatherers have been
concerned to exp lore the n etw orkin g d im en sio n o f in n ovation , and this has been
an im portant con cep tu al issue in su rvey design (see H ow ells 2000, for an o verview o f
research on this top ic).

6.3.1 Existing and New Indicators: What Can Be Measured,

and What are the Limitations?
What does it m ean to m easure qualitatively diverse phenom ena? C learly this is a
serious problem fo r R & D data. Research is a know ledge-creating process for which
both activities and ou tcom es are radically incom m ensurable— there is no m eaningful
way to assess the d issim ilar actions and events that feed into research, let alone to
compare the increm ents to know ledge that follow from research. This problem cannot
be overcom e— it can o n ly be circum vented b y carefully specifying aspects o f the
research process that are in som e serious sense m easurable. T h e solu tion adopted
by the fram ers o f the Frascati M anual (the O E C D ’s operatin g statistical m anual
for R & D data collection) has been to w rite d efinitions o f research -com prising
activities, an d then seek data on either exp en ditu re o r personnel resources devoted
to such activities. T h e m easu rem en t concept fo r R & D is therefore econom ic in
character, and the datasets that result are collections o f econom ic indicators
com patible w ith in d u strial datasets, and indeed w ith the national accounts,2
This app roach to m easu rem en t has also been taken w ith in n o vation surveys. The
problem is that in n o va tio n is u su ally conceptualized in term s o f ideas, learning, and
the creation o f k n o w led ge (m o reo ver know ledge creation o f a far w id er character
than research), o r in term s o f com petences and capabilities. A s w ith “ research,”
innovation is a m u ltid im e n sio n al process, w ith nothing clearly m easurable about
m any aspects o f the u n d erly in g process. M o st m o d ern in n o vation th eo ry rests on
some kind o f “ reso u rce-b ased ” th eo ry o f the firm , in w h ich firm s create physical and
intangible assets that u n d erp in capabilities (see Lazonick in this volu m e). In novative
learning can be seen as ch an ge in the know ledge bases on w h ich capabilities rest.
Neither learning, nor the capabilities which result, seem to be measurable in any
direct way. However, just as “ research” can be captured via expen ditures on certain
activities, or b y the use o f tim e b y certain research personnel, so learning processes
can to some extent be captured by activities such as design, training, market
152 KEITH SMITH

research, tooling up, etc. Expenditure on such activities can in principle be measured
(of course the practice may be difficult, since some o f these innovation-related
activities are not straightforwardly reflected in the accounting procedures of
firms). On the output side, the question is whether capability outcomes can be
measured by some tangible change in physical or economic magnitudes. Once again
there are also potential measurement areas— experience (with pilot or experimental
surveys in the 1980s) showed that firms can identify changes in their product mixes,
and can estimate sales from new or changed products (Smith 1992)- So it is possible
to define product change, in terms o f construction, use o f materials, technical
attributes, or performance characteristics, and then to look at the place of (differ­
ently) changed products in the sales o f the firm. These considerations lead to
expenditure measures o f inputs to innovation, and sales measures o f outputs of
innovation. These economic measures o f innovation are clearly analogous to the
measurement o f research. This similarity in approach incidentally suggests that it
makes no sense to use R&D data while rejecting the use o f more direct innovation
data.

6.4 C u r r e n t M a j o r I n d i c a t o r s

T his section outlines the m ajor established in d icators that have been used for
innovation analysis, and provides a b rie f guide to fu rth er analysis o f them. There
are three broad areas o f indicator use in ST I analysis: first, R & D data; second, data
on patent applications, grants and citations; and third, b ib liom etric data (that is
data on scientific publication and citation).
In addition to this there are three other im p o rtan t classes o f indicators:
• technom etric indicators, which explore the technical perfo rm an ce characteristics
o f products (see e.g. Saviotti 1996 and 2001 for a theoretical view o f this, and Grupp
1994 and 1998 for analysis and em pirical specifications);
* synthetic indicators developed for scoreboard pu rp oses m ain ly by consultants (see
W orld Econom ic Forum 2003);
databases on specific topics developed as research tools b y individuals or groups
(such as the large firm database used b y Pavitt and Patel, o r the M ERIT-CATI
database on technological collaboration developed b y Jo h n H agedoorn, or the
D ISKO surveys on technological collaboration em anating from the University of
A lborg (see Patel and Pavitt 1997 and 1999, H aged oo rn and Schakenraad i 99°>
and— for extensi ve reporting on the use o f collab oration data— O E C D 2001).
 Box 6.1 Bibliometric data

Bibliometric analysis, meaning the analysis o f the composition and dynamics o f

scientific publication and citation, revolves around the Science Citation Index and
the Institute for Scientific Information database. The Institute for Scientific Infor­
mation (ISI) was founded in 1958, and acquired by Thomson Business Information— a
subsidiary o f the Thomson Corporation— in 1992. The ISI National Science Indicators
database currently contains publication and citation statistics from more than 170
countries, and 105 subfields in the sciences, social sciences, and arts and humanities,
representing approximately 5,500 journals In the sciences, 1,800 in the social sciences,
and 1,200 in the arts and humanities.

The fo llo w in g d iscu ssion concentrates on R & D and patents, since b ib liom etric
analysis relates p rim a rily to the d ynam ics o f science rather than in n o vation (see
M oed et al. 1995, and K alo u d is 1997 fo r review s o f the state o f the art).

6.4.1 Research and Development (R&D) Statistics

and Indicators
By far the lo n gest-stan d in g area o f data collection is R & D .
The key O E C D d o cu m en t fo r the collection o f R & D statistics is the Standard
Practice fo r Surveys o f Research a n d E xp erim en ta l Development^ better k n o w n as the
Frascati M a n u a l T h e first ed itio n was the result o f an O E C D m eeting o f national
experts on R & D statistics in Frascati, Italy, in 1963. The m an u al has been co n tin u ­
ously m o n itored an d m o d ified th ro u gh the years: the current version o f the m anual,
the Frascati M a n u a l 2002y is the seventh ed itio n (O E C D 2002). The M an u al defines
R & D as co m p risin g b o th the p ro d u ctio n o f new know ledge and new practical
applications o f know ledge: R & D is conceived as covering three different kinds o f
activities: b asic research, app lied research, and experim ental developm ent— these
categories are d istin gu ish ed in term s o f their distance fro m application.
It is often d ifficu lt to d raw the d iv id in g line between w hat sh ou ld be counted as
R & D an d w h at sh o u ld be excluded: “ T h e basic criterion fo r distin gu ish in g R & D
from related activities is the presence in R & D o f an appreciable element o f novelty
and the resolution o f scientific and/or technological uncertainty, i.e. when the
solution to a problem is not readily apparent to someone familiar with the basic
154 Keith sm ith

stock o f commonly used knowledge and techniques in the area concerned” (OECD
2002:33). Education and training in general is not counted as R & D , Market research
is excluded. There are also many other activities with a scientific and technological
base that are kept distinct from R & D . These include such industrial activities related
to innovation as acquisition o f products and licenses? product design? trial produc­
tion, training and tooling up, unless they are a component o f research, as well
as the acquisition o f equipment and machinery related to product or process
innovations.
R & D is often classified according to m ultiple criteria, an d data is collected in
highly detailed form s. Beyond the distinction betw een basic research, applied
research and developm ent the data is classified into sector o f perform ance: business
enterprise, governm ent, higher education, and p rivate n o n -p rofit. It also distin­
guishes between sources o f finance, both dom estic an d in te rn a tio n a l Then there
is classification by socio-econom ic objectives, and a fu rth er classification by fields
o f research. These detailed classifications are u su ally ignored both by policy
analysts and researchers, who tend to focus on gross expenditure only (at industry
or country level), thereby m issing m ost o f the really interesting detail in the
data. For exam ple, a m ajor issue is that, w hen lo o k in g at R & D by fields of
research, IC T (inform ation and com m u n ication s technologies) turns out to be
the largest single category in all countries that classify R & D data in this way.
H ow ever m ost o f the IC T research is actually p erfo rm ed outside the IC T sector, in
the form o f systems and software developm ent b y users.3 O n the one hand, this raises
interesting questions about the cross-in d u stry significance o f the IC T sector;
but there are also questions about the extent to w h ich such activity should be
classified as R & D at all. C oncerns have also been expressed about whether
the R & D definitions are com prehensible to firm s (especially SM E s), and whether
or not there is system atic undercounting o f sm all-firm R&D (Kleinknecht,
M ontfort, and Brouw er 2002).
R & D data is always constrained as an in n ovation in d icator b y the fact that it
measures an input only (Kleinknecht et al. 2002). H ow ever, R & D also has funda­
m ental advantages. These include the long period over w hich it has been collected,
the detailed subclassifications that are available in m any countries, and the relatively
good harm onization across countries. U n fo rtu n ately a great deal o f the literature
consists essentially o fa n attempt to m atch aggregate R & D m easures across time and
across sectors or countries to som e m easure o f p ro d u ctivity (see G riffith, Redding,
and Van Reenen (2000) for a very thorough recent exam ple; D o w rick (2003) is a
recent survey o f this very large literature). H ow ever this research effort is limited in
two senses on the one hand it tends to im ply (along w ith the new grow th theory,
incidentally) that R & D is the prim ary source o f p ro d u ctivity grow th, and on the
other it fails to exploit the basic com plexity o f the data that is actually available. The
disaggregation processes that are possible with R & D data contin ue to offer rich and
unexploited opportunities for researchers.
6.4*2. Policy Pitfalls: The Use and Misuse of R&D Indicators
It is w orth sayin g som eth in g about the pitfalls o f R & D as a policy indicator,
especially via the m ost w idely-u sed indicator, that o f “ R & D In te n sity ” T h is is the
ratio o f R & D exp en d itu re to som e m easure o f output. For a firm , it is usually the
R & D /Sales ratio. For an in d u stry o r a co u n try it is the ratio o f business expendi ture
on R & D (often k n ow n as B E R D ) to total p ro d u ctio n or value added. F o r a co u n try it
is usually gross exp en d itu re on R & D (G E R D ) to G D R
The R & D /G D P ratio is used in two prim ary ways. First, it is used to characterize
industries— high B E R D /G D P ratios for an indu stry are held to identify high-
technology activities. Second, a h igh G E R D /G D P ratio for a country is often believed
to indicate technological progressiveness and com m itm ent to knowledge creation
(see G o d in 2004 for an account o f the historical background to these notions).
Fo r countries, there is a d istrib u tion o f G E R D /G D P intensities, as Table 6.1
indicates. Both analysts and p o licy m akers often treat a particu lar place in the
ranking, o r the O E C D average, o r som e particu lar G E R D /G D P ratio as desirable
in itself. So C an ad a, for exam ple, has the objective o f raising its ran kin g to fifth in the
O E C D table; N o rw a y has the target o f reaching the O E C D average fo r G E R D /G D P ;
and the E U as a w h o le has a target o f reaching a G E R D /G D P ratio o f 3 per cent (it
could be argu ed that this target d om inates E U tech n ology po licy m aking at the
present tim e). B u t w h at is the in d icato r really telling us?
A basic p ro b lem is that R & D in ten sity depends on the indu strial m ix. C urren tly
the O E C D uses a fo u r-tier m odel to classify industries, in w hich the basic criterion is
the B E R D /P ro d u ctio n ratio:

high-tech in du stries > 5% R & D /P ro d u ctio n

m edium h igh -tech indu stries > 5% > 3% R & D /P ro d u ctio n
low -tech indu stries > 3% > 1% R & D / P rod u ctio n
low -tech indu stries > 1% > 0% R & D / P rod u ctio n

Since indu stries v a ry co n sid erab ly in their B E R D /G D P ratios, the aggregate BE R D /

G D P ratio m a y sim p ly be an effect o f that fact that indu strial structures are different
across countries. A c o u n try o r region w ith large h ig h -R & D industries w ill naturally
have a high er aggregate B E R D /G D P ratio than one w ith m ost o f its activities in low -
R & D industries. T h ese stru ctu ral issues largely exp lain the differences in R & D
intensities across large an d sm aller econom ies (San dven and Sm ith 1997). T h e
question then is, does a specific in d u strial structure really m atter? T h is Is question
for debate, w h ich can n o t be addressed let alone settled here (it is interestingly
explored in Pol et a l 20 0 2); h ow ever the d esirab ility o f specific indu strial structures
is the real issue underlying use o f this aggregate indicator, though it is rarely
explicitly discussed. It is w o rth n o tin g also that within an in d u stry there tends to
be a w id e d istrib u tio n o f R & D intensities am o n g firm s, so it is co m m o n to find
156 KEITH SMITH

Table 6.1 6ERD/6DP ratios across countries

* ——.. -----
QERD/GDP2000 Percentage point^deviation from
Country
OECD mean
; ' ' Л.і'

3.65(1999) 1.40
Sweden
3.40 1.15
Finland
'■ '7;:
Japan 2.98 0.73
United States 2.72 0.47
.1;:•'•
$№■■;:}

Korea 2.65 0.40

Germany 2.49 0.29 ,■■■■' ;
France 2.18 -0.07
Netherlands 1.94 -0.26
Canada 1.87 -0.33
United Kingdom 1.85 -0.35
Austria 1.84 -0.36
Norway 1.65 -0.6
Australia 1.53 -0.72
Ireland 1.15 -1 .1
Italy 1.07 -1 .1 8
New Zealand 1.03 -1.22
Spain 0.94 -1 .3 1
Greece 0.67 -1.58
Total OECD 2.25

Source: OECD, M a in Scie n ce a n d T echnology Indicators D a ta b a se , accessed August 2003,

h igh -R & D firm s in lo w -R & D industries and vice versa (H u ghes 1988 discusses the
intra-industry distributions using U K data).
An im portant recent m odification o f this in d icator has been the addition of
“ acquired technology,” calculated as the R & D em b od ied in capital and intermediate
goods used b y an industry, and com puted via the m ost recent in p u t-o u tp u t table.
The m ethod for calculating acquired R & D is to assum e that the R & D em bodied in a
capital good is equal to the capital good's value m u ltiplied b y the R & D intensity of
the supplying industry. The m ost recent year for w hich relevant in p u t-o u tp u t data is
generally available is 1990. The overall structure o f the classification as currently used
can be seen in Table 6.2, which shows direct R & D intensities for the m ain industrial
groups for 1997, plus the proportion o f acquired to direct R & D for 1990, the last year
for which it was calculated.
 Table 6,2 Classification of industries based on RFtD intensity

ISIC Rev 3 Direct RFtD Acquired RFtD

Intensity intensity as °/o
1997 of direct RFtD
intensity, 1990

High technology Ind ustries

Aircraft and spacecraft 353 12.7 15
Pharmaceuticals 2423 11.3 8
Office, accounting and computing 30 10.5 25
machinery
Radio, television and communications 32 8.2 17
equipment
Medical, precision and optical 33 7.9 29
instruments

M e d iu m -h ig h -te ch n o lo g y in dustries
Electrical machinery and apparatus 31 3.8 42
Motor vehicles and trailers 34 3.5 29
Chemicals 24 exc 2.6 18
2423
Railroad and transport eqpt n.e.c. 352+359 2.8 88
Machinery and eqpt n.e.c. 29 1.9 104

M e d iu m -lo w -te ch n o lo g y industries

Coke, refined petroleum products and 23 0.8 30
nuclear fuel
Rubber and plastic products 25 0.9 127
Other non-metallic mineral products 26 0.9 285
Building and repairing of ships and 351 0.7 200
boats
Basic metals 27 0.7 289
Fabricated metals products 28 0.6 133

Low -tech nolog y in d u stries

Manufacturing n.e.c, and recycling 36-37 0.4 n.a.
Wood, pulp, paper, paper products, 20-22 0.3 167
printing and publishing
Food products, beverages and tobacco 15-16 0.4 267
Textiles, textile products, leather and 17-19 0.3 250
footwear

Sources: OECD, Science, Technology and industry Scoreboard 1999: Benchmarking Knowledge-Based
Econom ies (Paris:OECD 1999), Annex 1, p. 106; OECD, Science , Technology and Industry Scoreboard 20G1:
( Towards a Knowledge-Based Economy, Annex 1.1, pp, 13-139.
i? Mate: The JSiC elasslficatten was revised in 1996, though changes were relatively minor. 1990 data has been
reassigned to the most relevant Rev 3 category.
158 Keith smith

Table 6.2 shows that “ acquired tech n ology5' as a p ro p o rtio n o f direct R&D rises
dram atically as we m ove from high- to lo w -tech n o lo gy indu stries. This suggests that
technology intensity is likely to be very sensitive to h o w the measurement of
acquired technology is carried out. For exam ple, su p p o se we assum e that when a
firm buys a m achine it acquires not a p ro p o rtio n o f the R8dD that went into the
m achine (corresponding to the R & D /o u tp u t ratio) but all o f it? In other words,
purchasing a com puter gives the custom er access to all o f the R & D that was used to
produce it— this assum ption seems to be com patible w ith the know ledge externality
ideas o f the new grow th theory (for an o verview see Verspagen 1992, see also
Verspagen in this volum e). M aking this assu m ption w o u ld significantly alter the
rankings o f technology intensity in Table 6.2 by im p ro vin g the po sitio n o f industries
with substantial use o f R8cD em bodied in capital goods. A n o th er poin t to make here
is that so-called low -technology industries do not create o r access knowledge via
direct R8cD, and the classification is in effect biased against all industries that employ
n o n -R & D m ethods o f knowledge creation (H irsch -K rein so n et al. 2003). So the
indicator has drawbacks at the levels o f countries, indu stries and firm s; there are
therefore pitfalls in the uncritical use o f this apparently sim ple indicator.

6.4.3 Patent Data

A patent is a public contract between an inventor and a govern m en t that grants time-
lim ited m onopoly rights to the applicant for the use o f a technical invention (see
Iversen 1998 for a good review). The patentee m ust first dem onstrate a non-obvious
advance in the state o f the art after w hich the inventor enters into a binding
relationship with the state: in general, the inventor contracts to reveal detailed
inform ation about the invention in return for lim ited protection against others
using that invention for the time and geographical area for w hich the contract is in
force. In term s o f the concessions m ade by the parties, there is a trad e-o ff between
the disclosure o f detailed inform ation by the inventor against the perm ission of
lim ited m onopoly by the state. In this sense, the patent-system is designed as an
incentive-m echanism for the creation of new econ om ically valuable knowledge and
as a know ledge-dissem ination m echanism to spread this inform ation. There has
been a prolonged debate about whether the patent system w ou ld be w orth creating if
we did not have it (the usual answer is no), and w h eth er— since we do have it— it
should be abolished (again the usual answ er is no), o r w hether a reward system
would be superior (again, n o ).4
In general the patent system gathers detailed in fo rm atio n about new techno-
logies into a piotracted public record o f inventive activity, w hich is m ore or less
continuous. This gives it striking advantages as an innovation-indicator. These
include:
* Patents are granted for inventive technologies w ith com m ercial prom ise (i.e.
in n o vation ).
* The patent system system atically records im p o rtan t in fo rm atio n about these
inventions.
* The patent system collates these technologies according to a detailed and slow -to-
change classification system .
* The patent system system atically relates the invention to relevant technologies,
and also pro vid es links (via citations) to relevant technical and scientific literature.
* The patent system is an old institution , pro vid in g a lo n g h isto ry (see G ran stran d in
this v o lu m e )— it is the o n ly in n o vatio n in dicator extending back over centuries,
and this m eans that it is possible to use patents to explore quantitative issues over
very long p eriods (see B ru lan d and M o w e ry in this volu m e).
* The data is freely available.

The m ajo r sources o f patent data are the records o f the U S Patent O ffice and the
European Patent O ffice. Recent years have seen m ajo r increases in patenting activity,
as Figure 6 л show s. T h e causes o f this rise are an im p o rtan t issue: there does seem to
be grow th o f paten tin g extend in g back at least fifteen years, po ssib ly signifying
acceleration o f in n o vatio n efforts, or changes in strategic b eh avio r b y firm s; h o w ­
ever, the rise m ay also be shaped b y significant reductions in patent costs. (A n
analysis o f the issues here can be fo u n d in H all and Z ied o n is 2001; see also K o rtu m
and Lerner 1999.)

Fig, 6.1 USPTO utility patents 1 9 6 5 -2 0 0 2

Source: Hall 2QG3.
іб о KEITH SMITH

Patents also o f course have weaknesses, the m ost notable o f w h ich is that they are
an indicator o f invention rather than inn ovation : th ey m ark the em ergence o f a new
technical principle, not a com m ercial inn ovation . M a n y patents refer to inventions
that are intrinsically o f little technological or econ o m ic significance. M ore generally
Kleinknecht et a l have argued that

It is obvious that the patent indicator misses many non-patented inventions and innovations.
Some types of technology are not patentable, and, in some cases, it is still being debated
whether certain items (e.g. new business formulae on the internet) can be patented. On the
other hand, what is the share o f patents that is never translated into commercially viable
products and processes? And can this share be assumed to be constant across branches and
firm size classes? Moreover in some cases patent figures can be obscured by strategic behavior:
a firm will not commercialize the patent but use it to prevent a competitor patenting and
using it. (Kleinknecht et al. 2002:112)

But taking such qualifications into account, the analysis o f patent data has proven
very fruitful. Im portant achievem ents include the m ap p in g o f inventive activity over
long time periods (M acleod 1988; Sullivan 1990); assessing the im pacts o f economic
factors on the rate o f invention (Schm ookler 19 71); the elu cid ation o f the complexity
o f technological knowledge bases in large firm s (Patel and Pavitt 1999); the use and
roles o f science in industrial patenting (N arin and N o m a 1985; M eyer 2000); the
m apping o f inter-industry technology flows (Scherer 1982); the analysis o f spillovers
o f knowledge using patent citations (Jaffe, H enderson, and Trajtenberg 1993) and the
analysis o f patent values (Hall, Jaffe, and Trajtenberg 2001).

6.5 N ew In n o v a t io n In d ic a t o r s

Recent years have seen attem pts to create new and better-designed indicators
focused directly on innovation: for exam ple, the E uropean C o m m issio n has sup­
ported large-scale efforts to overcom e the absence o f direct data on industrial
innovation and there have been other attem pts to im p ro ve o u r knowledge of
outputs, sources, instrum ents and m ethods o f in n o vation (recent discussions are
Hansen 2001, Guellec and Pattinson 2001: Sm ith 2002).

6.5,1 Types of Innovation Survey

Innovation surveys divide into two basic types: those that focus on firm-level
innovation activity, asking about general innovation inputs (both R & D and non-
R8cD) and o u tp u ts (u su ally o f prod u ct in n o vation s), and those that focus on
significant technological in n o vation s (usu ally identified through expert appraisal,
or through n ew p ro d u ct announcem ents in trade jo u rn als o r other literature).
Som etim es the first o f these approaches is called a “ subject” app roach , since it
focuses on the in n o vatin g agent; the latter is referred to as the “ object” approach,
since it focuses on the objective o u tp u t o f the in n ovation process, on the tech nology
itself (A rchibugi and P ianta 1996). Both approaches can and do incorporate a t­
tem pts to exp lore aspects o f the in n ovation process itself: sources o f innovative
ideas, external inpu ts, users o f in n o vation , and so on. Both approaches define an
innovation in the Sch u m p eterian sense, as the com m ercialization o f a new product
or process. H ow ever the object app roach tends to focus on significantly new
products, w hile the subject ap p ro ach includes sm all-scale, increm ental change.

6.5.2 The “ Object” Approach to Innovation Indicators

Perhaps the m ost im p o rtan t exam p le o f the “ object” app roach is the S P R U database,
developed b y the Science P o licy Research U nit at the U n iversity o f Sussex, w h ich
collected in fo rm atio n on m a jo r technical in n o vation s in British industry, covering
sources and types o f in n o vatio n , in d u stry in n o vation patterns, cross-in d u stry lin k ­
ages, regional aspects, and so o n .5 T h e S P R U approach used a panel o f about 400
technical experts, d raw n from a range o f institutions, to id en tify m ajo r innovations
across all sectors o f the econom y, fro m 1945 th ro u gh to 1983. T h e database covered a
total o f ab ou t 4,300 in n o vation s. A n im p o rtan t related database is the U S Sm all
Business A d m in istratio n database, coverin g in n ovation s introduced to the m arket
by sm all firm s in the U S in on e year, 1982. T h is w as constructed through an
exam ination o f ab o u t on e h u n d red trade, engineering, and tech n ology jo u rn als—
a m ajor stu d y b y A cs an d A udretsch (1990) has been based u p on it. In add ition there
is a range o f sm aller literatu re-b ased surveys— based on searches o f trade litera­
ture— that have been u n d ertaken in recent years: the N etherlands, A ustria, Ireland,
and the U K fo r exam p le— K lein kn ech t an d B ain (1993) and K leinknecht (1996)
report the results fro m this w ork.
This type o f ap p ro ach has a n u m b er o f strong advantages. T echnology-oriented
approaches have the m erit o f fo cu sin g o n the tech n ology itself, and allow a fo rm o f
external assessm ent o f the im p o rtan ce o f an in n o vatio n — the fact that an in n o vation
is recognized b y an exp ert o r a trad e jo u rn a l m akes the co u n tin g o f an in n o vation
som ew hat in d ep en d en t o f p erso n al ju d g em en ts ab ou t w hat is o r is n o t an in n o v ­
ation. B o th exp ert-b ased an d literatu re-b ased approaches can be backw ard lo okin g,
thus g iv in g a h isto rical perspective o n tech nological developm ent.
Butthe approach also h as w eaknesses. T h e v e ry fact that in n o vation s m u st pass a
test of significance— that is, must b e su fficien tly in n o vative to be pu blicized .in trade
162 KEITH SMITH

jou rn als or the general press— also im parts a sam ple selection bias to the exercise* In
effect what these surveys cover is an im po rtan t subset o f the pop u lation o f innov­
ations: those that are new to an industry* W hat gets lost is the population of
innovation outputs w hich are “ routine,” increm ental, p art o f the norm al competi­
tive activity o f firm s, yet not strikingly new en ou gh to be reported.

6.5.3 Results from “ Object” Studies

One o f the m ost im portant results o f w ork using the S P R U database was to show the
existence o f quite different types o f innovative activ ity across different types of
industry. In a pioneering study, Pavitt (1984) d istin gu ish ed between four basic
firm types, which he called “ science based,” “ scale intensive,” “ specialized suppliers,”
and “ supplier d o m in ated ” He showed that these categories o f firm s were character­
ized by differences in sources o f technology, types o f users, m eans o f appropriation,
and typical firm size* This w ork was am on g the first to really dem onstrate empiric­
ally the im portance o f technological d iversity w ith in the econom y, with important
im plications for the design o f R & D p olicy in circum stances w here firm s have very
different technology creation patterns. O ther w o rk w ith the SP R U database has
em phasized the inter-sectoral flow o f in n ovation s (u sing the im p o rtan t data on first
users o f innovations w ithin the dataset), and gave an early em p irical insight into the
com plexity o f what is now called the system o f in n o vation (Pavitt, 1983; Robson et a l
1988). Geroski (1994:19) has sum m arized these intersectoral flow s as shown in Figure
6.2 where the key result is the im portance o f the three m ajo r engineering sectors
(m echanical engineering, instrum ents and electronic engineering) in terms o f the
flow o f innovations into other sectors. But it is im p o rtan t to note also the import­
ance o f flows w ithin this broad engineering com plex.

6.5.4 The “ Subject” Approach and the Community

Innovation Survey
In the early 1990s, the O E C D attem pted to synthesize the results o f earlier trial
innovation surveys, and to develop a m anual that m ight fo rm the basis o f a common
practice in this field, A group o f experts was convened, and over a period of
approxim ately fifteen m onths developed a consensus on an innovation manual
which becam e know n as the Oslo M anual (O E C D 1992).
I he European C om m ission, in a join t action betw een Eurostat and DG-
Enterprise, follow ed up the O E C D initiative in 19 9 2-3, im plem entin g the Commu­
nity Innovation Survey C IS was an innovative action in a n u m b er o f respects. First, it
 Food

Fig. 6.2 The SPRU innovation database: The intersectoral flow of innovations
Source: Geroski (1 9 9 4 ).

was a large-scale attem pt to collect internationally com parable direct m easures o f

innovation outputs. Second, it collected data at a highly disaggregated level and
made this data available in disaggregated fo rm to analysts. The survey has now been
carried out three tim es, m ost recently in 2002; in that year the survey covered
approxim ately 140,000 E uropean firm s.
C IS, in its vario u s versions, developed and incorporated data on the follow in g
topics:

* expenditure on activities related to the innovation o f new products (R & D , training,

design, m arket exploration, equipm ent acquisition and toolin g-up etc). There is
therefore a unique focus on n o n -R & D inputs to the innovation process;
* outputs o f increm entally and rad ically changed produ cts, and sales flow ing from
these products;
* sources o f in fo rm a tio n relevant to innovation;
* technological co llab o ratio n ;
* perceptions o f obstacles to in n o vation , and factors p ro m o tin g innovation.
In term s o f d efin ition s, the C IS follow ed the Oslo M anual in a n u m ber o f crucial
respects. Firstly, it fo cu sed o n tech nological in n o vation , particu larly in products.
But it then defin ed d ifferen t categories o f change, asking firm s to assign the product
range o f the firm to these different categories. T h e C IS also asked firm s to estim a te
the p ro p o rtio n s o f sales w h ich w ere co m in g from : new o r radically changed p ro d ­
ucts, fro m p ro d u cts w h ich had been changed in m in o r ways, o r fro m unchanged
products. T h e d efin itio n s o f tech nological in n o vatio n used in CIS-2, w hich have
Іб4 KEITH SMITH

been consistent throughout the variou s versions o f C IS , are sh ow n in Figure 6.3, It

should be noted that although both produ ct and process d efin ition s are offered, the
survey in fact concentrates on technologically changed p ro d u cts, m ain ly because of
the availability o f an econom ic m easure. M o st processes are o f course products of
capital good s-p rod u cin g firm s, although exp en d itu re on changing processes
extends well beyond ju st buying new equipm ent. Clearly, this lim its the scope of
the innovations on which data is being sought— ap art fro m processes, other aspects
o f innovative change, such as organizational change, u n d erlyin g learning processes,
and so on are excluded. However, this was done fo r considered reasons: focusing on
technologically changed products allow s a fairly rig o ro u s d efin itio n o f change to be
developed. Sales o f such products perm it at least a degree o f econ om ic commensur-
ability across firm s and even industries. It also p erm its reasonable definitions of
novelty: in deciding what was “ new ” about an in n o vatio n , the O slo M anual and CIS
identified different degrees o f produ ct in n o vation b y asking firm s to distinguish

Scope and impact of technological innovation and innovation a ctiv ity o f the enterprise

Technological innovations comprise implemented technologically new products and processes and
significant technological im provem ents in products and processes.

An innovation has been implemented if it has been introduced on the m arket (product innovation) or
used within a production process (process innovation). The product or process should be new (or
significantly improved) to the enterprise (it does not necessarily have to be new to the relevant market).

Technological innovation requires an objective improvement in the performance

of a product or in the way in which it is produced or delivered. The following
changes are not technological innovations:

* improvements of products that make them more attractive to the

Innovation activities are
ail those steps necessary
to develop and implement
technologically new or
improved products
purchasers without changing their ‘technological’ characteristics
. minor technological changes of products and processes or changes which
do not have the sufficient degree of novelty
* changes of products and processes, where the novelty does not concern
the use or objective performance characteristics of the products or the
way they are produced or delivered, but rather their aesthetic or
subjective qualities

A technologically new product is a product w hose technological characterist-ics or

intended uses differ significantly from those of previously produced products. Such
innovations can involve radically new technologies, can be based on com bining existing
technologies in new uses, or can be derived from the use of new know ledge

A technologically improved product is an existing product w hose perform ance has been
significantly enhanced or upgraded. A sim ple product m ay be im proved (in terms of better
performance or lower cost) th rou gh use of h igher-p erform ance com p onents or materials, or
a complex product w hich consist of a num ber of integrated technical subsystem s may be
improved by partial changes to one o f the subsystem s.

Fig. 6.3 Defining technological innovation—Community Innovation Survey (CIS)

Source: CIS-2 Questionnaire.
between sales o f p ro d u cts new to the firm only, products new to the industry, or
products that w ere w h olly new. So although the Oslo Manual!C IS approach co n ­
strains in n o vation to the field o f the technological, it does so in a w ay that allow s a
consistency betw een the concepts o f change, novelty, and com m ensurability. W ith ­
out such consistency, su rvey m ethods are not appropriate.

6.5.5 Innovation Activities and their Measurement

A second feature o f the Oslo Manual and o f C IS was the attem pt to estim ate
expenditures on categories o f in n o vatio n activity other than R & D . Six m ain cat­
egories o f in n o vatio n activities w ere identified, and the basic structure o f the
questions and d efin ition s w as as sh ow n in Figure 6.4. T h e basic idea here w as that
firm s invest in a w id e range o f n o n -R & D activities, resulting in both tangible and
intangible assets, and that these are likely to v a ry across firm s and industries. The
categories here are d raw n closely fro m K line and R osen berg (1986), w h ich provides
the general con cep tu al fo u n d atio n . B u t it can easily be seen that there are likely to be
problem s: these are co m p lex categories, in an area w here firm s do not necessarily
keep separate o r detailed records. In practice, in the first roun d o f the C IS , there were
m any firm s w h o d id n o t resp on d to the questions w hich were asked on this topic,
and m any w h o w ere clearly able to answ er on ly in term s o f b road estim ates. But there
are strong in terfirm variatio n s— som e firm s operate project m anagem ent system s
that perm it accurate answ ers in this area, and the data q u ality seem s to have
im proved o ver tim e.
One o f the im p o rta n t results to have em erged from this part o f C IS is that capital
expenditure related to in n o vatio n is the largest single com pon en t o f in n ovation
expenditure across all sectors (Evangelista et al. 1998). T h is em phasizes the im p o rt­
ance o f the em b od ied R & D in capital and interm ediate goods, discussed above.

6.5.6 CIS: Some Main Results

What have we learned so far from attempts to measure and map innovation? In this
section we look at some o f the results that have emerged from a range of studies using
CIS. The literature using innovation survey data is growing rapidly at the present
time, and it falls into three broad categories.

Descriptive overviews o f data results at national level. These studies are usually
written for policy makers, and typically consist o f tables and charts, accompanied
by commentary, showing results such as the distribution o f innovation expenditures
and their differences across industries, proportions o f firms introducing product or
l66 KEITH SMITH

Did your enterprise engage in the following innovation activities in 1996? txpeWiiart

— R esearch and development of new products and processes (re*d)

— A cquisition of machinery and equipment linked to product and process innovations
— A cquisition of external technology
— Industrial design , other production preparations for new products

— Training directly linked to innovations

— M arket introduction of innovations

Total expenditure

The expenditure items should cover current (labor If you have any RftD expenditure
costs, acquisition of services, materials etc.) and mentioned above, please
capital expenditure (instruments and equipment, in d icate...
computer software, land and buildings). If it is not
possible to estimate all expenditure items involved,
please at least indicate if your enterprise has been — percentage of RftD contracted out RftD
engaged in a particular innovation activity or not, — personnel in full tim e eqivaients in 1996
did your enterprise engage in R&D on a
~ continous basis (opposite to occasional)
between 1994 and 1996?

R esea rch and d evelop m en t o f p ro d u cts and p ro ce sse s (R ftD J com prises creative w ork undertaken on a
system atic basis in order to increase the stock of knowledge, and the use o f this stock o f know ledge to devise
new applications, Construction and testing of a prototype is often the m ost im portant phase o f REtD.
Softw are development is included as well. REtD can be carried out w ithin the enterprise or R8t0 services can
be acquired.

A cq u isitio n o f m a ch in ery and equipm ent lin ked to p ro d u c t a n d p ro c e ss in n o v a tio n s (including integrated
software) implemented by the enterprise.

A cq u isitio n o f extern a l tech n o lo g y in the form of patents, non-patented inventions, licenses, know -how ,
trademarks, draw ing plans and other consultancy services (excluding REtD), related to the im plem entation of
technological innovations, plus the acquisition of packaged softw are that is not classified elsewhere.

In du strial design and o th e r prod u ctio n p rep a ra tio n s fo r n ew p ro d u c ts include plans and d ra w in gs aimed at
defining procedures, technical specifications and operation features necessary to the production of
technologically new products and the im plem entation of new processes. D esign o f prototypes is a part of
RftD. This item also include changes in production and quality control procedures, m ethods and standards
and associated softw are required to produced the technologically new or im proved product or to use the
technologically new or improved process, Product or process m od ifications needed to start production,
including trial production (not included in R&D) is also included.

Training d ire ctly linked to in n o vatio n s is training for the im plem entation o f a technologically new or
improved product. Expenditure for training m ight include acquisition of external services and expenditure for
in-house training.

M a rk e t in tro d u ctio n o f in n ovation s includes activities in connection with the lau n ch in g of a technologically
new or improved product. These m ay include preliminary market research, m arket tests and launch
advertising, but will exclude the building of distribution netw orks to m arket innovations.

Fig. 6.4 Resources devoted to innovation activities in 1996

process innovations, the distribution o f different types o f new product sales across
industries, m ajor patterns o f technological collaboration, perceptions o f obstacles to
innovation, and data on objectives o f innovation. These studies tend to be important
not just in reaching p o licy m akers, but in em phasizing som e robust results w hich
em erge from this data in particu lar the conclu sion that in n o vation is pervasively
distributed across m o d ern econom ies, and that n o n -R & D inputs to in n ovation are
particularly im p o rtan t in n o n -h igh -tech sectors. In som e cases these reports are
sophisticated p ro d u ctio n s— the G erm an reports, for exam ple, rest on a substantial
panel dataset, and the C an ad ian analytical effort (sim ilar to but not identical
with C IS) is v e ry w ide ran gin g indeed (Jan z et al. 2002, and Statistics C anada:
ww w .statcan.ca ).6 M o st E U countries p rodu ce these reports and Eurostat in a d d ­
ition produces a E u ro p e-w id e o verview (Eurostat 2004).

Analytical studies sponsored by the European Commission. The E u ropean In novation

M o n itorin g System (w ith in D G -E n terp rise) has sponsored tw enty-five specific
studies add ressing a w id e range o f questions arising from the in n ovation data.
These cover, fo r exam ple, E u ro p e-w id e surveys o f in n ovation expen diture patterns,
innovation o u tp u ts across E u rop e, studies o f links betw een in n o vation and e m p lo y­
m ent patterns, an d sectoral studies (pharm aceuticals, telecom s, pulp and paper,
m achinery, m ach in e tools, service sector in n o vation , sp in -offs, and regional
im pacts). M o st o f these studies are substantial pieces o f w ork, often b o o k length.
An o verview o f the full range o f m aterial is pro vid ed in A p p en d ix 6.2 to this chapter
(reports are accessible v ia the E u rop ean In n o vatio n M o n ito rin g System on the E U s
C O R D IS website: w w w .co rd is.lu )

Econometric or statistical studies of innovation. The in n ovation su rvey data has a

m ore o r less u n iq u e feature, w h ich is that it is available in a h igh ly disaggregated fo rm
(as so-called “ m icro -ag g reg ated ” data). T h is m akes possible a w id e range o f m ic ro ­
level studies o f in n o vatio n processes and their effects, and the research o p p ortu n ities
this pro vid es are b ein g exp loited rather v ig o ro u sly at the present tim e. P u b lication in
this field has been b u ild in g rapidly, in the fo rm o f b oo ks (e.g. T h u riau x, A rn o ld , and
C ouch ot 20 0 1; K lein k n ech t and M o h n en 2002, and G au lt 2004), articles, jo u rn a l
STI Review 27 (20 0 1), and a fo rth co m in g special issue o f
special issues (such as
Economics of Innovation and New Technology), and so on. T h e b o o k edited b y
T h u rieau x et al. collects n o less than th irty-o n e chapters o n vario u s em pirical aspects
o f in n o vatio n u sin g p rim a rily C IS data. These covered m eth odological issues, the
extension o f the C IS ap p ro ach to services, m icro analysis o f in n o vatio n and firm
perform ance, in n o va tio n an d em ploym en t, in n o vatio n in trad ition al industries,
regional in n o vatio n , an d the use o f in dicators in p o licy decision m aking.

B y far the m o st ra p id ly g ro w in g area o f p u b licatio n is in sch olarly jo u rn als. A n on-

exhaustive rev ie w o f jo u rn a ls in 2002—4 reveals eighteen C IS-b ased pu blications.
These articles are b rie fly su m m arized in A p p e n d ix 6.1 to this chapter. Studies focus
on such to p ics as d eterm in an ts o f innovation, in n o vatio n and firm p erfo rm an ce
analysis, diversity (b oth in in n o va tio n patterns an d firm performance outcomes),
168 KEITH SMITH

the role o f science in innovation, sectoral p erfo rm an ce (such as employment

im pacts), inter-firm collaboration and in n o vation p erform an ce, as well as regional
and coun try studies, and m ethodological issues, T h ere is every sign that this pace of
publication will continue in years ahead. T h is is a rap id grow th in publication, and it
is w orth noting that it is occurring not o n ly in the fro n t-lin e jo u rn als o f innovation
studies, but also in the heart o f the econom ic m ain stream (n otab ly M airesse and
M ohnen in American Economic Review ).
Space lim itations prevent a detailed overview o f the results from the work
described above, but som e robust conclusions that seem to have em erged from the
literature as a w hole are as follows:
* Innovation is prevalent across all sectors o f the eco n o m y— it is not confined to
high-tech activities, and so-called low -tech activities contain high proportions of
innovating firm s, and often generate high levels o f sales from new and changed
products (SPRU , 1996; European C o m m issio n lo o t ) .7
* R & D is by no m eans the m ost im portant in n o vation inpu t. In all sectors, across all
countries, investm ent in capital equipm ent related to n ew p ro d u ct introduction is
the m ajor com ponent o f innovation expenditure, suggesting the need to focus on
the knowledge elements em bodied in such item s (STEP, 1997; Evangelista et al.
1998; Evangelista 1999).
* Across all sectors and countries innovation inputs and o u tp u ts are distributed
highly asym m etrically— sm all proportion s o f firm s account for large proportions
o f innovation outputs as m easured by the C IS.
* Collaboration is w idespread am ong innovating firm s, to such an extent that it
appears alm ost a sine qua non for innovation activity. T h is result from CIS has led
to a range o f specific subsidiary surveys, w hich have generated deeper detail and
have confirm ed the im portance o f collaboration suggested b y the C IS surveys (see
O EC D 2001 for papers on this).
* Extension o f the C IS form at to service sector activities is illustrative but problem­
atical, and deserves m ore attention (D jellal and G allo u j 2001; Tether and Miles
2001; Ch. 16 by M iles in this volum e).
* There continue to be significant differences in collection m ethodologies and
response rates across countries, im plying that the data appears to be m uch better
suited to w ithin-sector m icro studies than to cro ss-co u n try m acro comparisons.

6.6 C o n c l u s io n

W hile the C IS is clearly a step forw ard in term s o f the type and volu m e o f innovation
data that is available, it is ot course open to criticism . M o st criticism s focus on the
definitional restriction s in C IS w ith respect to in n ovation inputs and outputs, and
on whether an ap p ro ach that was o rigin ally adopted tor m an u factu rin g is extendable
to services. O n the ou tp u t side, the decisions m ade concerning the technological
definitions ot change o b vio u sly lim it the form s o f in n ovation that can be studied: it
seems to be the case that C IS w orks well fo r m anufactures, but not tor the extrem ely
heterogeneous services sector and its often intangible outputs. The analyses o f
Djellal and G allo uj (2001) and Tether and M iles (2001) suggest the need for quite
different approaches to data gathering on services. In defence o f the C IS approach it
can be argued that it is, and w as intended to be, m an u factu rin g-specific and that
extension to services w ou ld alw ays be problem atic. Sim ilar problem s arise w ith
other n on -tech n olo gical aspects o f in n o vation , such as organizational change (see
Lam, this volu m e, for an o verview o f organizational innovation ). It is very unclear
whether C IS, o r indeed any other su rvey-based m ethod, can grasp the dim ensions o f
this. The challenge fo r those w ho w ould go beyond this is w hether they can generate
definitional concepts, su rvey instrum ents, and collection m ethodologies that m ake
sense for other sectors o r other aspects o f innovation.
On the side ot R8cD and n o n -R & D in n ovation inputs, it is generally unclear just
how m uch o f a fir m s creative activity is captured by the types o f in n o vation outputs
that C IS m easures. A ru n d el has pointed ou t that “ W hen we talk about a firm
expending a great deal o f effort on in n o vation , we are not o n ly speaking o f financial
investm ents, b u t o f the use o f h u m an capital to think, learn and solve com plex
problem s and to p ro d u ce qu alitatively different types o f in n o vatio n s” (A rundel
1997: 6). T h is p o in t can n ot be argued w ith, but again the question arises as to w hat
can be done w ith su rvey qu estionnaires and w hat cannot. I f we w ant to explore
com plex p ro b lem so lvin g, fo r exam ple, then it is d ou btfu l w hether a su rvey in stru ­
ment is the righ t research to o l at all. Perhaps an u n d erlyin g issue here is the lo n g ­
standing ten sion betw een statistical m ethods, w ith their advantages o f generality but
lack o f depth, versus case stu d y m ethods, w h ich offer richness at the expense o f
generalizability.
N evertheless it is reasonable to conclu de that this data source is p ro vin g itself with
researchers. Both formal evaluations o f C IS as well as data tests by researchers have
been b ro ad ly p o sitive to the q u ality o f the data flo w in g from the su rvey (A alborg
U niversity 1995). O ne o f the p o sitive features o f C IS is that su rvey d efin ition and
construction, co llectio n m eth o d o logies, and general w o rk ab ility have been su b ­
jected to a degree o f evaluation, critique, and debate that goes far beyond anything
that has been carried out with other indicators (see Arundel et al, 1997, for one
contribution to the critical development o f C IS). This process is continuing, with
both positive and negative potential outcomes. On the positive side, the data source
m ay continue to be improved; on the negative, too much m ay be asked o f this
approach. But the real achievement is that C IS has produced results that have not
been possible with other data sources, and there is no doubt more to come as
researchers master the intricacies o f the data. In fact empirical studies using C IS
170 KEITH SMITH

data m ay well be the m ost rap idly grow ing su b -field o f pu b licatio n within innov­
ation studies at the present tim e. A n interesting feature o f the pu blications using CIS
is the breadth o f w ork being done— the data is being used fo r pu blic presentations,
tor policy analyses, and for a w ide range o f sch olarly research. It w as argued above
that researchers have yet to m ake full use o f the richness o f R & D data, and this
applies even m ore to the existing survey-based in n o vatio n data. T h is source will
continue to offer considerable scope to researchers in years ahead; issues such as
innovation and firm perform ance, the use o f science b y in n o vatin g firm s, the roles of
non-R8cD inputs, and the em ploym ent im pacts o f in n o vatio n are am on g likely areas
o f developm ent.
This chapter has concentrated on the C o m m u n ity In n o vatio n Survey, but future
developm ents are unlikely to rely on this source alone. O ne possible trend is for
greater integration o f existing data sources, and this can alread y be seen in multi-
indicator approaches to such issues as national com petitiveness. Another likely
trend is for the continued developm ent o f new su rvey in stru m en ts aligned to specific
needs, along the lines o f the D ISK O surveys on in terfirm collaboration (OECD,
2001). Such developm ents are m uch to be w elcom ed as In n o vatio n Studies seeks to
generalize its propositions beyond the lim its o f the case stu d y m ethod.

A P P E N D IX 6 .1

Recent (2002 onwards) journal publications using CIS data

Author(s) Data source Topic

Cox, Frenz, and Prevezer (2 0 0 2 ) CIS-2 Distinguishing high™ and low-tech

industries
Evangelista and Savona (2 0 0 2 ) CIS 1 and 2 Employment impacts of
innovation in service sector
H esselm an (2 0 0 2 ) CIS-2 Methodological issues and
response patterns
H inloo pen (2 0 0 3 ) C IS -1 and C IS -2 Determinants of innovation
performance at firm level across
Europe
fnzelt (2002) CIS-2 Service sector innovation m
Hungary
K lein kn ech t et al (2 0 0 2 ) CIS-2 Indicator choice and biases
L 6 6 f and H esh m ati (2 0 0 4 ) OS-2 Innovation and firm performance
Loof and H esh m ati (2 0 0 2 ) CIS-2 Performance diversity and
innovation
 Mairesse and Mohnen CIS-1 Determinants of innovation at firm
level
Mohnen and Horeau (2003) CIS-2 University-industry collaboration
Mohnen, Mairesse, and CIS-1 Expected vs. actual innovation
Dagenais {2003} output levels
Nascia and Perani (2002) CIS-1 Diversity of innovation patterns in
Europe
Guadros et al. (2001) Brazilian innovation Innovation in San Paulo region
survey
Selienthin and Hommen (2002) CIS-2 Innovation patterns in Swedish
industry
Tether (2002) CIS-2 Innovation and inter-firm
collaboration
Tether and Swann (2003) CIS-3 Role of science in innovation
Van Leeuwen and Klomp CIS-2 Innovation and multi-factor
(2004) productivity

A P P E N D I X 6.2

Publications using CIS data sponsored

by the European Commission

Publications are listed in chronological order, by topic and instituion.

Evaluation o f the Com m unity Innovation Su rvey (C IS)—Phase 1,

Aalborg University (Denmark), 1995
Europe's Pharm aceutical In du stry: An Innovation Profile (CIS),
SPRU (UK), 1996
Innovation O utputs in European Industry (CIS),
SPRU (UK), 1996
; Innovation in the European Food Products and Beverages Industry (CIS),
IKE (Denmark) and SPRU (UK), 1996
Technology Transfer, Inform ation Flow s and Collaboration (CIS),
Manchester School of Management Et University of Warwick (UK), 1996
The Im pact o f Innovation on Em ploym ent: A lternative interpretations and results o f the
Italian CIS,
University of Rome "La Sapienza" (Italy), 1996
^innovation in the European Chem ical in du stry (CIS),
[WZB (Germany) 1996
m . .. .. . ..... ................ .... ......... .
172 KEITH SMITH ____ _______ ____________________________ ____

Publications arc listed in dcronolojjical order, by topic and instituion. (cent*)

Innovation in the European Telecom Equipm ent Industry (CIS),

MERIT (Netherlands), 1996
Innovation A ctivities in Pulp, Paper and Paper Products in Europe (CIS),
STEP Group (Norway), 1996
The Im pact o f Innovation in Em ploym ent in Europe-А п A nalysis Using CIS Data,
Centre for European Economic R esearch/ZE W (Germany), 1 9 9 6
Com puter and Office M ach in ery-Firm s' external grow th & technological diversification:
analysis during CIS,
CESPRI (Italy) 1997
Innovation Expenditures in European Industry: analysis from CIS,
STEP Group (Norway), 1997
Manufacture of Machinery and Electrical Machinery (CIS),
Centre for European Economic Research/ZEW (Germany), 1997
Innovation M easurem ents and Policies: Proceedings o f International Conference,
20-21 May 1996, Luxembourg
Analysis o f CIS 2 Data on the Im pact o f Innovation on the Pharm aceuticals and
Biotechnology Sector,
SOFRES (Belgium), 2001
Analysis o f CIS 2 Data on the Im pact o f Innovation on Growth in the S e c to r o f Office
M achinery and Computer M anufacturing,
SOCINTEC (Spain), 2001
Analysis o f CIS 2 Data on the Im pact o f Innovation on Growth in M anufacturing of Machinery
and Equipment and o f Electrical Equipment,
STEP Group (Norway), 2001
Analysis of CIS Data on the Role of NTBFs, Spin-offs and Innovative Fast Growing SMEs in the
Innovation Process,
Institute for Advanced Studies and Johanneum Research (Austria), 2001
Innovation and the Acquisition and Protection o f Com petencies,
MERIT (Netherlands), 2001
Analysis of Empirical Surveys on Organisational Innovation and Lessons for Future Community
Innovation Surveys,
F rau n h o fer In s titu te (G erm any), 2 0 0 0

Regional Patterns o f Innovation: the Analysis o f CIS 2 R esu lts and Lessons from other
Innovation Surveys,
STEP SA S (Italy), 2000
Use of Multivariate Techniques to Investigate the Multidimensional Aspects of Innovation,
University of Newcastle Upon Tyne (ISRU) (UK), 2000
Statistics on Innovation in Europe,
European C om m ission, 2 0 0 1

Analysis of CIS 2 Data on Innovation in the Service Sector,

M a n c h e s te r U niversity (UK), 2 0 0 0

In n o v a tio n and enterprise creatio n : S tatistics and in d ic a to rs ," P roceedings o f th e International

C onference, 2 3
N otes

1. I would like to thank Ian Miles, Bart Verspagen, and Richard Nelson for comments on an
earlier draft, and in particular Bronwyn Hall for comments and advice. None are impli­
cated in the outcome, o f course.
2. The question o f what can be measured is an issue with all economic statistics. For example,
the national accounts do not cover all economic activity (in the sense o f all human activity
contributing to production or material welfare). They incorporate only activity that leads
to a measurable market outcome or financial recompense. This tends to leave out
economic activity such as domestic work, mutual aid, child rearing, and the informal
economy in general. Those services that are measured not by the value of output but by the
compensation o f inputs also provide problems for measurement o f output and product­
ivity.
3. In both Australia and Norway, each o f which collects data by field o f research for all
industrial sectors, roughly 25 per cent o f all R&D is in ICT.
4. An excellent overview o f the literature on these and other patent issues can be found on the
website o f Bronwyn Hall: http://emlab.berkely.edu/users/bhhall See also Granstrand in
this volume.
5. For analyses using the SPRU database, see e.g. Pavitt 1983,1984; Robson et al. 1988; the
most recent sustained analytical work using the SPRU database is Geroski 1994.
6. Canada is a leading site o f policy-related indicator work at the present time— see e.g. the
outstanding work o f the Canadian Science and Innovation Indicators Consortium which
can be found at the website given above.
7. On innovation in low-tech industries, see Ch. 15 by von Tunzelmann and Acha in this
volume.

R eferen ces*

Aalborg U n iv e r s it y (1995), Evaluation of the Community Innovation Survey (CIS)— Phase

i, 2 vols,, Report to the European Innovation Monitoring System.
Acs, Z., and A u d r et sc h , D. (1990), Innovation and Small Firms, Cambridge, Mass: M IT
Press.
*Archibugi and P ia n t a (1996), “ Innovation surveys and patents as technology indicators:
the state o f the art,” in OECD, Innovation, Patents and Technological Strategies, Paris:
OECD.
A rrow , K. J. (1984), “ Statistical Requirements for Greek Economic Planning,” Collected
Papers of Kenneth f Arrow, vol. 4: The Economics of Information) Oxford: Blackwell.
A rundel , A. (1997), “ Why Innovation Measurement Matters,” in A. Arundel and R. Garrelfs
(eds.), Innnovation Measurement and Policies) EU Luxembourg: EIMS 94/197.
Arundel, A., Patel , P,, Sirilli, G., and Smith , K. (1997)? The Future of Innovation
Measurement in Europe: Concepts) Problems and Practical Directions) IDEA Paper No 3,
STEP Group Oslo.

* Asterisked items are suggestions for further reading.

174 Ke it h sm ith

B iff l , Go and K n ell , M. (2001), “Innovation and Employment in Europe in the 1990s”
WIFO Working Paper 169/2001.
B rouw er , E,, and K lein kn ech t , A, (i997)> Measuring the Unmeasurable. A Country $
Expenditure on Product and Service Innovation, Research Policy 2 5 . 1235~ 4^*
________ (1999), "Innovative Output and a Firm's Propensity to Patent: An Exploration of
CIS Micro-Data;5 Research Policy 28: 615-24.
Cox, H., F ren z , M., and P r e v eze r , M. (2002), “ Patterns o f Innovation in UK Industry:
Exploring the CIS Data to Contrast High and Low Technology Industries;5 Journal of
Interdiciplinary Economics 13(1—3): 267-304.
*D je lla l , E, and G allo uj , R (2001), “ Innovation surveys for Service Industries: A Review,55
in Thurieaux, Arnold, and Couchot 2001: 70-6.
D ow rick , S. (2003), “A Review o f the Evidence on Science, R8cD and Productivity,55Working
Paper, Department o f Economics, Australian National University.
E uropean C ommission (1994), The European Report on Science and Technology Indicators
1994 (EUR 15897), Luxembourg.
—— (1997), Second European Report on Science and Technology Indicators 1997 (EUR 17639),
Luxembourg.
E u r o p e a n C o m m issio n (2001), Statistics on Innovation in Europe (EUR KS~32-oo~895-
EN-i), Luxembourg.
------ (2003), Third European Report on Science and Technology Indicators, Luxembourg.
eurosrat (2004), Innovation in Europe. Results for the EU, Iceland and Norway: Data 2998-
2ooi, (Luxembourg: Office for one official publications o f the European Communities)
------ San d ven , X , S ir il l i , G., and S m it h , K. (1998), “ Measuring Innovation in European
Industry,55 International Journal of the Economics of Business 5(3): 311-33.
------ and Savona , M. (2002), “ The Impact o f Innovation on Employment in Services:
Evidence from Italy,55 International Review of Applied Economics 16(3): 309—18,
E v a n g elist a , R. (1999), Knowledge and Investment, Cheltenham: Edward Elgar.
G au lt , E (ed.) (2004), Understanding Innovation in Canadian Industry, Montreal and
Kingston: McGill-Queens University Press.
G er o sk i , P. (1994), Market Structure, Corporate Performance and Innovative Activity, Oxford:
Clarendon Press.
G odin , B. (2004), “ The Obsession with Competitiveness and its Impact on Statistics: The
Construction o f High Technology Indicators,55 Research Policy (forthcoming).
G r if f it h , R., R edding , S., and V an R een en , J. (2000), “ Mapping the Two Faces of RM 3:
Productivity Growth in a Panel o f OECD Industries,55LSE Centre for Economic Performance
Discussion Paper, 2457: 1-74.
G r ilic h e s , Z. (1987), Comment,55 Washington, DC: Brookings Papers on Economic Activ­
ity, 3.
G r u p p , H. (1994), The Measurement o f Technical Performance o f Innovations by Techno-
me^rics and its Impact on Established Technology Indicators,55 Research Policy 2 3 :175“ 93-
* (Д998), Foundations of the Economics of Innovation, Theory, Measurement and Practice,
Cheltenham: Elgar.
G u ellec , D„ and Pattin so n , B, (2001), “ Innovation Surveys: Lessons from OECD Coun­
tries5 Experience,55 STI Review 27: 77—102.
H agedoorn , K., and Sch aken raad , J. (1990), “ Inter-firm Partnerships and Co-operative
Strategies in Core Technologies;5in C. Freeman and L. Soete (eds.), New Explorations in the
Economics of Technological Change, London: Pinter.
H all , В. H. and Ziedonis, R. H. (2002), “ The determinants o f patenting in the U.S,
semiconductor industry, 1980—1994 Rand Journal of Economics? 3 2 :10 1—28.
-----Jaffe, A., and Trajtenberg, M., 2001, “ The NBER Patent Citations Daa file: Lessons,
Insights and Methodological Tools^, NBER Working Paper No. 8498.
♦ H a n s e n , J. A. (2001), “ Technological Innovation Indicators: A Survey o f Historical Devel­
opment and Current Practice ” in M. R Feldmann and A. Link (eds.), Innovation Policy in
the Knowledge-Based Economy; Dordrecht: Kluwer, 73-103.
H esslem a n , L. (2002), A Description o f Responses to the UK Community Innovation
Survey 2,” The Journal of Interdiciplinary Economics 13(1-3): 243-66.
H inloopen , f. (2003), “ Innovation Performance across Europe," Economics of Innovation
and New Technology 12(2): 145-61,
H irsc h -K r e in se n , H., J a c o bsen , D., L a est a d iu s , S., and S m it h , K. (2003), “ Low Tech
Industries and the Knowledge Economy: State o f the Art and Research Challenges,"
Working Paper 2003:10? Dept o f Industrial Economics and Management, Royal Institute
ofTechnology (KTH), Stockholm.
Holbrook , J. (1991), “ The Influence o f Scale Effects on International Comparisons o f R&D
Expenditure," Science and Public Policy 18(4): 259-62.
H owells , J. (2000), “ Innovation Collaboration and Networking: A European Perspective,"
in Science Policy Support Group, European Research? Technology and Development Issues
for a Competitive Future? London.
H u gh es , K. (1988), “ The Interpretation and Measurement o f R&D Intensity: A Note,"
Research Policy 17: 301-7,
Inzelt , A. (2002), “Attempts to Survey Innovation in the Hungarian Service Sector," Science
and Public Policy 29(5): 367-83.
Iv er sen , E. (1998), “ Patents," in K, Smith (ed.), Science? Technology and Innovation Indica­
tors—a Guide for Policymakers? IDEA Report 5, STEP Group Oslo,
J aefe , A., and Henderson, R., and Trajtenberg, M., 1997, “ University Versus Corporate
Patents: A Window on the Basicness o f Invention", Economics of Innovation and New
Technology? 19-50.
J a n z , N., et al. (2002), Innovation Activities in the German Economy Report on Indicators from
the Innovation Survey 2000, Mannheim: ZEW.
K aloudis , A. (1998), “ Bibliometrics," in K. Smith (ed,), Science? Technology and Innovation
Indicators— a Guide for Policymakers? IDEA Report 5, STEP Group Oslo.
*K l ein k n ec h t , A. (ed.) (1996), Determinants of Innovation: The Message From New Indica­
tors? London: Macmillan.
------and В а ш , D. (eds.) (1993), New Concepts in Innovation Output Measurement? London:
Macmillan.
------and M o h n en , P. (eds.) (2002), Innovation and Firm Performance. Econometric Explor­
ations of Survey Data? Hampshire and New York: Palgrave.
------V an M o n tfo rt , K., and B ro uw er , E. (2002), “ The Non-Trivial Choice Between
Innovation Indicators," Economics of Innovation and New Technology 11(2): 109-21.
Kline , S., and R osenberg , N. (1986), “An Overview o f Innovation," in R. Landau (ed.), The
Positive Sum Strategy: Harnessing Technology for Economic Growth?Washington: National
Academy Press, 275-306,
Kortum, S„ and L e r n e r , J. (1999), “ What is Behind the Recent Surge in Patenting?" Research
Policy 28: 1-22,
I76 KEITH SMITH

L aesta d iu s , S. (2003), “ Measuring Innovation in the Knowledge Economy,” Paper pre­

sented to Pavitt Conference on Innovation, SPRU, Sussex, 13 Nov 2003*
Loof, H„ and H e s h m a t i , A. (2002), “ Knowledge Capital and Performance Heterogeneity:
A Firm-Level Innovation Study,” International Journal o f Production Economics 76(1): 61-85.
________ (2004), “ On the Relationship between Innovation and Performance: A Sensitivity
Analysis,” Economics o f Innovation and New Technology 13(1-2): forthcoming.
M c l e o d , C. (1988), Inventing the Industrial R evolution: The English Patent System, 1660-1800,
Cambridge: Cambridge University Press.
M a i r e s s e , J., and M o h n e n , P. (2001), “To be or not to be Innovative: An Exercise in
Measurement” STIReview 27:103-28.
*------- ---- (2002), “Accounting for Innovation and Measuring Innovativeness: An Illustra­
tive Framework and an Application,” American Economic Review 92(2): 226-30.
* M e y e r , M. (2000), “Does Science Push Technology? Patents Citing Scientific Literature,”
Research Policy 29: 409-34.
M o e d , H. E, D e B r u i n , R. E., and V a n L eeuw en ,T 1i . N. (1995), “ New Ribliometric Tools for
the Assessment o f National Research Performance: Database Description, Overview of
Indicators and First Application,” Scientometrics 33: 381-422.
M ohnen, P., and H oareau , C. (2003), “What Types of Enterprise Forge Close Links with
Universities and Government Labs? Evidence from CIS-2,” M a n a g e r ia l and Decision
Economics 24:133-45.
------M a i r e s s e , J., and D a g e n a i s , M. (2004), “ Innovativeness: A Comparison across Seven
European Countries,” Economics of Innovation and New Technology 13(1-2): forthcoming.
N a r i n , R, and N o m a , E. (1985), “ Is Technology Becoming Science?” Scientom etrics 3-6:
369-81.
N ascia , L., and Perani , G. (2002), “ Diversity of Innovation in Europe,” In te rn a tio n a l Review
of Applied Economics 16(3): 277-94.
OECD (1992, rev. 1997), In n o v a tio n M a n u a l: P ro p o sed G u id e lin e s f o r Collecting and Interpret-
ing In n o v a tio n D a ta ( O slo M a n u a l ), Paris: OECD, Directorate for Science, Technology and
Industry.
----- (1996), Technology, Productivity and Job C rea tio n , Paris: OECD.
(2001) , In n o v a tin g N etw o rks: C o lla b o ra tio n in N a tio n a l In n o v a tio n Systems, Paris:
OECD.
(2002) , T he M e a su erm en t o f S cien tific a n d T ech n o logical A ctvities. P ro p o sed Standard
P ractice fo r S u rveys on R esearch a n d E x p e rim e tn a l D e v e lo p m e n t: F ra sca ti M a n u a l, Paris:
OECD.
Pa tel , R, and Pa v itt , K. (1997), “The Technological Competencies of the World s Largest
Firms, Complex and Path-Dependent, but not Much Variety,” Research Policy 26:141-56.
(1999), The Wide (and Increasing) Spread of Technological Competencies in
the World s Largest Firms: a Cahllenge to Conventional Wisdom,” in A. Chandler et al
(eds. ), I he D y n a m ic F irm : I h e R o le o f Technology, Strategy, O rg a n iz a tio n a n d Regions,
Oxford: Oxford University Press.
Р а ѵ п r, K. (1983), Some Characteristics of Innovation Activities in British Industry,”
O m ega 11.
(1984), Sectoral Patterns of Technological Change: Towards a Taxonomy and a
Theory,” R esearch Policy 13: 343-73.
Pol , E., C arroll , R> and R obertson , P. (2002), “A New Typology for Economic Sectors with
a View to Policy Implications,” E con om ics of In n o v a tio n a n d N e w T ech n o logy 11(1): 61-76-
Q uadros , R., F u r i ado , A., B e r n a d e s , R., and E lia n e , F. (2001), “ Technological Innovation
in Brazilian Industry: An Assessment Based on the San Paulo Innovation Survey” Techno­
logical Forecasting and Social Change 67: 203-19.
R obson , M., T ow nsend , J., and P a v it t , K. (1988), “ Sectoral Patterns o f Production and Use
of Innovations in the UK: 1945-1983»” Research Policy 17(1): 1-15.
R o sen berg , N. (1976), Perspectives on Technology; Cambridge: Cambridge University Press.
------(1982), Inside the Black Box: Technology and Economics, Cambridge: Cambridge Univer­
sity Press,
Sandven , T., and S m it h , K (1997), “ Understanding R&D Indicators: Effects o f Differences in
Industrial Structure and Country Size,” IDEA Paper 14, STEP Group Oslo.
Sav io tti , P. P. (1996)* Technological Evolution, Variety and the Economy, Cheltenham: Elgar.
------(2001), “ Considerations about a Production System with Qualitative Change,” in
J. Foster and J, Stanley Metcalfe (eds.), Frontiers of Evolutionary Economics. Competition,
Self Organization and Innovation Policy>Aldershot: Edward Elgar, 197-227.
Sch erer , F. (1982), “ Inter-industry Technology Flows in the United States,” Research Policy
11(4): 227-45.
Schmookler , J. (1971), “ Economic Sources o f Inventive Activity,” in N. Rosenberg (ed.), The
Economics of Technological Change, London: Pelican, 117-36.
S ellen th in , M., and H o m m en , L. (2002), “ How Innovative is Swedish Industry? A Factor
and Cluster Analysis o f CIS II,” International Review of Applied Economics 16(3): 319-32.
Smith K. (1992), “ Technological Innovation Indicators: Experience and Prospects ” Science
and Public Policy 19(6): 24-34.
------(2001), “ Innovation Indicators and the Knowledge Economy: Concepts, Results and
Challenges,” in Thurieaux, Arnold, and Couchot (eds), Luxembourg: European Commis­
sion, 14-25.
SPRU (1996), Innovation Outputs in European Industry (CIS), Report to the European
Innovation Monitoring System.
STEP G roup (1997), Innovation Expenditures in European Industry: Analysis from CIS, Report
to the European Innovation Monitoring System.
Su l l iv a n , R. (1990), “ The Revolution o f Ideas: Widespread Patenting and Invention During
the Industrial Revolution”, Journal of Economic History 50(2): 340-62.
T eth er , B. (2002), “ Who Cooperates for Innovation and Why: An Empirical Analysis,”
Research Policy 31(6): 947-67.
------and M il e s , I. (2001), “ Surveying Innovation in Services— Measurement and Policy
Interpretation Issues,” in Thurieaux, Arnold, and Couchot (2001).
------and Swann , G. M. P. (2003), “ Services, Innovation and the Science Base: An Investi­
gation Into the UK's 'System, o f Innovation' Using Evidence from the Third Community
Innovation Survey,” paper to CNR/Univerity o f Urbino Workshop on Innovation In
Europe.
*T h u r iea u x , B., A rn o ld , E., and C ouchot , C. (eds.) (2001), Innovation and Enterprise
Creation: Statistics and Indicators, Luxembourg: European Commission (EUR 17038).
V an L eeu w en , G., and K lo m p , L. (2004), “ On the Contribution o f Innovation to Multi­
factor Productivity,” Economics of Innovation and New Technology 13(1—2): forthcoming.
V e r s p a g e n , B. (1992), “ Endogenous Innovation in Neo-classical Growth Models: A Survey,
Journal of Macroeconomics 631-62.
W orld Economic Forum (2003), Global Competitiveness Report 2002—2003, ed. P. Cornelius,
K. Schwab, and M. E. Porter, New York: Oxford University Press.
 PART II

THF SYSTFMTC
td ^ L» «ѵ іів тМ шт* ЛШшштИ * Jb ▼ ,я Д ь hJ H L

N A T U R E OF
ялкь *A * JKm » « А » «Л ^ т^шттш

INNOVATION
 In t r o d u c t i o n to P a r t II

A central finding in innovation research is that firm s seldom

innovate in isolation. Interaction w ith cu stom ers, suppliers, com ­
petitors and various other private and p u b lic o rgan ization s is very
im portant, and a “ system perspective” is useful in understanding
and analyzing such interaction. In C h ap ter 7, E d q u ist traces the
developm ent o f such an “ in n ovation system ” perspective, particu­
larly w ith respect to nations (so-called “ n ation al system s o f in n ov­
ation” ), and discusses the achievem ents, sh ortcom in gs and
potential o f this approach. H e also considers the role o f som e
central activities o f such system s, e.g., R & D and education.
Follow ing this, M o w ery and Sam pat (C h ap ter 8) exam ine one o f
the central organizations in national system s o f innovations: u n i­
versities. C hapter 9 by O 'Su llivan focuses on an o th er aspect o f
innovation system s, finance, discussing the varied approaches
adopted on the relationship between finance and in n ovation . An
essential feature in innovation system s is the co n d itio n s for app ro­
priation o f the econom ic returns to in n o vation . G ranstrand
(Chapter 10) provides a historical o verview o f intellectual property
rights and surveys the extensive literature on this topic. The next
two chapters deal w ith the bou n daries o f system s o f innovation. As
Edquist notes in C hapter 7 these bou n d aries d o not have to be
national, but m ay be regional, global or sectoral (“ sectoral system s”
are discussed in m ore detail in Part III o f this volu m e). Regional
system s o f innovation are analyzed b y A sh eim and G ertler (C h ap­
ter 11) and N arula and Zanfei in C h apter 12 exam in e the “ globaliza­
tion o f innovation and the role o f m u ltin atio n al enterprises in this
process.
 C H A P T E R 7

S Y S T E M S OF
INNOVATION
PERSPECTIVES
AND CHALLENGES

CHARLES EDQUIST

7.1 In t r o d u c t i o n 1

T h is chapter presents an o verview an d assessm ent o f the system s o f in n ovation

approach. I focus m a in ly o n n ation al system s o f in n o vation , but in add ition address
sectoral and region al system s o f in n o vatio n to a lim ited extent.2 T h e chapter
addresses the em ergence and d evelopm ent o f the system s o f in n o vation (SI)
approach, its strengths an d w eaknesses, the criticism that it is “ u n d erth eo rized ”
the constituents o f S is, the m ain fu n ctio n and activities in Sis, the b ou n daries o f Sis,
and proposals fo r fu rth er research. I also discuss h o w the rig o u r and sp ecificity o f the
SI approach co u ld be in creased.3 T h e m o st central term s used in this chapter are
specified in B o x 7.1.
182 CHARLES EDQUIST

Box 7 Л Systems of innovation: main terms used ____________

Innovations = product innovations as well as process innovations. Product innovations

are 0ew__0Г better— material goods as well as new intangible services. Process innov­
ations are new w a p o f producing goods and services. They may be technological or
organizational.
SI = system o f innovation = the determinants o f innovation processes — allimportant
economic, social, political, organizational, institutional, and other factors that influ­
ence the development, diffusion, and use o f innovations.
Constituents o f Sis = components + relations among the components.

Main components in Sis — organizations and institutions.

Organizations= formal structures that are consciously created and have an explicit
purpose. They are players or actors.
Institutions = sets of common habits, norms, routines, established practices, rules, or
laws that regulate the relations and interactions between individuals, groups, and
organizations. They are the rules of the game.
An SI has a function, i.e. it is performing or achieving something. The main function in
Sis is to pursue innovation processes, i.e. to develop, diffuse and use innovations.
Sis are those factors that influence the development, diffusion, and use of
Activities in
innovations. The activities in Sis are the same as the determinants of the main
function.

7.2 T he E m e r g e n c e a n d D e v e l o p m e n t
of t h e SI A p p r o a c h

The chapter b y Fagerberg in this volum e highlights the system ic nature o f innov­
ation processes, noting that firm s do not n o rm ally in n ovate in isolation, but in
collaboration and interdependence w ith other organizations. These organizations
m ay be other firm s (suppliers, custom ers, com petitors, etc.) o r non-firm entities
such as universities, schools, and governm ent m inistries. T h e beh avior o f organiza­
tions is also shaped by institutions— such as laws, rules, n orm s, and routines— that
constitute incentives and obstacles for innovation. These organizations and insti­
tutions are com ponents o f systems for the creation and com m ercialization of
knowledge. Innovations emerge in such "system s o f in n o v a tio n "
The innovation concept used in this chapter is w id e and include product innov­
ations as well as process innovations. Product in n o vation s are new — or better—
m aterial goods as well as new intangible services. Process in n o vation s are new ways
producing goods and services. I hey m ay be tech nological or organizational
(Edquist, H om m en, and M cK elvey 2001),
 S Y S T E M S OF I N N O V A T I O N 183

The expression 4 n ation al system o f in n o vation ” (N SI) was, in published form ,

first used in Freem an (1987). H e defined it as “ the netw ork o f institutions in the
public and p rivate sectors w hose activities and interactions initiate, im port, and
diffuse new technologies (Freem an 19 8 7 :1). Two m ajor books on national system s
o f innovation (N SI) are Lu n dvall (1992) and N elson (1993), w hich em ploy different
approaches to the stu d y o f N SIs. N elson (1993) em phasizes em pirical case studies
more heavily than th eo ry d evelopm en t4 and som e o f the studies focus narrow ly on
nations R8cD system s. B y contrast, Lundvall (1992) is m ore theoretically oriented
and seeks to develop an alternative to the neo-classical econom ics tradition by
placing interactive learn in g, u se r-p ro d u c e r interaction and innovation at the center
o f the analysis (Lu n d vall 19 9 2 :1) .
Lundvall argues that “ the structure o f p ro d u ctio n ” and “ the institutional set-u p ”
are the tw o m ost im p o rtan t dim en sion s that “ jo in tly define a system o f in n o vation ”
(Lundvall 1992; 10 ). In a sim ilar way, N elson and R osenberg single out organizations
supporting R & D , i.e. th ey em phasize those organizations that prom ote the creation
and d issem ination o f know ledge as the m ain sources o f innovation (N elson and
Rosenberg 1993: 5, 9 - 1 3 ) .5 LundvalTs broader approach recognizes that these
“ narrow ” o rgan ization s are “ em bedded in a m uch w ider so cio-econ om ic system
in which po litical and cu ltu ral influences as well as econom ic policies help to
determ ine the scale, d irection and relative success o f all innovative activities”
(Freem an 2 0 0 2 :19 5 ).
Both N elson and Lu n d vall define national system s o f innovation in term s o f
determ inants of, o r factors in flu en cin g, in n ovation processes.6 H owever, they single
out different d eterm inants in their actual definitions o f the concept, presum ably
reflecting w hat th ey believe to be the m ost im p o rtan t determ inants o f innovation.
Hence, they p ro p o se different defin itions o f the concept, but use the sam e term . T his
reflects the lack o f a generally accepted d efin ition o f a national system o f innovation.
A m ore general d efin ition o f (national) system s o f in n ovation includes “ all
im portant econ o m ic, social, political, organizational, institutional and other factors
that influence the d evelopm ent, d iffu sio n and use o f in n o vatio n s” (Edquist 1997b:
14). I f all factors that influence in n o vatio n processes are not included in a definition,
one has to argue w h ich po ten tial factors sh ou ld be excluded— and why. This is quite
difficult, since, at the presen t state o f the art, w e do not k n o w the determ inants o f
innovation system atically an d in detail. It seem s dangerous to exclude som e p o ten ­
tial d eterm inants, since these m igh t prove to be v ery im portan t, once the state o f the
art has advanced. F o r exam p le, tw enty-five years ago it w ou ld have been natural to
exclude the in teraction s betw een organizations as a determ inant o f in n ovation
processes. In clu d ed in this general d efin ition are the relationships am on g the factors
listed and the action s o f both firm s and governm ents.
There are other specifications of systems of innovation than national ones.
Carlsson and colleagues focus on “technological systems, arguing that these are
unique to technology fields (Carlsson 1995)* The sectoral approach of Breschi and
184 C H A R LE S EDQUIST

M alerba (1997) sim ilarly focuses on a grou p o f firm s that d evelop and manufacture
the products o f a specific sector and that generate an d utilize the technologies of
that sector. The concept o f “ regional in n o vation system ” has been developed and
used by C o oke et al. (1997) and Braczyk et al. (199^)» C o o k e (20 0 1), an d Asheim and
Isaksen (2002).
The three perspectives— national, sectoral and region al— m a y be clustered as
variants o f a single generic “ system s o f in n o vatio n ” ap p ro ach (Edquist 1997b: 3)
11- 12 ). M uch o f the discussion in this chapter is relevant fo r the generic approach,
and is based on the prem ise that the different varian ts o f system s o f innovation
coexist and com plem ent each other. W hether the m ost ap p ro p riate conception of
the system o f innovation, in a certain context, sh o u ld be national, sectoral or
regional, depends to a large extent on the questions one w ants to ask.7

7. 3 S t r e n g t h s a n d W e a k n e s s e s
o f t h e SI A p p r o a c h

7.3.1 The Diffusion of the SI Approach

The diffusion of the SI approach has been surprisingly rapid, and is now widely used
in academic circles. The approach also finds broad applications in policy contexts—
by regional authorities and national governments, as well as by international organ­
izations such as the OECD, the European Union, UNCTAD and UNIDO. In Sweden,
a public agency has even been named after the approach, i.e. the Swedish Agency for
Innovation Systems (VINNOVA). The practice of VINNOVA is strongly influenced
by the SI approach, an approach that appears to be especially attractive to policy
makers who seek to understand differences among economies5innovative perform­
ance, and develop ways to support technological and other kinds of innovation. The
next section briefly addresses some of the strengths of the generic SI approach.

7 -3-2 The Strengths of the SI Approach

Ih e SI approach places innovation and learning processes at the center o f focus. This
e m p h a s is o n le a rn in g a c k n o w le d g e s th a t in n o v a tio n is a m a tt e r o f p r o d u c in g new
k n o w le d g e o r c o m b in in g e x is tin g (a n d s o m e t im e s n e w ) e le m e n t s o f k n o w le d g e in
 S Y S T E M S OF I N N O V A T I O N I 85

new ways. T h is focus distinguishes the SI approach from other approaches that
regard technological change and other innovations as exogenous.

The S I a p p ro a c h a d o p ts a h o lis tic a n d in te r d is c ip lin a r y p e rs p e c tiv e . It is “ h olistic” in

the sense that it tries to encom pass a w ide array— or all— o f the im portan t determ in­
ants o f in n o vation , and allow s fo r the inclu sion o f organizational, social, and
political factors, as well as econom ic ones. It is “ in terdisciplin ary” in the sense that
it absorbs perspectives fro m different (social science) disciplines, including eco ­
nom ic history, econom ics, sociology, regional studies, and other fields.

The S I a p p ro a c h e m p lo y s h is to r ic a l a n d e v o lu tio n a r y p e rs p e c tiv e s , w h ic h m a ke s th e

n o tio n o f o p t im a lit y ir r e le v a n t. Processes o f in n ovation develop over tim e and
involve the influence o f m an y factors and feedback processes, and they can be
characterized as evolu tion ary. Th erefore, an o ptim al or ideal system o f innovation
cannot be specified. C o m p ariso n s can be m ade between different real system s (over
time and space), and betw een real system s and target system s, but not between real
systems and o p tim al ones. A lth o u gh this is a com plex view o f the innovation
process, it is far rich er and m o re realistic than its alternatives.

T he S I a p p ro a c h e m p h a s iz e s in te rd e p e n d e n c e a n d n o n - lin e a r ity . This is based on the

understanding that firm s n o rm ally do not innovate in isolation but interact w ith
other organ ization s th ro u gh co m p lex relations that are often characterized by
reciprocity and feedback m echanism s in several loops. In n ovation processes are
not only influenced b y the co m pon en ts o f the system s, but also b y the relations
between them . T h is captures the n o n -lin ear features o f in n ovation processes and is
one o f the m ost im p o rtan t characteristics o f the SI approach.

T he S I a p p ro a c h c a n e n c o m p a s s b o th p r o d u c t a n d process in n o v a tio n s , as w e ll as
s u b ca te g o rie s o f these typ e s o f in n o v a t io n . Traditionally, in n ovation studies have, to
a large extent, focused u p o n tech nological process innovations and to som e extent
upon pro d u ct in n o vatio n s, b u t less on n on-tech nological and intangible ones, i.e.
service p ro d u ct in n o vatio n s and organizational process innovations (as specified in
Section 7.2). A s argu ed in this H an d b o ok , there are go od reasons to use a com pre-
hensive in n o vatio n co n cep t,8 an d the system s o f in n ovation approach is well suited
to this co m preh en sive perspective, since all the categories o f in n ovation s specified in
this chapter can be analyzed w ith in it. T h at n on-tech nological form s o f in n ovation
deserve m o re atten tion is also argued in O E C D (2002a: 24.d).

T h e S I a p p ro a c h e m p h a s iz e s th e ro le o f in s titu tio n s , P ractically all specifications o f the

SI concept h igh ligh t the role o f institu tion s, rather than assum ing them aw ay fro m
the list o f d eterm in an ts o f in n o vation . T h is is im p o rtan t since institutions strongly
186 CHARLES EDQUIST

influence innovation processes* There is, how ever, no agreem ent about what the
term “ institutions” m eans (see Section 7.3*3)*

These six characteristics are often considered to be strengths o f the SI approach by

academ ic analysts, policy m akers, and increasingly b y firm strategists, and partly
explain its rapid diffusion. H owever, the SI approach also has weaknesses, which
represent challenges for future research o n system s o f in n o vation .

7,3.3 The Weaknesses of the SI Approach

The SI approach is still associated w ith conceptual diffuseness. O ne exam ple is the
term “ institution,” which is used in different senses b y different authors: it is
som etim es used to refer to organizational actors as w ell as to institutional rules.
Som etim es the w ord m eans different kinds o f organ ization s o r “ players” (according
to the definitions in Section 7.4.2). At other tim es, the term m eans laws, rules,
routines, and other “ rules o f the gam e.” For N elson and R osen berg (1993), insti­
tutions are basically different kinds o f organizations, w hile fo r Lu ndvall (1992) the
term “ institution” m eans prim arily the rules o f the gam e. H ence “ institution” is used
in several different senses in the literature (Lundvall 19 9 2 :10 ; N elson and Rosenberg
1993* 5) 9~i3; Edquist 1997b: 26-8).
A nother exam ple o f conceptual diffuseness is that the o rigin ato rs o f the SI
approach did not indicate what exactly shou ld be included in a “ (national) system
o f innovation” ; they did not specify the boun daries o f the system s (Edquist 1997k
13 -15 ). Nelson and Rosenberg provided “ no sharp guide to ju st w hat should be
included in the innovation system, and w hat can be left o u t” (N elson and Rosenberg
1993: 5-6 ). Lundvall insisted that “ a definition o f the system o f innovation must be
kept open and flexible” (Lundvall 19 9 2 :13 ).
W ith regard to the status o f the SI approach, it is certain ly not a form al theory, in
the sense o f providin g specific p ro po sitio n s regarding causal relations among
variables. It can be used to form ulate conjectures fo r em pirical testing, but this has
been done only to a lim ited degree (see Section 7.5). Because o f the relative absence of
well-established em pirical regularities, “ system s o f in n o vatio n s” should be labeled
an approach or a conceptual fram ew ork rather than a th eo ry (E dqu ist 1997b : 28-9)*
Scholars disagree on the seriousness o f these weaknesses o fth e SI approach and on
h o w they should be addressed. A ccording to som e, the app roach sh ou ld not be made
to o rigorous, the concept should not be “ overth eorized ” and it should remain an
inductive one.9 A nother position argues that the SI ap p ro ach is “ undertheorized”
th at conceptual clarity should be increased and that the app roach should be m ade
m ore “ theory-like.” 10
 S Y S T E M S OF I N N O V A T I O N 187

Hence, the in tern ation al co m m u n ity w ithin innovation studies is currently

divided on this issue. In w h at follow s, I try to increase the rigo r and specificity o f
the SI approach. T h is effort is intended as a step tow ards developing the approach
further. I f it reveals w eaknesses associated w ith the approach, this is a good thing.
Acknow ledging such w eaknesses m ay lead to addition al research and new insights
into the o p eratio n o f in n o vatio n system s.

7. 4 T h e C o n s t i t u e n t s , F u n c t i o n ,
A c t i v i t i e s a n d B o u n d a r i e s of S i s

7.4.1 What is a System?

In an effort to d evelop the SI approach, it m ight be useful to relate it explicitly to
“ general system s theory,” w h ich has been used m u ch m ore in the natural sciences
than in the social sciences. In everyd ay language, as well as in large parts o f the
scientific literature, the term “ system ” is used generously and w ith lim ited dem ands
for a precise d efin ition . To the question “ W hat is a system ?” there is, however,
a com m on answ er in everyd ay language as well in scientific contexts (Ingelstam
2 0 0 2 :19 ):

• A system consists o f tw o kinds o f constituents: There are, first, som e kinds o f

com ponents and, second, relations am on g them . The com ponents and relations
should fo rm a coherent w h ole (w hich has properties different from the properties
o f the constituents).
• The system has a fu n ctio n , i.e. it is p erfo rm in g o r achieving som ething.
• It m ust be p ossib le to d iscrim in ate betw een the system and the rest o f the w orld;
i.e. it m ust be possib le to id en tify the b ou n d aries o f the system . I f we, for exam ple,
want to m ake em p irical studies o f specific system s, we m ust, o f course, kn ow their
extent.11
M akin g the system s o f in n o vation s app roach m ore theory-like does not require
that all co m p o n en ts an d all relations am o n g them m ust be specified. Such an
am bition w o u ld certain ly be unrealistic. A t present, it is not a m atter o f tran sform in g
the SI ap p ro ach into a “ general th eo ry o f innovation,” but rather w e need to m ake it
clearer and m o re consisten t so it can better serve as a basis for generating hypotheses
about relations betw een specific variables w ithin S is (w hich m ight be rejected o r
supported th ro u g h em p irical w o rk ). Even the m uch m ore m odest objective o f
specifying the m a in fu n ctio n o f Sis, the activities and com ponents in them and
l 8g CHARLES EDQUIST

som e im portant relations am ong these, w o u ld represent a considerable advance in

the held o f innovation studies. Used in this way, the SI ap p ro ach can be useful for the
creation o f theories about relations between specific variables w ith in the approach,

7.4*2 The Main Components o f Sis

O rganizations and institutions are often considered to be the m ain com ponents of
Sis, although it is not always clear w hat is m eant b y these term s (as argued in Section
7.3.3). Let m e therefore specify what organizations and in stitu tion s m ean here.

O rg a n iz a tio n s a re f o r m a l s tru c tu re s t h a t a re c o n s c io u s ly c re a te d a n d h a v e a n e x p lic it

p u rp o s e (Edquist and John son 1997: 46-7)* T h ey are players o r actors.12 Some
im portant organizations in S is are hrm s, universities, ven tu re capital organizations
and public agencies responsible for in n ovation policy, co m p etition policy or drag
regulation.

In s titu tio n s a re sets o f c o m m o n h a b its , n o rm s , r o u tin e s , e s ta b lis h e d p ra c tic e s , rules, or

la w s th a t re g u la te th e re la tio n s a n d in te ra c tio n s b e tw e e n in d iv id u a ls , g ro u p s , and
o rg a n iz a tio n s (Edquist and John son 1997: 46). T h ey are the rules o f the game.
Exam ples o f im portant institutions in S is are patent laws, as well as rules and
norm s influencing the relations between universities and hrm s. O bviously, these
dehnitions are o f a “ N orthian” character (N orth 1990; 5), d iscrim in atin g between
“ the rules o f the gam e” and “ the players” in the gam e.
Sis m ay differ from one another in m an y respects. F o r exam ple, the set-ups of
organizations and institutions, constituting com pon en ts o f em p irically existing Sis,
vary am ong them. Research institutes and com pan y-b ased research departments
m ay be im portant R & D perform ers in one co u n try (e.g., Japan ) w hile research
universities m ay play a sim ilar role in another (e.g., the U n ited States). In some
countries, such as Sweden, m ost research is carried ou t in universities, while the
independent public research institutes are weak. In G erm any, the latter are much
m ore im portant. That the organizational set-up varies co n sid erab ly am ong NSIs is
shown in profiles o f the national system s in A ustria, B elgiu m , Finland, Germany,
Spain, Sweden, Switzerland, and the United K in gd o m , presented in O E C D (1999л:
A nnex 3).
Institutions such as laws, rules, and norm s also d iffer con siderab ly am ong na­
tional Sis. For exam ple, the patent laws are different betw een countries. In the USA,
an inventor can publish before patenting, whereas this is not possible according to
European laws. With regard to the patent rights o f u n iversity teachers, individuals in
this category own their patents outright in Sweden, thanks to the so-called “ univer­
sity teachers' privilege ” However, this is not the case in the U SA , w here different laws
 S Y S T E M S OF I N N O V A T I O N 189

apply. In D en m ark and G erm any, new laws have recently transferred ow nership
from the teachers to the universities, w hile in Italy, a transfer has occurred in the
opposite d irection. M an y O E C D governm ents are currently experim enting with
changes in the ow n ersh ip o f know ledge created in universities, in the b elief (based on
little evidence— see the chapter b y M o w ery and Sam pat) that such changes will
influence the p ro p en sity to patent and accelerate the com m ercialization o f eco n o m ­
ically useful know ledge.

In sum m ary, there seem s to be general agreem ent that the m ain com ponents in Sis
are organizations— am o n g w h ich firm s are often considered to be the m ost im p o rt­
ant ones— and institution s. H ow ever, the specific set-ups o f organizations and
institutions v a ry am o n g system s.

7.4.3 Functions and Activities in Sis

A lthough a system is n o rm ally considered to have a function, this was not addressed
in a system atic m an n er in the early w o rk on Sis. Som ew hat later, som e hints in this
direction w ere m ad e b y G alli and Teubal (1997: 34 6 -7). As we will see below, som e
recent co n trib u tio n s to the literature have started to address this theoretical gap.

7.4.3.1 Functions a n d A ctivities in Sis an d D eterm inants

o f In n ovation Processes
X ielin Liu and Steven W hite (2001) address w hat they call a fu ndam ental w eakness o f
national in n o vatio n system research, n am ely “ the lack o f system -level exp lan atory
factors'5 (L iu an d W h ite 2 0 0 1:10 9 2 ). To rem edy this, they focus u p on the “ activities55
in the system s, “ activities” b ein g related to “ the creation, d iffu sion and exploitation
o f tech nological in n o vatio n w ith in a system ” (Liu and W hite 2001: 1093). O n this
basis, they co m pile a list o f five fu ndam ental activities in innovation system s.13
Jo h n so n and Jaco b sso n (2003) em phasize that, for an in n ovation system “ to
sup port the grow th o f an industry, a n u m ber o f fu nctions have to be served w ithin
it, e.g, the su p p ly o f resou rces” (Jo h n so n and Jacobsson 20 0 3:2 ). T h ey suggest that “ a
tech nology o r p ro d u ct specific in n o vation system m ay be described and analysed in
term s o f its "functional pattern,5 i.e. in term s o f h ow these functions are served55
(Jo h n so n an d Jaco b sso n 2003: 3). These authors present a list o f five fu n ctio n s.14
Rickne (2000: 175) pro vid es a list o f eleven fu nctions that are im p o rtan t for new
tech n ology-b ased firm s (i.e. not for in n ovation s in an im m ediate sen se).15 Clearly,
there is n o con sen su s as to w h ich fu nctions o r activities should be included in a system
o f in n o vatio n an d this pro vid es abundan t o p p ortu n ities for fu rth er research.
190 CHARLES EDQUIST

One w ay o f addressing what happens in S is is the fo llo w in g. A t a general level,

the m ain function— or the overall fu n ctio n in S is is to pu rsu e innovation
processes, i.e. to develop, diffuse and use innovations. W h at I, fro m now on, call
activities in Sis are those factors that influence the d evelo pm en t, d iffusion, and use
o f in n o vatio n .16 Exam ples o f activities are R & D as a m eans o f the developm ent of
econom ically relevant knowledge that can p ro vid e a basis fo r innovations, or the
financing o f the com m ercialization o f such know ledge, i.e. its transform ation into
innovations.
A satisfactory explanation o f innovation processes alm ost certain ly will be multi-
causal, and therefore should specify the relative im p o rtan ce o f variou s determinants.
These determ inants cannot be expected to be independen t o f each other, but instead
m ust be seen to support and reinforce— or offset— one another. H ence, it is import­
ant to also study the relations am ong variou s determ inants o f in n o vation processes.
O ne way to try to do that w ould be to establish “ a h ierarch y” o f causes a la E. H. Carr.
C arr argues that the study o f h istory is a stu d y o f causes and that the historian
continuously asks the question, “ W hy?” Further, the h isto rian co m m on ly assigns
several causes to the sam e event (C arr 1986: 81, 83). H e continues: “ The true
historian, confronted with this list o f causes o f his ow n co m p ilin g, w ould feel a
professional com pulsion to reduce it to order, to establish som e hierarchy o f causes
which w ould fix their relation to one another, perhaps to decide w hich cause, or
which category o f causes, should be regarded fin the last resort' o r fin the final
a n a ly sis'... as the ultim ate cause, the cause o f all causes” (C a rr 1986: 84). I do not
believe that we will ever reach such an objective in a detailed and system atic manner
or that we will be able to identify all determ inants o f in n o vatio n — because o f the
com plexity o f the task. However, there are go od reasons to try to strive in this
direction b y developing theories about relations betw een specific variables within
the approach in a pragm atic w ay (as proposed in Section 7.4.x.).
I believe that it is im portant to study the activities (causes, determ inants) in Sis in
a system atic manner. The hypothetical list o f activities presented below is based
upon the literature, e.g. the lists m entioned earlier, and on m y ow n know ledge about
innovation processes and their determ inants. The activities listed are not ranked in
order o f im portance, but start w ith know ledge inputs to the innovation process,
continues with the dem and side factors, the pro vision o f constituent o f Sis, and ends
w ith support services for innovating firm s.
The follow ing activities can be expected to be im p o rtan t in m ost Sis:

(1) Provision o f Research and D evelopm ent (R8cD), creating new knowledge,
prim arily in engineering, m edicine, and the natural sciences.
(2) Com petence building (provision o f education and training, creation of
hum an capital, production and reproduction o f skills, in d ivid u al learning)
m the labor force to be used in innovation and R & D activities.
(3) Form ation o f new product markets.
 S Y S T E M S OF I N N O V A T I O N 191

(4) A rticu latio n o f q u ality requirem ents em anating from the dem and side with
regard to new produ cts.
(5) C reating and ch an gin g organizations needed for the developm ent o f new
fields o f in n o vation , e.g. enhancing entrepreneurship to create new firm s and
intrapreneu rship to diversify existing firm s, creating new research organ iza­
tions, p o licy agencies, etc.
(6) N etw orkin g th rou gh m arkets and other m echanism s, including interactive
learning betw een different organizations (potentially) involved in the in ­
n ovation processes. T h is im plies integrating new know ledge elem ents d e­
veloped in different spheres o f the SI and com in g from outside w ith elem ents
already available in the in n ovatin g firm s.
(7) C reating and ch an gin g institutions— e.g. IP R laws, tax laws, environm ent
and safety regulations, R8cD investm ent routines, etc— that influence in n o v­
ating organ izatio n s and in n o vation processes b y pro vid in g incentives o r
obstacles to in n ovation .
(8) In cub atin g activities, e.g. p ro vid in g access to facilities, adm inistrative su p ­
p ort, etc. fo r new in n ovative efforts.
(9) Fin an cin g o f in n o vatio n processes and other activities that can facilitate
co m m ercialization o f know ledge and its adoption.
(10) P ro visio n o f consu ltancy services o f relevance for in n ovation processes, e.g.
tech n ology transfer, co m m ercial in form ation , and legal advice.
This list is p ro visio n al and w ill be subject to revision as o u r know ledge about
determ inan ts o f in n o vatio n processes increases. In add ition to a set o f activities that
is likely to be im p o rtan t in m ost Sis, there are activities that are v e ry im po rtan t in
som e kinds o f Sis an d less im p o rtan t in others. For exam ple, the creation o f technical
standards is critically im p o rtan t in som e (sectoral) system s, such as m obile telecom ­
m u n icatio n s.17
The system atic app roach to S is suggested here does not im p ly that they are or can
be consciously designed o r planned. O n the contrary, just as in n ovation processes
are evolutionary, S is evolve over tim e in a largely unplanned m anner. Even i f we
knew all the determ inants o f in n o vatio n processes in detail (w hich we certainly do
not now, and perhap s never w ill), w e w o u ld not be able to control them and design
or “ b u ild ” S is o n the basis o f this know ledge. C entralized control over Sis is
im possible and in n o vatio n p o licy can o n ly influence the spon tan eous developm ent
o f Sis to a lim ited extent.

7.4.3.2 Three K in d s o f L ea rn in g in the S I A pproach

R egarding com peten ce b u ild in g as an im p o rtan t activity in S is— and given that
R & D has earlier been a central activity in SI studies— m eans that the SI approach
focuses o n three k in d s o f learning:
192 CHARLES EDQUIST

* Innovation (in new products as well as processes) takes place m ain ly in firms and
leads to the creation o f “ structural capital* w hich is a know ledge-related asset
controlled by firm s (as opposed to * h u m an capital ); it is a m atter o f organiza­
tional learning.
. Research and Development (R & D ) is carried ou t in universities and pu blic research
organizations as well as in firm s and leads to p u b licly available know ledge as well as
know ledge ow ned by firm s and other organizations, as w ell as b y individuals.
• Competence Building (e.g. training an d edu cation) w h ich o ccu rs in schools and
universities (schooling, education) as well as in firm s, and leads to the creation of
“ hum an capital.” Since hum an capital is controlled b y in d ivid u als, it is a matter of
individual learning.
A n im portant area for further research in the SI tradition concerns the relationship
am ong these three kinds o f learning, which appear to be closely related to one another.
One objective o f such studies would be to address w hat types and levels o f education
and training are m ost im portant for specific kinds o f innovations— e.g. process
innovations and product innovations, or increm ental and radical innovations.
As exem plification, I w ill now discuss tw o central learning activities— R& D and
com petence building— in som e greater detail. T h is d iscu ssion constitutes the be­
ginning o f an exam ination o f the relations am o n g activities and constituents in Sis
that is continued in section 7.4.4. It also serves as a guide to and synthesis o f some of
the recent w ork on these issues w ithin the system s o f in n o vatio n literature.

A. Research and D evelopm ent C onsiderable w o rk on N S Is has been carried out

w ithin the O ECD . However, although m ost o f the O E C D con trib u tion s mentioned
here have “ systems o f innovation” in the title, m an y o f them actually use this
approach m ore as a label than as an analytical took T h e first phase o f this work
included the developm ent o f quantitative indicators, c o u n try case studies, and work
within six focus groups on innovative firm s, in n ovative firm networks, clusters,
m ob ility o f hum an resources, organizational m app in g, and catch in g-u p economies.
Som e o f the em pirical results were presented in (O E C D 1998a) and a synthesis is
found in (O E C D 1999a).
The second phase provided a deepening o f the analysis in three areas: innovative
clusters (O E C D 20 0 1a and 1999b), cooperation in n ation al innovation systems
(O E C D 2001b), and m obility o f skilled personnel in N S Is (O E C D 20 0 1c). Yet another
study (O EC D 2002a) sum m arizes the findings o f the second phase o f the project and
derives policy im plications. In the studies m entioned, R & D as well as competence
building is addressed to som e extent.
In m ost countries, universities are the m ost im p o rtan t pu blic organizations
perform ing R & D (O EC D 1998a: ch, 3). G overn m ents fu nd u n iversity R & D activities
in a num ber o f ways. Traditionally, they have p ro vid ed general su p port via block
grants from the M inistries o f Education, part o f w hich was used b y university staff to
carry out R& D . Such funding is still very im portant in sm all, h igh ly R&D-intensive
 S Y S T E M S OP I N N O V A T I O N 193

countries such as the N etherlands, Sweden, and Sw itzerland, G overnm ents m ay also
provide grants to encourage research for the advancem ent o f know ledge” or grants
to obtain the know ledge needed fo r governm ent m issions such as defense or health
care. In m ost countries, b lock grants have declined and direct su p port has grow n in
im portance (O E C D 1998a: ch. 3).
In certain countries, universities fall under the responsibility of the national
government. In others, such as Germany, they are the responsibility of the regional
governments. Whatever the form of organization, a growing regional influence can
be observed in most countries. In Germany, the universities are financially very
autonomous. In the UK, financial support is provided by research councils to
individual projects selected on a competitive basis.
In m any countries, the science system also includes pu blic research institutes (or
“ national lab o rato ries” ) w h ich carry ou t the sam e type o f R & D activities as un iver­
sities, as w ell as m o re applied research and technical developm ent w ork. A lthough
the relative im p o rtan ce o f universities in term s o f p erform in g R & D has increased in
most countries (see M o w e ry and Sam pat in this volu m e), pu blic research organ iza­
tions rem ain im p o rtan t, These organizations m ay be linked to the universities and
included in the h igh er edu cation sector, or they m ay be independent o f them. The
largest single case in the O E C D area is the Centre N ation al de la Recherche Scien-
tifique (C N R S ) in France, w h ich receives the lion s share o f direct fu nding o f R & D in
the higher edu cation sector. T h e C N R S provides su p p o rt for projects that are
norm ally carried out in co llab oration w ith u n iversity researchers. In this regard,
the C N R S can be d e a rly d istinguished from its counterparts in G erm an y (M ax
Planck G esellschaft), Italy (C N R ) and Spain (C S IC ) (Laredo and M u star 2 0 0 1c $02).
In the U nited States, the h igh er education sector contains a large n um ber o f public
research laboratories. (O E C D 1998a: 8 3-4 ). O ther countries w ith а large institute
sector include N orw ay, Taiw an and G erm an y (e.g. M a x Planck Gesellshaft and
Fraunhofer). A n u m b er o f national governm ents have tried to change these o rgan ­
izations and p ro m o te their links w ith the rest o f the econ o m y and society. T h is has,
for exam ple, been d o n e in quite different w ays in France and the U K (Laredo and
M ustar 2 0 0 1c 503).
As this sh ort d iscu ssion suggests, different kinds o f pu blic organizations (such as
universities and p u b lic research institutes) can perform the sam e activity (R & D ) in
an N SI. N SIs d iffer sign ifican tly w ith regard to w hich organizations that perform
public R & D and w ith regard to the institutional rules that govern o r influence these
organizations (Lared o an d M u star 2001b: 6 -7 ).
In most NSIs, especially in low- and medium-income nations, only modest sums
are invested in R & D and most of the R & D is performed by public organizations. The
few countries that invest heavily in R & D are all rich, and much of their R & D is
carried out by private organizations. This group includes some large countries, such
as the U S A an d Jap an , b u t also som e sm all and m ediu m -sized countries, such as
Sweden, Sw itzerland, an d So u th Korea. There are also som e rich countries that do
194 CHARLES EDQUIST

rather little R & D , e.g. D enm ark and N orw ay. A s m en tion ed, a considerable part of
the R & D in m any rich countries is carried out and financed b y the private sector,
prim arily firm s (although there are also pu blic financial su p p o rt schem es to stimu­
late firm s to perform R & D ). The p ro p o rtion o f all R & D p erfo rm ed in high-income
O E C D m em ber states that is financed b y firm s ranged betw een 21 per cent (Portugal)
and 72 per cent (Japan) in 1999 (O E C D 2002 b). A ck n ow led gin g such differences may
help to distinguish between different types o f N SIs.
M ost o f the R & D carried out b y private organ ization s m ay be characterized as
developm ent w ork rather than research. In n o vatio n certain ly does not depend solely
on R & D results, but requires also other actions, such as technical experimentation,
technology adoption, m arket investigations, and entrepreneurial initiative. R&D
and innovation activities are norm ally driven b y d ifferent rationales and motiv­
ations— i.e. the advancem ent o f know ledge and the quest fo r profits, respectively
O ne im plication o f the com plex interface betw een “ research” an d “ innovation” is
that links between universities/public research o rgan ization s and innovating firms
are especially im portant to the perform ance o f N S Is.18 In n o vatin g firm s often need
to collaborate w ith public research organizations o r universities. Here, publicly
created institutions are im portant. G overnm ents m ay, fo r exam ple, support collab­
orative centers and program s, rem ove barriers to co op eration , and facilitate the
m obility o f skilled personnel am ong different kinds o f organizations. This might
involve the creation o f institutional rules, such as those in Sw eden stating that
university professors should perform a “ third task” in ad d itio n to teaching and
doing research— i.e. interact w ith the society su rro u n d in g the university, including
firm s. H owever, such “ linkage activity” is carried o u t in different ways and to
different extents in different N SIs.

B. Competence building. The early w ork w ithin the SI ap p ro ach largely neglected
learning in the form o f education and train in g .19 H ow ever, com petence building is
increasingly considered to be an im portan t activity in system s o f innovation,
reflecting the im portance o f skilled personnel fo r m ost innovative activities
(Sm ith 2001: 8).20 But no rigorous analyses o f com petence b u ild in g have, to my
knowledge, been conducted as part o f the analysis o f in n o vatio n systems.
Nevertheless, there is a large literature on variou s aspects o f com petence building
outside the SI context. Com petence building (e.g. train in g and education) is the
sam e as enhancem ent o f hum an capital and is carried ou t largely, though not
exclusively, in schools and universities, C om petence b u ild in g also occurs in firms
(in the form o f training, learning-by-doing, learn in g-b y-u sin g, and individual
learning) often throughout w orking life,
A recent O E C D study analyzed vocational and technical education and training in
som e detail in Australia, Austria, D enm ark, England (in clu d in g W ales and Northern
Ireland), France, G eim any, Italy, the N etherlands, Q uebec, and Switzerland (OECD
1998b). This study pointed out m any differences across countries with regard
 S Y S T E M S OF I N N O V A T I O N 195

to vocational and technical training. O ne difference concerns the stage prior to

vocational and technical training, i.e. the structure o f m iddle and lower secondary
education. T his structure is unified in m ost countries, but is divided into distinct
program s in G erm any, A ustria, the N etherlands, and Switzerland. O ne o f these
program s is the begin n in g o f the academ ic pathway, w hile the others lead essentially
to vocational and technical training. A n oth er difference concerns the relative n u ­
meric im portan ce o f vocatio n al and technical training as opposed to academ ic
pathways in u p per secon dary education. In the countries o f the British C o m m o n ­
wealth— A ustralia, C an ad a, and the U K — the academ ic pathw ay is very m uch in the
majority, w hile in the countries o f continental Europe, vocational and technical
training d om in ates (O E C D 1998 b).
The ways in which people access skilled jobs (and then climb the career “ladders”
of enterprises) differ greatly among NSIs:
This may occur at a certain time after recruitment, when the young person has proved
himself; after a fixed and codified period o f service, according to a specific labour contract; or
on recruitment, depending on the qualifications previously acquired. For vocational
training, these three modes o f access lead to three broad traditions: on-the-job training,
formal apprenticeship, school training. (OECD 1998b: 12)

These practices coexist in various countries, but their relative importance varies
considerably; frequently, one of them dominates and determines training policy.
The m odels for transition fro m education to em ploym ent also differ across coun ­
tries. Apprenticeship is im portan t in som e countries— e.g. in Germ any, where it caters
to about tw o-thirds o f the age group (O E C D 1996; 48). In other countries, school-
based learning and productive w o rk are com bined in alternative ways— e.g. in
Sweden, Australia, France, the U nited K ingdom , and Korea (O E C D 1996; 146).
The organizational and institutional contexts of competence building thus vary
considerably among NSIs. There are particularly significant differences between the
systems in the English-speaking countries and continental Europe. However,
scholars and policy makers lack good comparative measures on the scope and
structure of such differences. There is little systematic knowledge about the ways
in which organization of education and training influences the development, diffu­
sion and use of innovations. Since labor, including skilled labor, is the least mobile
production factor, domestic systems for competence building remain among the
most enduringly “national” elements of NSIs.

7.4.4 The Relations Between Activities and Components

and among Components
This chapter has placed greater em phasis on “ activities than m uch of the early
w ork on S is. N onetheless, this em ph asis does n o t m ean that we can disregard o r
19б CHARLES EDQUIST

neglect the “ com ponents” o f S is and the relations am o n g them . Organizations or

individuals perform the activities and institutions p ro vid e incentives and obstacles
influencing these activities. In order to understand an d exp lain innovation pro*
cesses, we need to address the relations between activities an d com ponents, as well as
am ong different kinds o f com ponents.
W hat then are the relations between the co m pon en ts an d the activities in Sis? As
we saw in Section 7.4.3.2.A, the activity o f research (the creation o f new knowledge)
can be carried out b y research institutes, universities, o r research-oriented firms.
M ost o f the other activities m entioned earlier can also be p erfo rm ed by different
organizations. Further, m any categories o f o rganizations can perfo rm m ore than
one activity. For exam ple, universities provide new know ledge and educate people
(hum an capital). Hence, there is not a o n e-to -o n e relation betw een activities and
organizations.21 H owever, there are lim its to this flexib ility— fo r instance, primary
schools cannot carry out basic research. The relations betw een activities and insti­
tutions are less direct, since institutions influence w hether o r not, and how, certain
organizations perform certain activities. It seem s that the set-up o f activities can be
expected to vary less across N SIs than the set-up o f o rgan ization s perform ing them
and the set-up o f institutions influencing those organizations. H ow ever, the “ quan­
tity" o f each activity and the efficiency w ith w hich each activity is perform ed might
vary considerably am ong N SIs.22
As we saw in section 7.3.2, the SI approach em phasizes the relations or interactions
am ong the com ponents in Sis. Interactions am o n g different organizations maybe of
a m arket o r non-m arket kind. That m arket, as well as n o n -m ark et, relations should
be addressed in SI research is stressed in a recent O E C D report. There the concept of
interaction is specified as including:

* C om petition, which is an interactive process w herein the actors are rivals, and
which creates or affects the incentives for innovation.
* Transaction, which is a process by w hich goods and services, including technol­
ogy-em bodied and tacit knowledge, are traded betw een econ o m ic actors.
* N etw orking, which is a process by w hich know ledge is transferred through collab­
oration, cooperation and long term netw ork arrangem ents, (O E C D 2002a: 15)
W ith regard to interactions am ong organizations in their pu rsu it o f innovations,
em pirical w ork inspired by and designed on the basis o f the SI approach has been
carried out in m any countries. O ne exam ple is the C o m m u n ity In novation Surveys
(CIS) that have been coordinated by Eurostat o f the E u rop ean U n io n and carried out
in all EU countries and in several additional countries (see Sm ith , C hapter 6 in this
volum e, for a detailed discussion of C IS). The C IS results include data on collabor­
ation am ong innovating organizations, and indicate that such collaboration is very
com m on and im portant. This result is supported by oth er surveys which have
shown that between 62 per cent and 97 per cent o f all p ro d u ct innovations
were achieved in collaboration between the in n o vatin g firm and some other
 S Y S T E M S OF I N N O V A T I O N

organizations in A u stria, N orw ay, Spain , D enm ark, and the region o f East G othia in
Sweden (Christensen et al. 1999; O rstavik and Nas 1998; Edquist, Ericsson, and
Sjogren 2 0 0 0 : 4 7 )*
These findings constitute em p irical su p p o rt for one o f the m ain tenets o f the SI
approach, i.e. that interactive learning am o n g organizations is crucial for innovation
processes. This also illustrates the d ynam ics o f this field o f research over tim e, The
emergence o f the SI ap p ro ach in its D anish version (Lundvall 1992) took inspiration
from case studies in d icatin g that u se r-p ro d u ce r interaction was very im portan t for
innovations, e.g. in the D an ish d airy in du stry; the SI approach was form ulated
partly on this basis. O ne o f its central elem ents— the im portance o f relations o f
interactive learning am o n g organizations— has since been verified by system atic
empirical research in D en m ark and elsewhere. T h is is a good exam ple o f fruitful
interaction between theoretical and em pirical w ork.
Another exam ple o f em p irical w o rk partly based on the N S I approach is Furm an,
Porter, and Stern (2002), w h o introduce the concept o f national innovative capacity,
which is the ability o f a c o u n try to p ro d u ce and com m ercialize a flow o f new -to-the-
world technologies o ver the lo n g term . T h is concept is explicitly based u p on ideas-
driven endogenous grow th th eo ry a la R o m er (see Verspagen, Ch. 18 in this volum e),
the cluster-based ap p ro ach a la Porter (1990), and the N SI approach. O n this basis,
they estim ate the relation sh ip betw een international patenting (patenting by foreign
countries in the U SA ) an d observable m easures o f national innovative capacity.
Their results suggest that a sm all n u m b er o f observable factors describe a co u n try’s
national innovative cap acity— i.e. they identify determ inants o f the produ ction o f
new -to-the-w orld technologies. T h ey find that a great deal o f variation in patenting
across countries is due to differences in the level o f inputs devoted to innovation
(R&D m an p ow er and sp en din g). T h ey also find that an extrem ely im portan t role is
played by factors associated w ith differences in R & D productivity, e.g. po licy choices
such as the extent o f p ro tectio n o f intellectual p ro p erty and openness to inter­
national trade, the share o f research p erfo rm ed b y the academ ic sector and funded
by the private sector, the degree o f technological specialization, and each in divid ual
country’s know ledge “ sto ck ” (Fu rm an et al. 2002).
The relations betw een o rgan izatio n s an d institutions are im portant for in n o v­
ations and fo r the o p eratio n o f Sis. O rganizations are strongly influenced and
shaped b y in stitution s, so that o rganizations can be said to be “ em bedded in an
institutional en viron m en t o r set o f rules, w hich include the legal system , norm s,
routines, standards, etc. B u t in stitu tion s are also em bedded in and develop w ithin
organizations. E xam p les are firm -sp ecific rules w ith regard to bookkeeping or
concerning n o rm s w ith regard to the relations between m anagers and em ployees.
Hence, there is a co m p licated relationship o f m u tu al em beddedness between in sti­
tutions and o rgan izatio n s (E d q u ist and Jo h n so n 1997* 59 "6 o).
Some organizations create institutions that influence other organizations.
Examples are organizations that set standards and public organizations that
Ifi C H A R L E S EDQUIST

formulate and implement those odes that we might call Innovation, policy. Sam ples
are the N M T 450 and the GSM mobile telecom standards. The NM T 4501«as created,
by the Nordic public telephone operators, which were state-owned monopolies at
the time. The development and implementation o f NM T 450 was an example oft he
importance o f user-producer relations in innovation processes, which is stressed 90
strongly in the SI approach. The public organizations provided a technical frame­
work for private equipment producers and thereby decreased uncertainty,. The
Nordic equipment producers, Ericsson and Nokia, greatly benefited from tills, and
it was an important factor contributing to their leading role In mobile telecommuni­
cations equipment production today. In essence, the NM T 450 provided the cradle
for the development of mobile telecommunications in Europe (Edquist 2003; 21-3).
Institutions may also be the basis for the creation o f organizations, as Mien a
government makes a law that leads to the establishment o f an organization.
Examples of such organizations include patent offices or public innovation policy
agencies.
There may also be important relations between different institutions, for example,
between patent laws and informal rules concerning exchange o f information
between firms. Institutions o f different kinds may support and reinforce each
other, but may also contradict and be in conflict with each other, as discussed in
some detail by Edquist and Johnson (1997). This work has- teen carried forward by
Coriat and Weinstein (2002), who discuss different levels o f institutions and focus
on the principle o f a hierarchy among rules themselves (Coriat and Weinstein
2002: 2B0).23
Our knowledge about the complex relations—characterized by reciprocity and
feedback—between organizations and institutions Is limited. Since the relations
between two phenomena cannot be satisfactorily analyzed if they are not conceptu­
ally distinguished from each other, it is important to make a clear distinction
between organizations and institutions when specifying the concepts.24

7.4.5 Boundaries of Sis: Spatial? Sectoral and

in Terms of Activities
The distinction between what is inside and outside a system is crucial— Le., the issue
of the boundaries o f Sis cannot be neglected (see Section 7.4.1). It is therefore
necessary to specify the boundaries if empirical studies o f specific Sis— o f a national,
regional, or sectoral kind are to be carried out. As will be discussed later, one way
to identify the boundaries o f Sis is to identify the causes or determ inants of
innovations.
Although national systems of innovation” is only one o f several possible spe­
cifications o f the generic SI concept, It certainly remains one o f the most relevant
 S Y S T E M S OP I N N O V A T I O N 199

O ne reason is the fact that the variou s case studies in N elson (1993) show that there
are sharp differences am o n g vario u s national system s in such attributes as in sti­
tutional set-up, o rgan ization al set-up, investm ents in R & D , and perform ance. For
exam ple, the differences in these respects between D en m ark and Sweden are
rem arkable— in spite o f the fact that these tw o sm all countries in northern Europe
are very sim ilar in m an y other respects (Edquist and Lu ndvall 1993; 5-6 ).
A n oth er reason to focus on national Sis is that m ost public policies influencing
innovation processes o r the econ o m y as a w hole are still designed and im plem ented
at the national level. F o r very large countries, the national SI approach is less relevant
than for sm aller countries, but institutions such as laws and policies are still m ainly
national, even in a co u n try such as the U SA . In other w ords, the im portance o f
national Sis has p artly to do w ith the fact that they capture the im portance o f the
policy aspects o f inn ovation . It is not o n ly a m atter o f geographical delim itation, as
the state, and the po w er attached to it, is also im portant.
Sis m ay be supran ation al, national, or subnational (regional, local)— and at the
same tim e they m ay be sectoral w ith in any o f these geographical d em arcation s.25 A ll
these approaches m ay be fruitful, but for different pu rposes or objects o f study.
Generally, the varian ts o f the generic SI approach com plem ent rather than exclude
each other and it is useful to consider sectoral and regional S is in relation to— and
often as parts o f — n ation al ones.
There are three ways in which we can identify boundaries of Sis:
(1) spatially / geograph ically;
(2) sectorally; and
(3) in term s o f activities.
1. To define the spatial b ou n d aries is the easiest task, although it also has its
problem s. Such b ou n d aries have to be defined for regional and national Sis, and
som etim es also for sectoral ones. The problem o f geographical boun daries is so m e­
what m ore com plicated for a regional than for a national SI. O ne question is w hich
criteria sh ould be used to id en tify a “ region.”
F o r a region al SI, the specification o f the bou n daries sh ou ld not only be a question
o f ch oosin g o r u sing adm inistrative b ou n daries betw een regions in a m echanical
m anner (alth ough this m igh t be useful fro m the p o in t o f view o f availability o f d ata ).
It sh ould also be a m atter o f ch oosing geographical areas for w hich the degree o f
“ coherence” o r “ inw ard o rien tation ” is high w ith regard to in n ovation processes.
O ne possible operatio n alization o f this criterion could be a sufficient level o f
localized learning sp ill-overs (am on g organizations), w hich is often associated
w ith the im p o rtan ce o f transfer o f tacit know ledge am on g (in divid u als and) o rg an ­
izations. A second cou ld be localized m o b ility o f skilled w orkers as carriers o f
know ledge, i.e. an o p eration alization w hich show s that the local lab or m arket is
im portant. A th ird p o ssib ility could be that a m in im u m p ro p o rtio n o f the in n o v ­
ation-related co llab o ration s am o n g organizations should be w ith partners w ithin
200 CHARLES EDQUIST

the region. This is a m atter o f localized netw orks, i.e. the extent to which learning
processes am ong organizations are contained w ith in regions.
For a national SI, the country's borders n o rm ally p ro vid e the boundaries. How­
ever, it could be argued that the criteria fo r regional S is are as valid fo r national ones.
In other w ords, if the degree o f coherence o r inw ard o rien tatio n is very low, the
country m ight not reasonably be considered to have a n ation al SI. It was also
m entioned above that the national SI ap p ro ach is less relevant for large than for
sm aller countries. In Germ any, for exam ple, the ap p ro p riate u n it o f analysis maybe
“ Lander." The choice o f approach m ay not o n ly be a q u estio n o f size o f the country,
but also whether it is federally organized o r not.
2. Leaving the geographical dim ension, w e can also talk about “ sectorally"
delim ited Sis, i.e., systems that include o n ly a p art o f a regional, national or
international system. These are delim ited to specific tech n ological fields (generic
technologies) or product areas. The “ tech nological system s" approach belongs to
this category (although it did not initially use language associated w ith systems of
innovation) (Carlsson and Stankiewicz 1995; 49).
A ccording to Breschi and M alerba, “ a Sectoral In n o va tio n System (SIS) can be
defined as that system (group) o f firm s active in d evelo p in g and m aking a sector s
products and in generating and utilising a se c to rs tech n o lo gies" (Breschi and
M alerba 1997; 131; see also Ch. 14 by M alerba in this v o lu m e ). Specific technologies
or product areas are used to define the boun daries o f sectoral system s, but they must
also norm ally be geographically delim ited ( if they are n o t glo b al). However, it is not
self-evident what a sector is, i.e., the sectoral b ou n daries are p a rtly a theoretical— or
social— construction, which m ay reflect the specific p u rp o se o f the study. It should
also be noted here that the specification o f sectoral b o u n d aries is particularly
difficult with regard to new sectors or sectors goin g th ro u gh radical technological
shifts.
3, W ithin a geographical area (and perhaps also lim ited to a technology field or
product area), the whole socio-econ om ic system cannot, o f cou rse, be considered to
be included in the SI. The question is, then, which parts sh ou ld be included? This is a
m atter o f defining the boundaries o f Sis in term s o f activities. These have to be
defined for all kinds o f Sis: national, regional, and sectoral. T h is is m ore complicated
than in the cases o f spatial and sectoral boundaries.
Early w ork in the SI approach did not address the activities in S is in a systematic
way, and therefore failed to provide clear guidance as to w h at sh ou ld be included in a
system of innovation. N or have the boundaries o f the system s in term s o f activities
been defined in an operational w ay since then.
In Section 7.2, a system of innovation was defined as in clu d in g “ all important
econom ic, social, political, organizational, institution al, and other factors that
influence the developm ent, diffusion, and use o f in n o vation s." I f the concept of
innovations has been specified (e.g., as in the beginn in g o f Section 7.2), and if we
know the determ inants o f their developm ent, d iffu sion , and use, we will be able to
 S Y S T E M S OF I N N O V A T I O N 201

define the boun daries o f the S is in term s o f activities. T h is is one reason w hy it is so

im portant to id en tify the activities in Sis. Adm ittedly, this is not as easy in practice as
in principle, since we sim p ly do not kn ow in detail and system atically all the
activities in S is o r determ inants o f innovation processes. As pointed out in Section
7*4*3*T any list o f activities in an SI m ust be treated as provision al and subject to
change as o u r know ledge increases.

7.5 R e s e a r c h G a p s a n d Opportunities

In innovation studies, there has trad ition ally been a tendency to focus m uch m ore
on technological process in n ovation s and goods produ ct innovations than on
organizational process in n ovation s and service prod u ct innovations. There are
strong reasons to use a com prehensive innovation concept and give m ore attention
to non -tech n ological and intangible kinds o f in n ovation (as proposed in Section
7.3.2). Such an o rien tation is im plicit in the fact that we talk about system s o f
innovation and not system s o f technological change.
M ore research sh ou ld be done on the activities in Sis, i.e., on the determ inants o f
the developm ent, d iffu sio n , and use o f innovations. O ne particu lar task m ay be to
revise and restructure the p relim in ary list o f im portan t activities in S is presented in
Section 7.4.3л. Su ch a list can p ro vid e an im po rtan t point o f en try for em pirical
innovation studies.
A stronger focus on activities w o u ld increase o u r know ledge of, and capacity for,
explaining in n o vatio n processes. G iven o u r lim ited system atic know ledge about
determ inants o f in n o vation s, case studies o f the determ inants o f specific innovations
o r specific (and n arrow ) categories o f in n ovation s w o u ld be very useful. In p articu ­
lar, I believe that comparative case studies have great potential, co m p arin g in n o v­
ations system s o f v ario u s kinds as w ell as the determ inants o f in n o vation processes
w ithin them . Relevant questions to ask w ou ld include: W hich activities o f w hich
organizations are im p o rtan t fo r the developm ent, d iffu sion , o r use o f specific
innovations? Is it possible to distinguish between im p o rtan t activities and less
im p o rtan t ones? W h ich institu tion al rules influence the organizations in carryin g
out these activities? Such w o rk cou ld fu rth er develop the SI approach and contribute
to the creation o f p artial theories about relations between variables w ith in Sis. Such
theories w o u ld also im p ro ve o u r ab ility to sp ecify the bou n daries o f in n ovation
system s.
In this chapter, I have accounted fo r many o f the existing empirical studies that
claim to have been carried out w ith in a SI fram ew ork. T h e result has, on the whole,
202 CHARLES EDQUIST

been rather d isappointing in the sense that m an y o f the studies cited have not been
related to the SI approach in a p ro fo u n d way, alth o u gh there are exceptions. The SI
approach has often been used m ore as a label than as an analytical tool. It has not
influenced the em pirical studies in depth; fo r exam p le, it has not been used to
form ulate hypotheses to be confronted to em p irical o bservation s. T his has made a
virtuous fertilization between conceptual and em p irical w o rk , that is so important
to scholarly progress in this and other fields o f research, d ifficu lt to achieve. The state
o f the art o f the SI approach is partly responsible fo r this: it is often presented in too
vague and unclear a way.
C learly defined concepts are necessary in o rd er to id en tify em pirical correspond­
ents to theoretical constructs and to identify the data that should be collected.
Therefore, conceptual specifications are crucial fo r em p irical studies and it is
im portant to increase the rigor and sp ecificity o f the SI approach. This can be
done by clarifying the m eaning o f key concepts such as innovation, function,
activities, com ponents, organizations, and in stitu tion s, as w ell as the relations
am ong them. M ovin g in this direction does not m ean tran sfo rm in g social science
into som ething sim ilar to natural science. Fo r exam ple, on e cannot abstract from
tim e and space, since there are no universal laws in the social sciences. It is also
Im portant to continue the w o rk o f specifying the b o u n d aries o f S is o f various kinds.
There are strong reasons to integrate conceptual and theoretical work with
em pirical studies in an effort to identify d eterm inants o f the development,
diffusion, and use o f innovations. Such integration can be expected to lead to
cross-fertilization— just as in the case o f w o rk on in teractive learning referred to
in Section 7.4A* The SI approach should be used as a conceptual framework in
specific em pirical analyses o f concrete conditions. Testable statem ents o r hypotheses
should be form ulated on the basis o f the app roach and these sh ou ld be investigated
em pirically, b y using qualitative as w ell as quantitative observation s. Theoretically
based em pirical w o rk is the best w ay to straighten up the SI app roach conceptually
and theoretically; the em pirical w ork w ill, in this way, serve as a “ d isciplining” device
in an effort to develop the conceptual and theoretical fram ew o rk. Such work would
increase o u r em pirical knowledge about relations betw een the m ain function,
activities, organizations, and institutions in Sis. T h is kn ow led ge could then be a
basis for further em pirical generalizations to develop the fram ew ork— including
theory elements. In other words, em pirically based theoretical w o rk is also very
fruitful. Independently o f where one starts, the im p o rtan t th in g is that there should
be a close relationship between theoretical and em pirical w ork.
The array o f determ inants o f innovations and the relations am o n g them can be
expected to vary over time and space, i.e. between in n o vatio n system s, as well as
am ong different categories o f innovation. For exam ple, the determ inants will
probably vary between process and product in n o vation s as w ell as between incre­
m ental and radical innovations (and between subcategories o f these). It is therefore
im portant to pursue the explanatory w ork at a m eso- o r m icro -level o f aggregation.
 S Y S T E M S OF I N N O V A T I O N 203

Taxonomies o f different categories o f innovations can therefore be expected to be an

im portant basis fo r this w ork.
In novation studies have trad ition ally included research on R & D and its signifi­
cance for in n o vation processes. A w ell-educated labor force is necessary for both
R & D and in n ovation , and com petence bu ildin g therefore should receive greater
em phasis in in n o vation studies and in the SI approach. We should not only address
those learning processes that lead directly to process and product innovations, but
also address the know ledge infrastructure and learning in a m ore generic way.
This “ w id en in g ” m igh t eventually transcend the SI approach and m ove into
thinking along the lines o f “ System s o f Learn in g” rather than “ System s o f In n o v ­
ation.” System s o f L earn in g w ould include in divid u al learning (leading to the
creation o f h u m an capital) as well as organizational learning (leading to the creation
o f structural capital, e.g. innovations). It w ould include w ork on three kinds o f
learning: R & D , in n o vation , and com petence building, and, above all, the relations
between them . T h is also points out one direction in w hich the SI approach is
currently developing.

N o t e s

1. For comments on previous versions o f this chapter, I want to thank the editors o f this
Handbook, my discussant at the Lisbon workshop (John Cantwell), and my discussants at
the Roermond workshop (Jan Fagerberg, Bill Lazonick, and Rikard Stankiewicz). I have
also greatly profited from comments by other participants in the workshops and from
Pierre Bitard, Susana Borras, David Doloreux, Leif Hommen, Bjorn Johnson, Rachel
Parker, Lars Mjoset, and Annika Rickne. I also want to thank The Swedish Agency for
Innovation Systems (VINNOVA) for supporting my work with this chapter. However,
I remain responsible for the contents.
2. The regional and sectoral versions are dealt with in more detail in Ch. 11 by Asheim and
Gertler and in Ch. 14 by Malerba in this volume.
3. In this sense, this chapter is a continuation along the same trajectory as earlier attempts,
e.g. Edquist (1997b), Edquist and Johnson (1997) and Edquist (2001).
4. “ [T]he orientation o f this project has been to carefully describe and compare, and try to
understand, rather than to theorise first and then attempt to prove or calibrate the theory”
(Nelson and Rosenberg (1993: 4).
5. They mention organizations such as firms, industrial research laboratories, research
universities, and government laboratories.
6. Their definitions o f NSIs do not include, for example, consequences o f innovation—
which does not, o f course, exclude the fact that innovations, emerging in innovation
systems, have tremendously important consequences for socio-economic variables such
as productivity growth and employment. To distinguish between determinants and
consequences does not, o f course, exclude feedback mechanisms.
7. It should also be mentioned that the publications mentioned in Section 7,2 by no
means exhaust the stock o f literature addressing or using the SI approach. Edquist and
204 CHARLES EDQUIST

McKelvey (2000) is a reference collection containing forty-two articles on Sis, some of

which are reviewed in this chapter. Other contributions will be addressed in later sections
of this chapter.
8. There are chapters in the Handbook on service product innovations (Ch. 16 by Ian Miles)
and on organizational process innovations (Ch. 5 by Alice Lam).
9. See Lundvall et al. (2002: 221) and Lundvall (2003: 9), where it is argued that the
pragmatic and flexible character o f the concept may be seen to be a great advantage.
However, Lundvall et a l (2002: 221) also argue that efforts should be made to give the
concept a stronger theoretical foundation.
10. Such a view has, for example, been expressed by the OECD: “ There are still concerns in
the policy making community that the NIS approach has too little operational value and
is difficult to implement” (OECD 2002a: 11). A similar position is taken by Fischer (2001:
213-14).
11. Only in exceptional cases is the system closed in the sense that it has nothing to do with
the rest o f the world (or because it encompasses the whole world). Like the SI approach,
“general systems theory” might rather be considered to be an approach than a theory,
12. Although there are other kinds o f actors than organizations— e.g. individuals—the
terms “ organizations” and “actors” are used interchangeably in this chapter.
13. The five activities are R8cD, implementation, end-use, education, and linkage,
14. These are: to create new knowledge, to guide the direction o f the search process, to
supply resources, to create positive external economies, and to facilitate the formation of
markets (Johnson and Jacobsson 2003: 3-4). Anna Johnson— now Anna Bergek—
previously discussed these issues in Johnson (1998). There she identified functions
mentioned or implicitly addressed in various previous contributions to the development
of the SI approach. She also listed and stressed various benefits o f using the concept of
“ function” in SI studies.
15. These functions are to create human capital, to create and diffuse technological oppor­
tunities, to create and diffuse products, to incubate (provide facilities, equipment, and
administrative support), to facilitate regulation that may enlarge the market and facili­
tate market access, to legitimize technology and firms, to create markets and diffuse
market knowledge, to enhance networking, to direct technology, market and partner
search, to facilitate financing, and to create a labor market that the new technology-based
firms can utilize,
16. The activities in Sis are the same as the determinants o f the main function. An alternative
term to activities could have been “ subfunctions.” I chose “ activities” in order to avoid
the connotation with functionalism” or “ functional analysis” as practiced in sociology,
which focuses on the consequences of a phenomenon rather than on its causes, which are
in focus here,
17. The activities in this sectoral system are discussed in Edquist (2003:11).
18. Specific ways in which knowledge transfer takes place between universities and firms are
analyzed in detail for the case of Austria in Schibany and Schartinger (2001).
19. When designing the anthology edited by Lundvall (1992), the Aalborg group planned to
have a chapter on the education system. However, in the end it was not included
(Lundvall and Christensen (1999: 3).
20. Competence building has also been addressed in some OECD publications, including a
study on knowledge management in the learning society, managed by the Centre for
Educational Research and Innovation (CERI) (OECD 2000). Another CERI study
 includes a conceptual framework which tries to integrate “ individual learning” (e.g.
education) and organizational learning” (e.g. innovation) into a generic conceptual
framework on learning. It also contains empirical studies of the respective roles of
education and innovation for economic growth at a regional level (OECD 200 id).
Another contribution is the DISKO project in Denmark as reported by Lundvall (2002).
21. In Rickne 2000: ch. 7, there is a more detailed discussion of the relations between
activities and organizations.
22. As we saw in Section 7.4,3л, there are also important relations between activities, he.
relations between determinants o f innovations processes.
23. Coriat and Weinstein address the relations between organizations and institutions as
well, although they consider firms to be both institutions and organizations (Coriat and
Weinstein 2002: 279).
24. The so-called “ varieties o f capitalism” literature has a wider perspective and focuses on a
broader range o f institutions and organizations. Examples are Hollingsworth and Boyer
(1997), Whitley (1999), Hall and Soskice (2001), and Whitley (2002). Space limitations
prevent me from going into this literature here. The same applies to “ the social systems o f
innovation” perspective (e.g. Amable 2000) and the Triple-Helix perspective (e.g.
Etzkowitz and LeydesdorfF (2000).
25. An “ industrial complex” or “cluster” as used by Porter (1990,1998) can, if it is regionally
delimited, be seen as a combination o f a sectoral and a regional SI.

R eferen ces

A m able , B. (2000), “ Institutional Complementarity and Diversity o f Social Systems o f

Innovation and Production,” Review of International Political Economy 7(4): 645-87.
A sh eim , B., and I sa k se n , A. (2002), “ Regional Innovation Systems: The Integration o f Local
"Sticky’ and Global "Ubiquitous’ Knowledge,” Journal of Technology Transfer 27: 77-86.
*B ra czyk , H.-J., C ooke , P., and H e id e n r e ic h , M. (eds.) (1998), Regional Innovation
Systems: The Role of Governance in a Globalised Worlds London and Pennsylvania: UCL,
*B resc h i , S„ and M a l e r b a , F. (1997), “ Sectoral Innovation Systems: Technological
Regimes, Schumpeterian Dynamics, and Spatial Boundaries,” in Edquist, 19 97a: 130-56.
*C arlsso n , B. (ed.) (1995), Technological Systems and Economic Performance: The Case of
Factory Automation, Dordrecht: Kluwer.
------and S t a n k ie w ic z , R. (1995), “ On the Nature, Function and Composition o f Techno­
logical Systems,” in Corlson 1995: 21-56.
C a r r , E. H. (1986), What is History? Harmondsworth: Penguin.
C h r isten sen , J, L., R o g aczew ska , A. L., and V in d in g ,A. L., (1999) Summary Report of the
Focus Group on Innovative Firm Networks, OECD home page.
C ooke , R (2001), “ Regional Innovation Systems, Clusters, and the Knowledge Economy,”
Industrial and Corporate Change 10(4): 945-74.
*------ G om ez U r a n g a , M., and E t x e b a r r ia , G. (1997). "Regional Systems o f Innovation:
Institutional and Organisational Dimensions,” Research Policy 26: 475-9**-

* Asterisked items are suggestions for further reading.

20 6 C H A R LE S EDQUIST

C o r ia t , B., and W ein st e in , O, (2002), “ Organisations, Firms and Institutions in the

Generation o f Innovation ” Research Policy 31(2): 273-90.
*E dquist , G (ed.) (1997a), Systems of Innovation: Technologies, Institutions and Organiza­
tions, London: Pinter.
____ (1997b), “ Systems o f Innovation Approaches their Emergence and Characteristics” in
Edquist 1997a: 1-35* (The book is out o f print, but this chapter has been republished in
Edquist and McKelvey 2000.)
____ (2001), “ The Systems o f Innovation Approach and Innovation Policy: An Account of
the State o f the Art,” Lead paper presented at the DRUID Conference, Aalborg, June 12-15
2001. Unpublished
____(2003), “ The Fixed Internet and Mobile Telecommunications Sectoral System of In­
novation: Equipment, Access and Content,” in C. Edquist (ed,), The Internet and Mobile
Telecommunications System of Innovation: Developments in Equipment, Access and Content,
Cheltenham: Edward Elgar, 1-39.
------ E ricsson , M.-L., and S jogren , H. (2000), “ Collaboration in Product Innovation in the
East Gothia Regional System o fln n o v a tio n Enterprise & Innovation Management Studies, 1.
------ H ommen , L„ and M c K e lv e y , M. (2001), Innovation and Employment: Process versus
Product Innovation, Cheltenham: Edward Elgar.
------and J ohnson , B. (1997), “ Institutions and Organisations in Systems oflnnovation,” in
Edquist 1997a: 41-63. (The book is out o f print, but this chapter has been republished in
Edquist and McKelvey 2000.)
------ and L und vall , B.-A. (1993), “ Comparing the Danish and Swedish systems of innov­
ation,” in Nelson 1993: 265-98.
------ and M c K e lv ey , M. (eds.) (2000), Systems of lnnovation: Growth Competitiveness and
Employment, Cheltenham: Edward Elgar.
Etzko w itz , H., and L eydesdo rff , L. (2000), “ The Dynamics oflnnovation: From National
Systems and "Mode z7 to Triple Helix o f University-Industry-Government Relations,”
Research Policy 29:109-23.
F isc h er , M. E (2001), “ Innovation, Knowledge Creation and Systems o f Innovation”
Regional Science 35:199-216.
F reem an , C. (1987), Technology Policy and Economic Performance: Lessons from Japan,
London: Pinter.
* (2002), Continental, National and Sub-national Innovation Systems— Complemen­
tarity and Economic Growth,” Research Policy 31(2): 191—211.
F urm an , J. L , P orter , M. E., and S t er n , S. (2002), “ The Determinants of National
Innovative Capacity,” Research Policy 31: 899—933.
G a l l i , R., and T eu ba l , M. (1997), “ Paradigmatic Shifts in National Innovation Systems,” in
Edquist 1997a: 342-70,
H a ll , P, A., and S o skice , D. (eds.) (2001), Varieties of Capitalism. The Institutional Founda­
tions of Comparative Advantage, Oxford: Oxford University Press.
H ollingsw orth , J. R„ and B o yer , R., (eds.) (1997), Contemporary Capitalism: The Embed­
dedness of Institutions, Cambridge: Cambridge University Press.
I n gelstam , L. (2002), System—att tanka over samhalle och teknik (Systems: To Reflect over
Society and Technology— in Swedish), Energimyndighetens forlag.
Jo h n s o n , A . (1998), “ Functions in Innovation System Approaches” Mimeo, Department of
Industrial Dynamics, Chalmers University o f Technology.
 S Y S T E M S OF I N N O V A T I O N 207

* —and J aco bsso n , S. (2003), The Emergence o f a Growth Industry: A Comparative

Analysis o f the German, Dutch and Swedish Wind Turbine Industries,” in S. Metcalfe and
U. Cantner (eds,), Transformation and Development: Schumpeterian Perspectives, Heidel­
berg: Physica/Springer.
Laredo , P., and M u st a r , P. (eds.) (2001a), Research and Innovation Policies in the New Global
Economy: an International Comparative Analysis, Cheltenham: Edward Elgar.
-------------(2001b), General Introduction: A Focus on Research and Innovation Policies,” in
Laredo and Mustar 2001a: 1-13.
-------------(2001c), General Conclusion: Three major Trends in Research and Innovation
P o lic ie s in Laredo and Mustar 2001a: 497-509.
*Liu, X. and W h it e , S. (2001), “ Comparing Innovation Systems: A Framework and Appli­
cation to Chinas Transitional Context” Research Policy30:1091-114.
*L und vall , B.-A. (ed.) (1992), National Systems of Innovation: Towards a Theory of Innov­
ation and Interactive Learning, London: Pinter.
------(2002), Innovation, Growth and Cohesion: The Danish Model, Cheltenham: Edward
Elgar.
------(2003), “ National Innovation Systems: History and Theory,” in Elgar Companion to
Neo-Schumpeterian Economics, Cheltenham: Edward Elgar.
------J ohnson , B „ A n d e r se n , E. S., and D a lu m , B. (2002), “ National Systems o f Produc­
tion, Innovation and Competence Building,” Research Policy 31(2): 213-31.
------and L in d g a a rd C h r ist e n se n , J. (1999), “ Extending and Deepening the Analysis of
Innovation Systems— with Empirical Illustrations from the DISKO-project,” Paper for the
DRUID Conference on National Innovation Systems, Industrial Dynamics and Innov­
ation Policy, Rebild, 9-12 June 1999.
*N elson , R. R. (ed.) (1993), National Systems of Innovation: A Comparative Study; Oxford:
Oxford University Press.
------ and R o sen b er g , N. (1993), “ Technical Innovation and National Systems,” in Nelson
1993: 3“ 21.
N orth , D. C. (1990), Institutions, Institutional Change and Economic Performance, Cam ­
bridge: Cambridge University Press.
OECD (1996), Lifelong Learning for All, Paris: OECD.
------(199Sa), Technology>Productivity and Job Creation: Best Policy Practices, Paris: OECD.
------ (1998b), Pathways and Participation in Vocational and Technical Education and Training,
Paris: OECD.
------(1999a), Managing National Innovation Systems, Paris: OECD.
------ (1999b), Boosting Innovation: The Cluster Approach, Paris: OECD.
------(2000), Knowledge Management in the Learning Society, Paris: OECD (CERI).
------(2001a), Innovative Clusters: Drivers of National Innovation Systems, Paris: OECD.
------ (2001b), Innovative Networks: Co-operation in National Innovation Systems, Paris:
OECD.
------(2001c), Innovative People: Mobility of Skilled Personnel in National Innovation Systems,
Paris: OECD.
------ (looid), Cities and Regions in the N e w Learning Economy, written by C. Edquist, G. Rees,
K. Larsen, M. Lorenzen, and S. Vincent-Lancrin, Paris: OECD (CERI).
------(2002a), Dynamising National Innovation Systems, Paris: OECD.
------(2002b), Main Science and Technology Indicators, vol. 2002/2, Paris: OECD.
208 CHARLES EDQUIST

O r s t a v ik , E, and NAs, S.-O. (1998), “ The Norwegian Innovation— Collaboration Survey,

Oslo, the STEP Group STEP Working Paper A-10.
Po rter , M. E. (1990), The Competitive Advantage of Nations, New York: Free Press.
------ (1998), “ Clusters and the New Economics o f Competition,” Harvard Business Review
77-90.
R ic k n e , A. (2000), New Technology-Based Firms and Industrial Dynamics: Evidence from the
Technological Systems of Biomaterials in Sweden, Ohio and Massachusetts, Department of
Industrial Dynamics, Chalmers University o f Technology.
S ch iba n y , A., and Sch artin g er , D. (2001), “ Interactions between Universities and Enter­
prises in Austria: An Empirical Analysis at the Micro and Sector Levels,” in OECD 2001k
235—52.
Sm ith , K. (2001), “ Human Resources, Mobility and the Systems Approach to Innovation,” in
OECD 2001c: ch. 1.
W h it le y , R. (1999), Divergent Capitalisms: The Social Structuring and Change of Business
Systems, Oxford: Oxford University Press.
(ed.) (2002), Competing Capitalisms: Institutions and Economics. An Elgar Reference
Collection, 2 vols., Cheltenham: Edward Elgar.
 C H A P T E R 8

UNIVERSITIES
IN N A T I O N A L
INNOVATION
SYSTEMS

D A V I D C. M O W E R Y
B H A V E N N. S A M P A T

8.1 In t r o d u c t io n

T he research university plays an important role as a source o f fundamental know­

ledge and, occasionally, industrially relevant technology in modem knowledge-
based economies. In recognition o f this fact, governments throughout the industri­
alized world have launched numerous initiatives since the 1970s to link universities
to industrial innovation more closely. Many o f these initiatives seek to spur local
economic development based on university research, e.g., by creating “ science
parks” located nearby research university campuses, support for “business incuba­
tors” and public “seed capital” funds, and the organization of other forms of
“bridging institutions” that are believed to link universities to industrial innovation.
Other efforts are modeled on a US law, the Bayh-Dole Act o f 1980, that is widely (if
210 D A V I D C. M O W E R Y A N D B H A V E N N. S A M P A T

perhaps incorrectly) credited w ith im p ro vin g u n iv e rsity -in d u stry collaboration and
technology transfer in the U S national in n ovation system .
This chapter exam ines the roles o f universities in in d u strial-eco n o m y national
innovation system s, the com plex institutional landscapes that influence the cre­
ation, developm ent, and dissem ination o f in n o vation s (fo r fu rth er discussion see
Edquist, Ch. 7 in this volum e). The inclusion o f a chapter on u n iversity research in a
volum e on innovation is itself an in n ovation — it is likely that a sim ilar handbook
published tw o decades ago w ou ld have devoted far less attention to the role of
universities in industrial in n o vatio n .1 But scholarship on the role o f universities in
the innovation process, as opposed to their role in basic research, has grow n rapidly
since 1970. One im portant them e in this research is the reconceptualization of
universities as im portant institutional actors in n ation al and regional systems of
innovation. Rather than “ ivo ry tow ers” devoted to the p u rsu it o f knowledge for its
ow n sake, a grow ing num ber o f in d u strial-econ o m y and developing-econom y
governm ents seek to use universities as instrum ents for know ledge-based economic
developm ent and change.
Governm ents have sought to increase the rate o f tran sfer o f academ ic research
advances to indu stry and to facilitate the app lication o f these research advances by
dom estic firm s since the 1970s as part o f b road er efforts to im prove national
econom ic perform ance. In the “ know ledge-based e c o n o m y ” according to this
view, national systems o f higher education can be a strategic asset, if links
w ith industry are strengthened and the transfer o f tech n o lo gy enhanced and
accelerated. M any if not m ost o f these “ tech n o logy-tran sfer” initiatives focus on
the codification o f p ro p erty rights to in d ivid u al inventions, and rarely address the
broader m atrix o f in d u stry-u n iversity relationships that span a broad range of
activities and outputs.
U niversities throughout the O E C D also have been affected b y tighter constraints
on public funding since 1970. G row th in pu blic fu n d in g fo r h igh er education has
slowed in a num ber o f O E C D m em ber states. In the U n ited States, Cohen et al
(1 998) note that federal research funding per fu ll-tim e academ ic researcher declined
by 9,4 per cent in real term s during i979“ 9L in the face o f significant upw ard pressure
on the costs o f conducting state-of-the-art research in m an y fields o f the physical
sciences and engineering. Financial su p port from state govern m en ts for US public
universities operating budgets (which o b v io u sly include m o re than research)
declined from nearly 46 per cent o f total revenues in 1980 to slightly more than
40 per cent in 1991 (Slaughter and Leslie 1997: Table 3.2), w hile the share o f federal
funds in U S public university operating budgets declined fro m 12.8 to 10 per cent
during the sam e period (the share o f operating revenues d erived from tuition and
fees rose from 12.9 to 15 per cent). The U K govern m en t reduced its institutional
f unding o f universities (as opposed to targeted, com petitive p ro gram s for research)
during the 1980s and 1990s, as did the governm ent o f A ustralia (Slaughter and
Leslie 1997).
 U N I V E R S I T I E S IN N A T IO N AL INN O V AT IO N S YS T EM S 211

Faced w ith slow er grow th in overall public funding, increased com petition for
research fu n d in g, and con tin u in g cost pressures w ithin their operating budgets
during the past tw o decades, at least som e universities have becom e m ore aggressive
and “ e n tre p re n eu ria l" in seeking new sources o f funding. U niversity presidents and
vice-chancellors have p ro m oted the regional and national econom ic benefits flo w ­
ing from academ ic research and have sought closer links w ith indu stry as a m eans o f
expanding research support.
Both internal and external factors thus have led m any nations" universities to
prom ote stron ger linkages w ith in d u stry as a m eans o f publicizing and/or
strengthening their contribu tions to in n ovation and econom ic grow th. In som e
cases, these initiatives b u ild on lo n g histories o f collaboration between u niversity
and in d u stry researchers that reflect unique structural features o f national u niversity
system s and their in du strial environm ent. In other cases, however, these initiatives
are based on a m isu n d erstan d in g o f the roles played b y universities in national
innovation system s, as well as the factors that u nderpin their contributions to
industrial in n ovation .
A lth ough universities fu lfill b roadly sim ilar fu nctions in the in n ovation system s
o f m ost indu strial and indu strializing econom ies, the im portance o f their role varies
considerably, and is influenced b y the structure o f d om estic industry, the size and
structure o f oth er p u b licly fu nded research perform ers, and nu m erou s other factors.
Follow ing a discu ssion o f the (lim ited) evidence on the contrasting im portance o f
universities w ith in R8cD perfo rm an ce and em ploym ent in national innovation
systems, we exam in e other evidence on the contribu tions o f universities to industrial
innovation. Based on this discussion, we critically exam ine recent initiatives b y
governm ents in a n u m b er o f O E C D nations to enhance the contributions o f
universities to in n o vatio n and econom ic grow th. We conclude w ith a discussion
o f the b road agenda fo r future research.

8.2 W h a t F u n c t io n s d o U n iv e r s it ie s
P e r f o r m w i t h i n N a t io n a l
In n o v a t io n S y s t e m s ?

In v aryin g degrees, universities th ro u gh o u t the O E C D now com bine the fu n ction s o f

education and research. T h is jo in t p ro d u ction o f trained personnel and advanced
research m a y be m o re effective than specialization in one o r the other activity.2
For exam ple, the m o vem en t o f trained personnel into industrial and other o ccu pa­
tions can be a p o w erfu l m echan ism for the d iffu sion o f scientific research, and
212 D A V I D C, M O W E R Y A N D B H A V E N N . S A M P A T

dem ands from students and their prospective em ployers fo r relevance in the
curriculum can strengthen links between the acad em ic research agenda and
the needs o f society.
The econom ically im portant "o u tp u ts” o f u n iversity research have come in
different form s, varyin g over tim e and across in d u stries.3 T h ey include, among
others; scientific and technological in fo rm atio n 4 (w h ich can increase the efficiency
o f applied R & D in indu stry by guiding research tow ards m o re fru itfu l departures),
equipm ent and instrum entation5 (used b y firm s in their p ro d u ctio n processes or
their research), skills or hum an capital (em b o d ied in students and faculty
m em bers), networks o f scientific and tech nological capabilities (which facilitate
the diffusion o f new know ledge), and prototypes fo r new p ro d u cts and processes.6
Universities are w idely cited as critical in stitu tion al actors in national innovation
system s (see Nelson 1993; Edquist, C h. 7 in this vo lu m e, an d n u m erou s other works).
As Edquist notes in his chapter, the precise d efin ition o f “ national innovation
system s” rem ains som ew hat hazy, but m ost o f the large literature on the topic
defines them as the institutions and actors that affect the creation, development,
and diffusion o f innovations. The literature on n ation al in n o vatio n systems empha­
sizes the im portance o f strong linkages am on g these v a rio u s institution s in improv­
ing national innovative and com petitive perform an ce, and this em phasis applies in
particular to universities w ithin national in n o vatio n system s.7 The “ national”
innovation system s o f the industrial econom ies app ear m o re and m o re interdepend­
ent, reflecting rapid grow th during the post-1945 p erio d in cross-border flows of
capital, goods, people, and knowledge. Yet the u n iversity system s o f these economies
retain strong “ national” characteristics, reflecting sign ifican t contrasts among na­
tional university system s in structure, and the influence o f historical evolution on
contem porary structure and policy.
O ne influential conceptualization o f the role o f academ ic research w ithin national
innovation systems and econom ies was the so-called “ lin ear m o d el” o f innovation
w idely associated w ith Vannevar Bush and his fam ou s “ b lu ep rin t” fo r the US post-
1945 R8cD system, Science: The Endless Frontier. B ush argu ed for expanded public
funding for basic research w ithin U S universities as a critical contributor to eco­
nom ic growth, and argued that universities w ere the m ost app ro p riate institutional
locus for basic research. This “ linear m o d el” o f the in n o vatio n process asserted that
funding o f basic research was both necessary and sufficient to pro m ote innovation.
Bush s argum ent anticipated parts o f the “ m arket failu re” ration ale for the funding
o f basic academ ic research subsequently developed b y N elson (1959) and Arrow
(1962), This portrayal o f the innovation process has been w id ely criticized (see Kline
and Rosenberg 1986, for one such rebuttal o f the lin ear m o d el). M any US policy
m akers during the 1970s and 1980s cited the Japanese eco n o m y as evidence that basic
research m ay not be necessary or sufficient for a nation to im p ro ve its innovative
perform ance.
 U N I V E R S I T I E S IN N A T I O N A L I N N O V A T I O N S Y S T E M S 213

Yet another view o f the role o f university research focuses on the contrasting
“ norm s o f academ ic and industrial research, M erely contrasting the “ fu ndam ental”
research activities o f academ ics w ith the applied research o f industrial scientists and
engineers obscures as m uch as it illum inates— after all, there are abundant exam ples
o f university researchers m aking im portan t contributions to technology develop­
ment, as well as n u m erou s cases o f im portan t basic research advances in industrial
laboratories. Paul D avid and colleagues (D asgupta and D avid 1994; D avid, Foray,
and Steinm ueller 1999) argue that the norm s o f academ ic research differ significantly
from those observed w ithin industry. For academ ic researchers, professional recog­
nition and advancem en t depend crucially on being first to disclose and publish their
result. P rom pt disclosure o f results and in m ost cases, the m ethods used to achieve
them, therefore is central to academ ic research. Industrial innovation, b y contrast,
relies m o re h eavily on secrecy and lim itation s to the disclosure o f research results.
The significance o f these “ cultural differences” for the conduct and dissem ination o f
research m ay assum e greater significance in the face o f closer links between un iver­
sity and industrial researchers (see below ).
But these contrasts also can be overstated, as D avid et al. (1999) acknow ledge. The
h istory o f science is replete w ith exam ples o f fierce com petitions (“ discovery races” )
between team s o f researchers in a given field that system atically seek to m islead one
another through the d isclosure o f false inform ation. A n d recent research b y H en ­
derson and colleagues (H enderson, O rsenigo, and Pisano 1999; H enderson and
C ockburn 1998) o n ph arm aceu tical in d u stry R8cD highlights the increased em phasis
by a n um ber o f large p h arm aceu tical firm s on pu blication b y industrial researchers
as a m eans o f im p ro vin g their basic science capabilities. N evertheless, the potential
for clashes betw een the disclosure norm s o f academ ia and industry, and in p a rticu ­
lar, the poten tial risks posed b y m ore restrictive disclosure norm s for the educational
functions and the b ro ad er pace o f advance in scientific understanding, rem ains
significant.
Still another conceptual fram ew ork that has been applied recently to descriptions
o f the role o f academ ic research in “ p o st-m o d ern ” industrial societies is the “ M od e
2” concept o f research identified w ith M ichael G ib b on s and colleagues (G ib bon s
et al. 1994). “ M o d e 2” research is associated w ith a m ore interdisciplinary, pluralistic,
“ netw orked” in n o vatio n system , in contrast to the previou s system in w hich m ajo r
corporate o r academ ic research institutions w ere less closely linked w ith other
institutions. G ib b o n s and other scholars argue that the grow th o f “ M ode 2” research
reflects the increased scale and d iversity o f know ledge inputs required fo r scientific
research, a p o in t echoed in the chapter b y Pavitt in this volum e. Increased diversity
in inputs, in this view , is associated w ith greater interinstitutional collaboration and
m ore in terd iscip lin ary research. Because “ M o d e 2” involves the interaction o f m any
m ore co m m u n ities o f researchers and other actors w ithin any given research area,
21 4 D A V I D C. M O W E R Y A N D B H A V E N N . S A M P A T

purely academ ic research norm s m ay prove less in flu en tial even in such areas of
fundam ental research as biom edical research.
The “ M ode 2” fram ew ork assuredly is consistent w ith som e characteristics of
m odern innovation systems, notably the increased in terin stitu tion al collaboration
that has been rem arked upon by n um erous scholars. B u t this fram ew ork's claims
that the sources o f knowledge w ithin m odern in n o vatio n system s have become more
diverse need not im ply any decline in the role o f u niversities as fundam ental research
centers. Several studies (G od in and G ingras 2000; H icks an d H am ilton 1999; see
below for further discussion) su p port the “ M o d e 2 " assertion that cross-
institutional collaboration and diversification in know ledge sources have grown,
but indicate no such decline.
Still another conceptual framework for analyzing the changing position of uni­
versities within national innovation systems is the “Triple Helix” popularized by
Etzkowitz and Leytesdorff (1997)* Like the “Mode 2” framework, the triple helix
emphasizes the increased interaction among these institutional actors in industrial
economies' innovation systems. Etzkowitz and co-authors (Etzkowitz et al. 1998)
further assert that
In addition to linkages am ong institutional spheres, each sphere takes the role o f the other.
Thus, universities assum e entrepreneurial tasks such as m arketing knowledge and creating
com panies even as firm s take on an academ ic dim ension, sharing knowledge among each
other and training at ever-higher skill levels, (p. 6)

The “ triple h elix” scholarship devotes little attention to the “ transform ations” in
industry and governm ent that are asserted to com plem ent those in universities. The
helix's em phasis on a m ore “ industrial” role for universities m a y b e valid, although it
overstates the extent to w hich these activities are o ccu rrin g th ro u gh o u t universities,
rather than in a few fields o f academ ic research. B u t the “ triple h elix ” has yet to yield
m ajor em pirical or research advances, and its value as a guide for future empirical
research appears to be lim ited.
The national systems ” “ M ode 2,” and “ triple h elix ” fram ew o rks for conceptual­
izing the role o f the research university w ith in the in n o vatio n processes o f know­
ledge-based econom ies em phasize the im po rtan ce o f stron g links between
universities and other institutional actors in these econom ies. A n d both “ Mode 2”
and the Triple H elix argue that interactions between universities and industry, in
particular, have grow n. According to the “ Triple H elix ” fram ew ork, increased
interactions are associated with change in the internal cu ltu re an d norm s o f univer­
sities (as noted, this fram ew ork has m uch less to say about the change in the
characteristics of industrial and governm ental research institution s). What is
lacking in all o f these fram ew orks, however, is a clear set o f criteria by which to
assess the strength o f such linkages and a set o f in dicators to guide the collection of
data.
 U N I V E R S I T I E S IN N A T IO N AL INN O V AT IO N S YS T EM S 215

8.3 T h e R o le o f U n iv e r s it ie s in

N a t io n a l In n o v a t io n S y s t e m s :
C r o s s -N a t io n a l D a ta

8.3.1 Comparative Data on the Structure of National Systems

The first universities appeared du ring the M id dle Ages in B ologn a and Paris, and
were autonomous, self-governing institutions recognized by both church and local
governmental authorities.8 This situation persisted through much of the period
prior to the eighteenth century. But the rise of the modern state was associated with
the assertion by governments of greater control over public university systems in
much of continental Europe, notably France and Germany, as well as Japan.9 Such
centralized control was lacking, however, in the British and especially, the US higher
education systems throughout the nineteenth and twentieth centuries. Throughout
the twentieth century, US universities retained great autonomy in their adminis­
trative policies. Rosenberg (1999) and Ben-David (1968) argue that this lack of
central control forced American universities to be more “entrepreneurial” and
their research and curricula to be more responsive to changing socio-economic
demands than their European counterparts. Data allowing for systematic cross­
national comparisons of the structure of the higher educational systems of major
industrial economies are surprisingly scarce.
This section sum m arizes and assesses the lim ited com parative data on the
training and research roles o f h igher educational system s, as w ell as their relation ­
ships w ith industry. E nrollm en t data (sum m arized in G eiger 1986, and G raham
and D iam o n d 1997) indicate that the U S system enrolled a larger fraction o f the
i8-22»year-old p o p u latio n than those o f any E u ropean nations throughout the
19 0 0 -19 45 period . N o t until the 1960s did E uropean enrollm ent rates exceed 10 per
cent o f the relevant age cohorts, by w hich tim e U S enrollm ent rates w ithin this group
were reaching 50 p er cent (B u rn et al. 19 71). These contrasts in enrollm ent rates are
reflected in en d u rin g differences betw een the U nited States and E uropean nations in
the shares o f their po p u latio n s w ith u n iversity education. The share o f the U S
po p u lation w ith u n iversity or “ tertiary” educational degrees exceeded that o f any
other O E C D eco n o m y as late as 1999.
These data also reveal that the U S un iversity degreeholder share is follow ed closely
b y that o f N orw ay, at 25 p er cent (O E C D Edu cation D atabase 2001). Surprisingly,
Austria, w ith 6 per cent o f the relevant po p u lation h oldin g university o r tertiary
degrees, exhibits the low est degreeholder share in this database. A s Fagerberg and
G o d in h o note in C h ap ter 19 o f this volu m e, however, the large output o f university
degreeholders in the U n ited States includes a significantly sm aller share o f n atural
science and engineerin g degreeholders than is true o f such other nations as the
2l6 D A V I D C. M O W E R Y A N D B H A V E N N . S A M P A T

United K ingd om , Singapore, Finland, South K orea, and France. T h e share o f 24-
year-olds in the United States w ith “ first degrees” fro m universities in natural
sciences and engineering also lags well behind these and o th er n a tio n s.10
The lim ited data on the role o f national h igh er ed u catio n system s as R&D
perform ers highlight other cross-national contrasts, in clu d in g differences in their
significance w ithin the overall national R & D enterprise, th eir scale, their roles as
em ployers o f researchers, and their relationships w ith indu stry. A s Figure 8.1 shows,
the role o f universities as R & D perform ers (m easured in term s o f the share of
national R & D perform ed w ithin higher education) is greatest in Italy, the Nether­
lands, and Canada, all o f which sh ow universities p e rfo rm in g m o re than 25 per cent
o f total national R & D b y 1998-2000 (Figure 8.1). T h e share o f national R&D
perform ance accounted for b y U S and Japanese universities, b y contrast, was slightly
m ore than 14 per cent during the sam e period.

-♦—USA
-•-C a n a d a
France
Germany
Italy
Netherlands
— Norway
— Sweden
UK
-•-Japan
-*-EU

Fig. 8.1 Universities' performance share of total national R&D, 1981-99

Source. OECD, Main Science and Technology indicators, 2001.
 U N I V E R S I T I E S IN N A T I O N A L I N N O V A T I O N S Y S T E M S 217

C ro ss-n atio n al data high lighting differences in the “ division o f lab o r” between
universities and governm ent laboratories in basic research indicate that the higher
education sector’s share o f basic research perform ance is sim ilar in m ost W estern
European econom ies and the U nited States, although higher than in m ost o f the
Eastern E uropean and A sian countries for w hich data are available (O E C D 2001b;
A n nex Table A .6 .4 .1). B u t a key difference between the U nited States and m ost
European countries for w hich data are available is that a relatively low share o f
basic research outside the academ ic sector in the United States is perform ed b y the
governm ent, and a relatively high share b y industry. n
The data also reveal considerable variation am ong O E C D m em ber nations in the
scale o f the higher education research enterprise. A lth ough the U S higher education
system is larger in absolute term s than those o f other O E C D m em ber states, U S
universities’ perform an ce o f R & D in fact accounts for a sm aller share o f G D P than is
true o f Sw eden, France, C anad a, the N etherlands, and N o rw ay (Figure 8.2). Indeed,
Figure 8.2 indicates that U S universities’ R & D as a share o f G D P has in fact declined
slightly d u rin g the 1989-99 period. A t least a p o rtio n o f this decline reflects the rapid
growth in indu strially funded R & D perform ed w ithin U S industry, especially d urin g
the 1995-9 period.
C o m p ariso n o f the share o f “ em ployed researchers” in various n ation s’ R & D
system s that w o rk in universities reveals that the U nited States and Jap an rank
very low, reflecting the fact that a m uch higher share o f researchers in both nations

— Sweden
— France
.....N etherlands
Norway
— Japan
—*■Canada
G erm any
— EU
USA
--U K
— - -Italy

Fig. 8.2 R&D performed by the higher education sector as a percentage of GDP
S o u rc e : OECD, M a in S c ie n c e and T ech n o lo g y In d ica to rs, 2001
218 DAVID €. MOWERY AND BHAVEN N. S A MP A T

are em ployed b y in d u stry rather than h igh er ed u cation . In 1997, the last year for
w hich reasonably com plete data are available, 82.5 p e r cent o f researchers were
em ployed b y in d u stry in the United States (O E C D 2 0 0 1c Table 39), significantly
higher than in any other O E C D nation. K orea ranks secon d (68.1 per cent) and
Jap an third (64.6 per cent), w hile the overall average fo r E U countries is m uch lower
(4S.4 per cent).
Figure 8.3 depicts the share o f R8cD fu n d in g w ith in n ation al higher education
system s that is provided by industry. D espite the w id ely rem arked closeness o f US
u n iversity-in d u stry research ties and co llab oration (see R o sen b erg and Nelson 1994;
M o w ery et al. 2004), the share o f R & D in high er ed u catio n that is financed by
in du stry is higher for C anada, G erm any, and the U n ited K in g d o m than for the
U nited States in the late twentieth century.

* 4 - USA
-« -C a n a d a
France
Germany
—'"Ж..Italy
Netherlands
— Norway
—— Sweden
— UK
—4 - Japan
-*~EU

year *

Fig. 8.3 Share of higher education R&D financed by industry, 1991-2000

Source: OECD, Main Science and Technology Indicators, 2001
 U N I V E R S I T I E S IN N A T I O N A L I N N O V A T I O N S Y S T E M S 219

O ther qualitative data from the O E C D 2002 study o f “ scien ce-in d u stry relation ­
ships” (2002: 37) com pare the labor m o b ility and other “ netw ork relationships”
linking universities and in d u stry for A ustria, Belgium , Finland, Germ any, Ireland,
Italy, Sw eden, the U K , the U S, and Japan. “ R8cD consulting w ith firm s by university
researchers” is greater than the EU average (the basis for these characterizations is
not pro vid ed b y the O E C D study) in Austria, Germ any, the U K , U S, and Japan; such
consulting is rated as “ lo w ” in Belgium , Finland, Ireland, and Italy. The annual flow
o f university researchers to industrial em ploym ent, another potentially im portant
channel for know ledge exchange, is significantly higher than the E U average in
Belgium , Finland, G erm any, Sweden, the U K , and the U nited States. Finally, the
“ significance o f n etw orks” linking universities and indu stry is rated as above the EU
average for Finland, G erm any, Sweden, the U K , the U S, and Japan.
Surprisingly, in view o f the frequency w ith w hich the U nited States is cited
ap p ro vin gly fo r the close links between university and industrial researchers, the
evidence that u n iversity^ in d u stry relationships are “ stronger” in the U S than
elsewhere is m ixed: the qualitative data on labor m o b ility support this characteriza­
tion, w hile the data on industrial su p port o f academ ic research do not. A n im portant
gap in research on the role o f universities in national innovation system s and a
corresponding research o p p o rtu n ity is the developm ent o f better quantitative
m easures or indicators o f the scope and im portance o f this role. I f the stereotypical
view o f U S universities as m ore closely linked w ith industrial research and in n o v­
ation is indeed valid (and we believe that it is), it is striking that the available
indicators shed so little light on the dim ensions o f these closer links.
A lth o u gh universities serve sim ilar fu nctions in m ost industrial econom ies, these
indicators suggest that their im portance in training scientists and engineers and in
research perfo rm an ce differs considerably am on g O E C D m em ber nations. These
differences reflect cross-nation al differences in in d u stry structure, especially the
im portance o f such “ h igh -tech n o lo gy” industries as electronics o r in form ation
tech n ology that are h igh ly research-intensive and (at least since the end o f the
C old W ar) rely h eavily on private-sector sources for R8cD finance. In addition , o f
course, the role o f n o n u n iversity pu blic research institutions differs am ong these
econom ies, and is reflected in the contrasts in universities as perform ers o f pu blicly
funded R8cD. These structu ral contrasts are the result o f a lengthy, path-dependent
process o f h istorical developm ent, in w hich institutional evolution interacts w ith
industrial grow th and change.

8.3.2 Recent Trends in University-Industry Linkages

A lth o u gh co m p arative cross-sectional data reveal substantial differences in the
sources o f fu n d in g an d other characteristics o f the national system s o f higher
220 D A V I D C. M O W E R Y A N D S H A V E N N. S A M P A T

education am ong O E C D m em ber states, lo n gitu d in al data reveal an increase in

co-authorship between university and in d u stry researchers in m a n y o f these nations.
A m on g other things, this evidence on increased co -au th o rsh ip m a y indicate some
grow th, rather than decline, in the role o f universities as centers fo r knowledge
production w ithin national innovation system s, the argu m en ts o f the M ode 2”
m odel notw ithstanding. A recent paper b y C alvert and Patel (2002) based on an
exam ination o f slightly m ore than 22,000 papers reveals a threefold increase in
co-authorship between U K in d u stry and u n iversity researchers d u rin g 1981-2000.
Papers co-authored b y industrial and u n iversity researchers expanded from
approxim ately 20 to nearly 47 per cent o f all U K scientific papers published by
industrial researchers during the 19 8 1-2 0 0 0 period. T h e share o f papers with UK
university authors that were co-authored b y indu strial and u n iversity scientists also
grew during this period, from 2.8 per cent in the early 1980s to 4.5 p er cent in 2000,12
C o -au th ored papers in com puter science grew b y m ore than eigh tfold, although the
fields o f chem istry, m edicine, and b io lo gy accounted fo r the largest shares of
co-authored papers (respectively, 20, 20, and 14 per cent).
C alvert and Patel found that the 1981/5-1986/90 p erio d w as characterized by the
m ost rapid growth in such co-authorship. T h is fin d in g is p articu larly interesting
since the 1980s were characterized b y cuts in U K central govern m en t spending on
higher education, and the 1990s were a p erio d o f m o re aggressive governmental
pro m otion o f u n iversity-in d u stry collaboration and tech n o lo g y transfer. In other
w ords, the grow th in co-authorship m easured b y these scholars appears to have
occurred w ithout any specific encouragem ent (beyond fu n d in g cuts) from govern­
m ent policy. The U K universities responsible for the m a jo rity o f the co-authored
papers were am ong the m ost distinguished research universities in Great Britain.
A n oth er study o f co-authorship between u n iversity and in d u stry researchers is
that b y H icks et al. (1995), w hich com pares trends d u rin g the 19 8 0 -9 period in
co-authorship in Japan and W estern Europe. O verall co -au th o rsh ip rates (covering
all industrial sectors and including both dom estic and foreign universities) were
sim ilar (rough ly 20 per cent for E uropean papers an d sligh tly less for Japanese
papers) for Western Europe and Japan in 1980. B y 1989, how ever, co-authorship
rates for W estern Europe had risen to nearly 40 per cent o f pu blish ed papers, while
Japanese co-authorship rates only slightly exceeded 20 per cent.
There is surprisingly little em pirical w o rk on co -au th o rsh ip in the United States.
A study by H icks and H am ilton (1999) reports that betw een 1981 and 1994, the
num ber o f U S papers co-authored by university and in d u stry researchers m ore than
doubled, considerably exceeding the 38 per cent increase in the total num ber of
scientific papers published by US researchers d u rin g this p eriod. The authors also
suggest that these co-authored papers are less “ basic5" than academ ic articles without
industrial co-authors.
O v e ra ll, th e se b ib lio m e tr ic s t u d ie s p re se n t a ric h d e s c r ip t iv e a n d a re lativ e ly weak
e x p la n a t o r y a n a ly sis o f an im p o r t a n t ty p e o f u n iv e r s it y - in d u s t r y c o lla b o r a tio n ,
 U N I V E R S I T I E S IN N A T I O N A L I N N O V A T I O N S Y S T E M S 221

inasm uch as they p ro vid e little explanation for trends or cross-national differences.
N onetheless, these data highlight a broad trend o f grow th in such co-authorship,
and this area rem ains a very fruitful one for future research that spans m ore fields,
nations, and types o f pu blications. The results o f the bibliom etric w ork in this area
provide som e su p p o rt for the “ M ode 2” and “ Triple H elix” fram ew orks' argum ents
that research collaboration between universities and indu stry is grow ing throughout
the industrial econom ies, in university system s w ith very different structures
(see C h apter 7 b y E dquist in this volu m e, as well as the studies in Laredo and
M ustar 2001).

8.4 How U niversity Research

does
A f f e c t In d u s t r i a l In n o v a t i o n ?
A S u m m a r y o f S o m e US S t u d i e s

The quantitative indicators discussed in the previou s section provide som e in fo r­

m ation o n the structure o f universities w ithin the O E C D and their links w ith
national in n o vatio n system s. But these data shed very little light on the characteris­
tics o f the know ledge flow s between university research and the indu strial in n o v­
ation process. T h is issue is especially im portan t in light o f the num erous
governm ent p o licy initiatives that seek to enhance or exploit such know ledge
flow s (see b elow ). A lth o u gh their coverage is lim ited to U S universities and industry,
a n u m ber o f recent studies based on interview s o r surveys o f senior industrial
m anagers in indu stries ran gin g from pharm aceuticals to electrical equipm ent have
exam ined the influence o f u n iversity research on industrial innovation, and thereby
provide ad d ition al insigh t into the role o f universities w ithin the U S national
in n o vation system .
A ll o f these studies ( G U I R R 1991; M ansfield 1991; Levin et al. 1987; C oh en, N elson,
and W alsh 2002) em phasize the significance o f interin du stry differences in the
relationship betw een u n iversity and indu strial innovation. The biom edical sector,
especially b io tech n o lo g y and pharm aceuticals, is unusual, in that university research
advances affect indu strial in n o vation m ore significantly and directly in this field
than in other sectors. In these other technological and indu strial fields, universities
o ccasion ally contribu ted relevant “ inventions," b u t m ost com m ercially significant
inventions cam e fro m n on academ ic research. The increm ental advances that w ere
the p rim a ry fo cu s o f the R & D activities o f firm s in these sectors w ere alm ost
exclusively the d o m ain o f indu strial research, design, pro b lem -so lvin g, and devel­
opm ent. U n iversity research contributed to technological advances b y enhancing
222 D A V I D C. M O W E R Y A N D B H A V E N N . S A M P A T

know ledge o f the fundam ental physics and ch em istry u n d erlyin g manufacturing
processes and product innovation, an area in w h ich train in g o f scientists and
engineers figured prom inently, and experim ental techniques.
The studies hy Levin et al. (1987) and C o h en et al. (2002) sum m arize industrial
R & D m anagers’ view s on the relevance to indu strial in n o va tio n o f variou s fields of
university research (Table 8.1 sum m arizes the results discu ssed in Levin et al. 1987).
V irtually all o f the fields o f university research that w ere rated as “ im portant” or
“ v ery im portant” for their innovative activities b y su rvey respondents in both
studies were related to engineering o r applied sciences. A s we noted above, these
fields o f U S university research frequently d eveloped in close collaboration with
industry. Interestingly, w ith the exception o f chem istry, v e ry few basic sciences
appear on the list o f university research fields deem ed b y in d u stry respondents to
be relevant to their innovative activities.
The absence o f fields such as physics and m athem atics in Table 8.1, however,
should not interpreted as indicating that academ ic research in these fields does not
contribute directly to technical advance in industry. Instead, these results reflect the
fact that the effects on industrial in n ovation o f basic research findings in such areas
as physics, m athem atics, and the physical sciences are realized o n ly after a consider­
able lag. M oreover, application o f academ ic research results m ay require that these
advances be incorporated into the applied sciences, such as chem ical engineering,
electrical engineering and m aterial sciences. The su rvey results summarized in
Cohen et al. (2002) indicate that in m ost industries, u n iversity research results
play little i f any role in triggering new indu strial R & D projects; instead, the stimuli
originate w ith custom ers or from m an u factu rin g operation s. H ere as elsewhere,
pharm aceuticals is an exception, since u n iversity research results in this field often
trigger industrial R & D projects.
Cohen et a l (2002) further report that the results o f “ p u b lic research” performed
in governm ent labs or universities were used m ore frequently b y U S industrial firms
(on average, in 29.3 per cent o f industrial R8cD projects) than prototypes emerging
from these external sources o f research (used in an average o f 8.3 per cent o f indus­
trial R & D projects). A sim ilar portrait o f the relative im p o rtan ce o f different outputs
o f university and public laboratory research em erges fro m the responses to ques­
tions about the im portance to industrial R & D o f vario u s in fo rm atio n channels
(Table 8.2). A lthough pharm aceuticals once again is u n u su al in its assignment of
considerable im portance to patents and license agreem ents in vo lvin g universities
and public laboratories, respondents from this in d u stry still rated research public­
ations and conferences as a m ore im portant source o f in fo rm atio n . For most indus­
tries, patents and licenses involving inventions fro m u n iversity or public
laboratories were reported to be o f very little im portan ce, co m pared with publica­
tions, conferences, inform al interaction with u n iversity researchers, and consulting.
Data on the use by industrial R&D managers of academic research results are
needed for other industrial economies. Nonetheless, the results of these US studies
 U N I V E R S I T I E S IN N A T IO N A L I N N O V A T IO N S Y S T E M S 223

Table 8.1 The relevance of university science to industrial technology

Science Number of Industries Selected industries for which

with "relevance" scores the reported "relevance" of
university research was large

>5 >6 (> 6 )

Biology 12 3 Animal feed, drugs, processed

fruits/vegetables
Chemistry 19 3 Animal feed, meat products,
drugs
Geology 0 0 None
Mathematics 5 1 Optical instruments
Physics 4 2 Optical instruments,
electronics
Agricultural science 17 7 Pesticides, animal feed,
fertilizers, food products
Applied math/operations 16 2 Meat products, logging/
research sawmills

Computer science 34 10 Optical instruments, logging/

sawmills, paper machinery

Materials science 29 8 Synthetic rubber, nonferrous

metals

Medical science 7 3 Surgical/medical instruments,

drugs, coffee

Metallurgy 21 6 Nonferrous metals, fabricated

metal products

Chemical engineering 19 6 Canned foods, fertilizers, malt

beverages

Electrical engineering 22 2 Semiconductors, scientific

instruments

Mechanical engineering 28 9 Hand tools, specialized

industrial machinery

Source: Data from the Yale Survey on Appropriability and Technological Opportunity in Industry. For a
description of the survey, see Levin e t al. (1987).

consisten tly em phasize that the relationship betw een academ ic research and in d u s­
trial in n o vatio n in the b io m ed ical field differs fro m that in other know ledge-
intensive sectors. In ad d ition , these studies suggest that academ ic research rarely
224 D A V I D C. M O W E R Y A N D B H A V E N N . S A M P A T

Table 8 .2 Importance to industrial REtDo f sources o f

RftD (including university research)
------- ------ ,— — ----------- ------ — -------*............ ........-........-■ .1; ^

Information source Rating it as “very important" for industrial REtD '

Publications Et reports ■41.2 ■ ■

Informal Interaction 35.6 '

Meetings Et conferences 35Л ѵ'й |

Consulting 31.8 А';

Contract research 20.9

Recent hires 19.8

Cooperative REtD projects 17.9

Patents 17.5

Licenses 9.5

Personnel exchange 5.8

Source: Cohen et a l. {2002).

produces ""prototypes” o f inventions for d evelopm ent and com m ercialization by

industry— instead, academ ic research in fo rm s the m ethods and disciplines
em ployed by firm s in their R8cD facilities. Finally, the channels rated b y industrial
R & D m anagers as m ost im portant in this co m plex interaction between academic
and industrial innovation rarely include patents and licenses. Perhaps the most
striking aspect o f these survey and interview results is the fact that they have not
inform ed the design o f recent policy initiatives to enhance the contributions of
university research to industrial innovation.

F rom “ S c i e n c e P u s h ” to
8.5
“ T echnology C o m m er c ia liza tio n ”

As we suggested in Section 8.1, since 1980 a num ber o f industrialized countries have
im plem ented or considered policies to strengthen ""linkages” between universities
(and public research organizations) and industry, in order to enhance the con­
tributions o f university-based research to innovation and econom ic performance.
 U N I V E R S I T I E S IN N A T I O N A L I N N O V A T I O N S Y S T E M S 225

These initiatives all share the prem ise that universities su p port innovation in in d u s­
try p rim arily through the p ro d u ction by universities o f “ deliverables” for co m m er­
cialization (e.g., patented d iscoveries)>despite the m odest su pport for this prem ise
in the research discussed above. We illustrate these points in this section with case
studies o f two types o f policies: (1) policies encouraging the form ation o f regional
econom ic “ clusters” and sp in -offs based on university research, and (2) policies
attem pting to stim ulate u n iversity patenting and licensing activities.
The global d iffu sion o f these “ tech nology com m ercialization” policies illustrates a
phenom enon that has received too little attention in the literature on innovation
policy— the efforts b y p o licy m akers to “ b o rro w ” policy instrum ents from other
econom ies and ap p ly these instrum ents in a very different institutional context. As
Lundvall and B o rras p o in t out in their chapter, history, path dependence, and
institutional “ em beddedn ess” all m ake this type o f “ em ulation” very difficult.
N onetheless, such em u latio n has been especially w idespread in the field o f technol­
ogy policy. In ternational p o licy em ulation o f this sort is characterized by tw o key
features: (1) the “ learn in g ” that underpins the em ulation is h igh ly selective; and (2)
the im plem entation o f p ro gram designs based on even this selective learning is
affected b y the different institutional landscape o f the em ulator.

8.5.1 Universities and Regional Economic Development

In m any O E C D countries, efforts to increase the national econom ic returns from
public investm ents in university research have attem pted to stim ulate the creation o f
“ regional clusters” o f innovative firm s around universities. These undertakings seek to
stimulate regional econom ic developm ent and agglom eration via facilitating the
creation o f “ sp in -o ff ’ firm s to com m ercialize university technologies (O E C D 2002).13
These p o licy initiatives are m otivated by the h igh -tech n olo gy regional clusters in
the U nited States, n otab ly Silicon Valley in C alifo rn ia and Route 128 in the Boston
area. B oth o f these h igh -tech n o lo gy clusters have a spaw ned a large n u m ber o f new
firm s and have m ajo r research universities in their m idst (in C alifo rn ia, the U n iver­
sity o f C a lifo rn ia at Berkeley, Stanford U niversity, and the U niversity o f C alifo rn ia at
San Francisco; in B oston , H arvard U n iversity and M IT ). A t least som e o f the
successful new firm s in these regions have been involved in com m ercializing tech­
nologies d eveloped at regional universities.
O ther evidence (notably, Trajtenberg, Jaffe, and H enderson 1997) suggests that the
“ know ledge sp illo vers” fro m u n iversity research w ithin the U nited States, m easured
b y the lo cation o f inventors citing u n iversity patents, tend to be localized at the
regional level. Recent w o rk b y H icks et al. (2001) sim ilarly indicates that patents filed
by U S inventors d isp ro p o rtio n ately cite scientific papers from research institutions
located in the sam e state as these inventors.
226 D A V I D C* M O W E R Y A N D B H A V E N N. S A M P A T

But little evidence su pports the argum ent that the presence o f universities some­
h ow “ causes” the developm ent o f regional h ig h -tech n o lo g y agglom erations. And
even less evidence supports the argum ent that the region al o r in n o vatio n policies of
governm ents are effective in creating these agglom eration s. O ne can point to high-
technology clusters w ith highly produ ctive research universities in a num ber o f areas
in the United States and other industrial econom ies; b u t there are also a number of
research universities that have not spaw ned such agglom eratio n s. M oreover, efforts
to replicate the “ Silicon Valley m o d el” in other econ om ies have proven difficult and
the results o f these efforts have been m ixed (a fascinating h istorical account o f the
efforts by Frederick Term an o f Stanford U n iversity to p ro m o te such “ exports” may
be found in Leslie and Kargon 199b).
N ation al and local governm ents in m an y O E C D cou n tries have attempted to
stim ulate the form ation o f these clusters v ia fu n d in g fo r “ science p ark s” (occasion­
ally also called incubators, technology centers, o r centers o f excellence). Interest­
ingly, there is considerable disagreem ent about exactly w hat a “ science park” is and
what they do; the International A ssociation o f Science Parks characterizes them as
follows:

A Science Park is an organisation managed by specialised professionals, whose main aim is to

increase the wealth of its community by promoting the culture of innovation and the
competitiveness of its associated businesses and knowledge-based institutions... To enable
these goals to be met, a Science Park stimulates and manages the flow of knowledge and
technology amongst universities, R&D institutions, companies and markets; it facilitates the
creation and growth of innovation-based companies through incubation and spin-off
processes; and provides other value-added services together with high quality space and
facilities, (http://www.iaspworld.org/information/definitions.php)

Despite the w idespread interest in science parks, there is little evidence that they
positively affect universities' contributions to in n o vatio n or sp u r regional economic
developm ent. Using data on U S science parks, Felsenstein (1994) finds no evidence
that firm s located on university-based science parks are m o re in n ovative than other
local firm s, and Wallsten (2001) finds that science parks have little effect on regional
econom ic developm ent and rates o f innovation.
The research on science parks” in other indu strial econ om ies is also limited. One
exam ination ol science parks” in the U K (M assey et al. 1992) is dated, but presents
interesting evidence on the characteristics o f nearly 200 firm s in tw enty U K science
parks. 1 he study found that startup firm s represented 2 5 -3 0 per cent o f the tenants in
the science parks surveyed; in the absence o f som e kin d o f “ con trol population,” it is
difficult to reach conclusions about whether startup firm s are over-represented or
under-represented in these U K science parks. Perhaps m ore su rprising was the
study's finding that

formal research links between academic institutions and establishments on science parks
were no more evident than similar links with firms located off-park... Formal research links
 U N I V E R S I T I E S IN N A T I O N A L I N N O V A T I O N S Y S T E M S 227

such as “employment of academics” “sponsoring trials or research,” “ testing and analysis,”

“student project” work and “graduate employment” were fairly similar for park firms and
off-park firms. However, significantly more park firms than off-park firms mentioned
“ informal contacts with academics” and the use of academic facilities such as computers,
libraries or dining facilities as being important. (Massey et al. 1992: 38)

This and other evidence on the results o f governm ent policies to prom ote university-
based regional agglom erations suggests that such policies have a m ixed record o f
success. A n d even successful regional agglom erations m ay require considerable tim e
to em erge. Recent w o rk by Sturgeon (2000) argues that Silicon Valley's h istory as a
center for n ew -firm fo rm atio n and innovation dates back to the early decades o f the
twentieth century, suggesting that m uch o f the region s innovative “ culture" de­
veloped over a m uch longer period o f tim e and predates the ascent to global research
em inence o f Stanford University. Sim ilarly, the N orth C arolin a “ Research Triangle,"
which w as p ro m oted m uch m ore aggressively b y the state governm ent, was estab­
lished in the late 1950s and becam e a center for n ew -firm fo rm atio n and innovation
only in the late 1980s.
Still other w o rk on the developm ent o f Silicon Valley b y Leslie (1993, 2000) and
Saxenian (1988) em phasizes the m assive increase in federal defense spending after
1945 as a catalyst for the fo rm atio n o f new h igh -techn ology firm s in the region. In
this view , the presence o f leading research universities m ay have been necessary, but
was b y no m eans sufficient, to create Silicon Valley du ring the 1950s and 1960s.
Saxenian in particu lar em phasizes the v ery different structure o f British defense
procurem ent policies in explain in g the lack o f sim ilar d ynam ism in the C am b ridge
region.
The links between university research and the emergence of regional high-
technology agglomerations thus are more complex than is implied by the correlation
between the presence of high-technology firms and research universities in a
number of locales. The US experience suggests that the emergence of such agglom­
erations is a matter of contingency, path-dependence, and (most importantly) the
presence of other supporting policies (intentional or otherwise) that may have
little to do with university research or the encouragement of university-industry
linkages.
The policy initiatives in the United States and other OECD economies that seek to
use university research and “science parks" to stimulate regional economic develop­
ment suffer from a deficiency that is common to many of the other recent efforts to
stimulate university-industry linkages in OECD countries, i.e. a lack of attention to
supporting institutions, a focus on “success stories" with little attention to system­
atic evidence on the casual effects of the policies, and a narrow focus on commercial­
ization of university technologies, rather than other more economically important
outputs o f university research. These characteristics are also seen in recent efforts
elsew here w ith in the O E C D to em ulate the B a y h -D o le Act.
228 D A V I D C, M O W E R Y A N D B H A V E N N . S A M P A T

8.5.2 Patenting the Results of Publicly Funded

Academic Research
As we noted above, this increased interest by governments in “Bayh-Dole type”
policies is rooted in motives similar to those underpinning policy initiatives that
seek to create “high-technology3'regional clusters. But the “emulation" of Bayh-
Dole in other industrial economies overlooks the importance and effects on univer­
sity-industry collaboration and technology transfer of the many other institutions
that support these interactions and the commercialization of university technolo­
gies in the United States. In addition, these “emulation" initiatives are based on a
misreading of the empirical evidence on the importance of intellectual property
rights in facilitating the “transfer" and commercialization of university inventions,
as well as a misreading of the evidence on the effects of the Bayh-Dole Act.

8.5.2.1 O rigins o f the B a y h -D o le A ct

A lthough som e U S universities were patenting facu lty in ven tion s as early as the
1920s, few institutions had developed form al patent p olicies p rio r to the late 1940s,
and m any o f these policies em bodied considerable am bivalen ce tow ard patenting.
Public universities were m ore heavily represented in paten tin g than private univer­
sities during the 1925-45 period, both w ithin the top research universities and more
generally.
These characteristics o f university patenting began to change after 1970, as private
universities expanded their share o f US u n iversity patenting, universities generally
expanded their direct role in m anaging patenting and licensing, and the share of
biom edical patents within overall university patentin g increased. Lobbying by US
research universities active in patenting was one o f several factors behind the passage
o f the B a yh -D o le Act in 1980.
The B a yh -D o le Patent and Tradem ark A m en d m en ts A ct o f 1980 provided blanket
perm ission for perform ers o f federally funded research to file for patents on the
results o f such research and to grant licenses for these patents, including exclusive
licenses, to other parties. The Act facilitated university paten tin g and licensing in at
least two ways. First, it replaced a web o f Institutional Patent A greem ents (IPAs) that
had been negotiated between in divid u al universities and federal agencies with a
uniform policy. Second, the Act's provision s expressed C o n gression al support for
the negotiation of exclusive licenses between universities and indu strial firm s for the
results o f federally funded research.
1 he passage of the Bayh-Dole Act was one part o f a broader shift in US policy
toward stronger Intellectual property rights.14 Am ong the most important of these
policy initiatives was the establishment o f the Court o f Appeals for the Federal
Circuit (CAFC) in 1982, Established to serve as the court o f final appeal for patent
cases throughout the federal judiciary, the CAFC soon emerged as a strong
 U N I V E R S I T I E S IN N A T I O N A L I N N O V A T I O N S Y S T E M S 229

cham pion o f patentholder rights. But even before the establishm ent o f the C A F C ,
the 1980 U S Suprem e C o u rt decision in D iam on d v. C h akrab arty upheld the valid ity
o f a b road patent in the new indu stry o f biotechnology, facilitating the patenting and
licensing o f inventions in this sector.
Rather than em phasizing pu blic fu nding and relatively liberal disclosure and
dissem ination, the B a y h -D o le Act assum es that restrictions on dissem ination o f
the results o f m any R8cD projects will enhance econom ic efficiency by supporting
their com m ercialization. In m any respects, the B a y h -D o le Act is the ultim ate
expression o f faith in the “ linear m o d el” o f in n ovation — if basic research results
can be purchased b y w ou ld-be developers, com m ercial innovation will be
accelerated.

8.5.2.2 The Effects o f B a y h -D o le

H ow did the B a y h -D o le A ct affect technology transfer b y U S universities? Figure 8.4
depicts U S research u n iversity patenting as a share o f dom estically assigned U S
patents d u rin g 19 6 3-9 9 , in order to rem ove the effects o f increased patenting in the
United States b y foreign firm s and inventors d u rin g the late twentieth century.
U niversities increased their share o f patenting from less than 0.3 per cent in 1963 to
nearly 4 per cent by 1999, but the rate o f grow th in this share begins to accelerate
before rather than after 1980. The grow th rate o f the ratio o f research university
patents to academ ic research spen ding rem ains surprisingly constant through the
1963-93 p erio d , suggesting no structural break in trends in universities' “ patent
p ropen sity” after passage o f the B a y h -D o le Act in 1980.
Figure 8.5 displays trends d u rin g 19 6 0-9 9 in the distribution am o n g tech nology
classes o f U S research u n iversity patents, h ighlighting the grow ing im portance o f
biom edical patents in the patenting activities o f the leading U S universities d urin g
the period. N o n -b io m ed ical university patents increased b y 90 per cent from the
196 8 -70 p erio d to the 19 78 -8 0 period, but biom edical university patents increased
b y 295 per cent. T h e increased share o f the biom edical disciplines w ithin overall
federal academ ic R8cD fu n din g, the dram atic advances in biom edical science that
occurred d u rin g the 1960s and 1970s, and the strong industrial interest in the results
o f this b io m ed ical research d u rin g this period all contributed to this shift in the
co m p o sitio n o f u n iversity patent portfolios.
D u rin g the late 1990s and early tw enty-first century, m an y com m en tators and
po licy m akers p o rtrayed the B a y h -D o le A ct as a critical catalyst to grow th In U S
universities' in n ovative and econ om ic contributions. Indeed, the O E C D went so far
as to argue that the B a y h -D o le A ct w as an im portan t factor in the rem arkable grow th
o f incom es, em ploym en t, and p ro d u ctivity in the U S econ o m y o f the late 19 9 0 5 .°
R em arkably, v irtu a lly none o f these characterizations o f the positive effects o f the
B a y h -D o le A ct cite an y evidence in su p p o rt o f their claim s beyond the clear grow th
in patenting and licensin g b y universities. N o r does evidence o f increased patenting
230 D A V I D C. M O W E R Y A N D B H A V E N N . S A M P A T

Fig. 8.4 US research university patents as a percentage of all

domestic-assignee US patents, 1 9 6 3 -9 9

and licensing b y universities by itself indicate that u n iversity research discoveries are
being transferred to indu stry m ore efficiently o r com m ercialized m o re rapidly as
C olyvas et a l (2002) and M o w ery et al. (2001) p o in t out.
These assessm ents” o f the effects o f the B a y h -D o le A ct also fail to consider any
potentially negative effects o f the Act on U S u n iversity research o r innovation in the
broader econom y. Som e scholars have suggested that the “ com m ercialization mo­
tives created b y B a y h -D o le could shift the orien tation o f u n iversity research away
from “ basic” and tow ards “ applied” research (H end erson et a l 1998), but thus far
there is little evidence o f substantial shifts since B a y h -D o le in the content of
academ ic research.
A second potentially negative effect of increased u n iversity patentin g and licens­
ing is the potential weakening o f academ ic researchers' co m m itm en ts to “ open
scien ce” leading to publication delays, secrecy and w ith h o ld in g o f data and
 U N I V E R S I T I E S IN N A T I O N A L I N N O V A T I O N S Y S T E M S 231

Issu e Y e a r
И Chemicals and Chemical Processes (Excluding Drugs) О Mechanical
■ Drugs and Medical Technology □ Other
■ Electronic, Optical, and Nuclear

Fig. 8.5 Technology field of US “research university" patents, 1 9 6 0 -9 9

m aterials (D asgu p ta and D avid 1994; Liebeskind 2001). In view o f the im portance
assigned b y indu strial researchers to the “ nonpatent/licensing” channels o f inter­
action w ith universities in m ost indu strial sectors, it is crucially im portan t that these
channels n ot be constricted or im peded b y the intensive focus on patenting and
licensing in m an y universities. The effects o f any increased assertion by institutional
and in d ivid u al inventors o f p ro p erty rights over inputs to scientific research have
o n ly begun to receive serious scholarly attention. Patenting and restrictive licensing
o f in pu ts into fu tu re research (“ research too ls” ) could hinder dow nstream research
and p ro d u ct d evelopm ent (H eller and Eisenberg 1998; M erges and N elson 1994).
A lth o u gh there is little evidence as yet that the B a y h -D o le A ct has had significant,
negative consequences fo r academ ic research, technology transfer, and industrial
in n o vatio n in the U nited States, the data available to m o n itor any such effects are
v e ry lim ited. M oreover, such data are necessarily retrospective, and in their nature
232 D A V I D C. M O W E R Y A N D R H A V E N N . S A M P A T

<urc likely to reveal significant changes in the n orm s an d b eh a vio r o f researchers or

universities o n ly w ith a long lag. A n y negative effects o f B a y h —D ole accordingly are
likely to reveal them selves on ly well after they first appear.

8 .5.2.3 International “ E m u lation ” o f the B ayh —D ole A ct

The lim ited evidence on the A c ts effects (both p ositive and negative) has not
prevented a num ber o f other O E C D governm ents fro m p u rsu in g policies that
closely resem ble the Bayh—D ole Act. Like the B ayh —D o le A ct, these initiatives
focus narrow ly on the "d eliverab le" outputs o f u n iversity research, and typically
ignore the effects o f patenting and licensing on the other, m o re economically
im portant, channels through w hich universities co n trib u te to in n o vation and eco­
nom ic grow th. M oreover, such em ulation is based on a m isread in g o f the limited
evidence concerning the effects o f B a y h -D o le and on a m isunderstandin g o f the
factors that have encouraged the lon g-stan d in g and relatively close relationship
between U S universities and industrial innovation.
The po licy initiatives that have been debated o r im plem en ted in m ost OECD
econom ies have sought to shift ow nership o f the intellectual p ro p erty rights (IPR) for
academ ic inventions to either the academ ic institu tion o r the researcher (see OECD
2002, for an excellent su m m ary). In som e u n iversity system s, such as those of
G erm an y or Sweden, researchers have long had ow n ersh ip rights fo r the intellectual
pro perty resulting from their w ork, and debate has centered on the feasibility and
advisability o f shifting these ow nership rights fro m the in d ivid u al to the institution.
In Italy, legislation adopted in 2001 shifted ow nership fro m u niversities to individual
researchers. In Japanese universities, ow nership o f intellectual property rights
resulting from pu blicly funded research is determ ined b y a com m ittee, which on
occasion awards title to the researcher. N o single national p o licy governs IP R owner-
ship within the British or C anadian university system s, althou gh efforts are underway
in both nations to grant ow nership to the academ ic in stitu tion rather than the
in divid ual researcher or the fu nding agency. In ad d ition , the Sw edish, Germ an, and
Japanese governm ents have encouraged the fo rm atio n o f external "technology
licensing organizations," which m ay o r m ay not be affiliated w ith a given university
These policy proposals and initiatives display the classic signs o f international
em ulation described above— selective "b o rro w in g " from anoth er nation s policies
for im plem entation in an institutional context that differs sign ifican tly from that of
the nation being em ulated. Inasm uch as patenting and licensin g are rated by indus­
trial R8cD m anagers as relatively un im p ortan t fo r tech n o lo gy transfer in m ost fields,
em ulation o f the Bayh—Dole Act is insufficient and perhaps even unnecessary to
stim ulate higher levels o f u n iversity -in d u stry interaction and tech nology transfer.
Instead, reform s to enhance interinstitutional com petition and au to n o m y within
national university systems, as well as su p port for the external institutional con­
tributors to new -firm form ation and tech nology co m m ercializatio n , appear to be
m ore im portant. Indeed, em ulation o f B a y h -D o le could be counterproductive in
other industrial econom ies, precisely because o f the im portance o f other channels
for technology transfer and exploitation b y industry.

8.6 C o n c l u s i o n

U niversities p lay im p o rtan t roles in the “ know ledge-based” econom ies o f m odern
industrial and indu strializing states as sources o f trained “ know ledge w orkers” and
ideas flo w in g fro m both basic and m ore applied research activities. B u t conventional
(and, perhaps, evo lu tio n ary) econom ic approaches to the analysis o f institutions are
very difficult to apply to universities, for several reasons. First, w ith the exceptions o f
the U S and B ritish u n iversity system s, inter-un iversity “ co m petition ” has been
lim ited in m ost national system s o f higher education. Inter-university com petition
was a very im p o rtan t historical influence on the evolution o f U S universities and
their links w ith in d u stry; but this aspect o f the “ selection environ m en t” is lacking in
m ost other national system s o f higher education.
Second, analyzing universities as econom ic institutions requires som e defin ition
o f the objectives pu rsued by in d ivid u al universities. Partly because universities
perform m u ltiple roles in m an y national system s and partly because the internal
structure o f m ost research universities m ore closely resem bles that o f a cooperative
organization rather than the hierarchical structure associated w ith industrial firm s,
characterizing “ the objectives o f the u n iversity” is difficult if not m eaningless. The
m odern u n iversity has its roots in the M id dle Ages, rather than the Industrial
R evolution, and its m edieval origin s continue to influence its organization and
operation. I f universities are to be conceptualized as econom ic institutions for
purposes o f analyzing their evolution, the current analytic fram ew orks available in
neo-classical o r evo lu tio n ary econom ics are insufficient.
The developm ent o f such an analytic fram ew ork is im portant, not least for under­
standing the consequences for academ ic research o f governm ent policies that seek to
accelerate the transfer o f research results to industrial firm s. The intensified dem ands
from governm ents to raise the (m easurable) econom ic returns to their substantial
investm ents in academ ic research and education m akes the developm ent o f better
tools for understanding and m easuring the operations and outputs o f universities all
the m ore im portant. A s w e argued above, m any o f the current initiatives in the United
States and other industrial econom ies to enhance the econom ic returns from univer­
sity research are based on a p o o r understanding o f the full spectrum o f roles fulfilled by
research universities in industrial econom ies, as well as a tendency to em phasize the
outputs o f university research that can be easily quantified.
234 D A V I D C . M O W E R Y A N D B H A V E N N . S A M P A T _________________

A lth ough the analytic fram ew orks pro vid ed b y the ‘ n atio n al in n ovation systems,”
“ M ode 2,” and “ Triple H elix” m odels o f scientific research an d innovation shed
som e light on the roles o f universities and largely agree in th eir assessm ent o f these
roles, these fram ew orks provide lim ited guidance fo r p o licy o r evaluation. More­
over, these fram ew orks tend to d ow n p lay the v ery real ten sions am on g the different
roles o f research universities w ithin know ledge-based econom ies. Such tensions are
likely to intensify in the face o f pressure fro m p o licy m akers and others on
universities to accelerate their pro d u ction and tran sfer to co m m ercial interests of
tangible, m easurable research outputs.
The developm ent o f useful theoretical o r conceptual tools o r m odels for analyzing
universities as econom ic or other institutions w ith in kn ow ledge-b ased economies is
seriously ham pered by the lack o f data on the roles o f universities that enable
com parisons across tim e or across national in n o vatio n system s. Indicators that
enable longitudinal analysis o f the roles o f universities in train in g scientists and
engineers, contributing to “ public know ledge,” o r tran sferrin g inventions to indus­
trial firm s are scarce if not entirely lacking fo r m ost n ation al system s o f higher
education. Few o f these indicators in corporate in fo rm a tio n on the geographic
dim ensions o f u n iversity -in d u stry interactions, despite the im p o rtan ce o f agglom­
eration econom ies in the current po licy approaches o f m an y governm ents in this
area. M oreover, such indicators as do exist rarely are co m parab le across national
system s o f higher education.
The absence o f broader longitudinal and cross-n atio n ally com parable indicators
o f u n iversity-in d u stry interaction thus im pedes b oth the fo rm u latio n and the
evaluation o f policies. A nd the lack o f better indicators reflects the lack o f a stronger
analytic fram ew ork for understanding the roles o f universities w ithin national
innovation system s. Such a fram ew ork m ust ad o p t a m o re evolu tion ary, historically
grounded approach to the understanding o f the roles o f universities, especially the
influence o f the structure o f national higher edu cation system s o n these roles. As we
have argued in this chapter, m any o f the efforts b y O E C D govern m en ts to encourage
technology transfer and to increase the econom ic payoffs to investm ents in univer­
sity research are ham pered b y a lack o f such und erstand in g. M ore comparative
institutional w ork on the evolution and roles o f research universities, including the
contrasting division o f lab or” am ong universities and other p u blicly supported
research institutions in both industrial and indu strializing econom ies, is an indis­
pensable starting point for analysis o f the current and likely fu tu re position o f the
research university w ithin national innovation system s.
The developm ent o f better indicators o f the lu ll array o f channels through which
industries and universities interact w ithin know ledge-based econom ies represents
another im portant research opportunity. In add ition , m o re in fo rm atio n is needed
on m easures o f firm -level absorptive cap acity” and investm ents in its creation and
m aintenance how do existing firm s develop these vario u s channels o f interaction?
Flow and why are new firm s form ed to exploit u n iversity research advances, and how
 U N I V E R S I T I E S IN N A T I O N A L I N N O V A T I O N S Y S T E M S 235

does this sp in o ff process v ary across tim e, geographic space, and national in n o v ­
ation systems? The extensive discussion o f all o f these im portant econom ic ph e­
nom ena still lacks a stron g eviden tiary basis for m aking com parisons am ong the
higher education system s o f the industrialized econom ies. The current em phasis on
the countable rather than the im portan t aspects o f u n iversity -in d u stry interactions
could have un fortu n ate consequences for innovation p o licy in the industrial and
industrializing w orld.

N otes

1. Godin and Gingras (2000:273) note that “After having been left out of major government
policies centered on industrial innovation, universities seem, over the past 5 years, to
have become the object of a renewed interest among students of the system of knowledge
production”
2. Mowery and Rosenberg (1989: 154) note that the conduct of scientific research and
education within many research universities “exploits a great complementarity between
research and teaching. Under the appropriate set of circumstances, each may be per­
formed better when they are done together”
3. This list draws from Rosenberg (1999), Cohen et al. (1998), and other sources.
4. David, Mowery, and Steinmueller (1992) and Nelson (1982) discuss the economic
importance of the “ informational” outputs of university research.
5. See Rosenberg s (1994) discussion of universities as a source of innovation in scientific
instruments.
6. See Rosenberg (1999).
7. Thus, Nelson's concluding chapter in his 1993 collection of studies of national innovation
systems argues that “ One important feature distinguishing countries that were sustain­
ing competitive and innovative firms was education and training systems that provide
these firms with a flow of people with the requisite knowledge and skills. For industries in
which university-trained engineers and scientists were needed, this does not simply
mean that the universities provide training in these fields, but also that they consciously
train their students with an eye to industry needs” (1993: 511).
8. According to Ben-David (1971: 48), “ To create order among the turbulent crowds of
scholars and to regulate their relationships with the environing society, corporations
were established. Students and scholars were formed into corporations authorized by the
church and recognized by the secular ruler. The relationships of their corporation with
that of the townspeople, with the local ecclesiastical officials, and with the king were
carefully laid down and safeguarded by solemn oaths.... The important result of this
corporate device— which was not entirely unique to Europe but which attained a much
greater importance there than elsewhere—was that advanced studies ceased to be
conducted in isolated circles of masters and students. Masters and/or students came to
form a collective body The European student of the thirteenth century no longer went to
study with a particular master but at a particular university”
236 D A V I D C, M O W E R Y A N D B H A V E N N . S A M P A T

9** The Japanese higher education system has a large number o f private universities*
although the bulk o f these are devoted primarily to undergraduate education.
10. US “ science and engineering” degreeholders also account for a smaller share of all
advanced degrees awarded in the United States in 1999 than is true o f France, Taiwan*
and the United Kingdom, although this share in the United States exceeded those for
Finland and South Korea (National Science Foundation 2002).
11. These data must be interpreted with caution, since the definitions used by the national
statistical agencies whence they are drawn often differ. For example, in France CNRS is
classified as part o f the Higher Education Sector, whereas in Italy similar organizations
are treated as part o f the government sector. See OECD 2001k Annex 2.
12. By comparison, the share o f US university publications co-authored by industrial and
university researchers grew from 4.9 per cent in 1989 to 7.3 per cent in 1999, while this
share in Canadian university publications grew from 1.4 per cent in 1980 to 3.5 per cent in
1998.
13. A recent OECD report notes that “ Spinning off is the entrepreneurial route to commer­
cializing knowledge developed by public research and as such is attracting a great deal of
attention, given the 'start-up' fever in many countries” and that governments “have a
special interest in this specific type o f industry-science linkage because it maybe one of
the factors that explain differences in performance in newfast-growing science based
industries” (OECD 2002: 41).
14. According to Katz and Ordover (1990), at least fourteen Congressional bills passed
during the 1980s focused on strengthening domestic and international protection for
intellectual property rights, and the Court o f Appeals for the Federal Circuit created in
1982 has upheld patent rights in roughly 80 per cent o f the cases argued before it, a
considerable increase from the pre-1982 rate o f 30 per cent for the Federal bench.
15. “ Regulatory reform in the United States in the early 1980s, such as the Bayh-Dole Act,
have [sic] significantly increased the contribution o f scientific institutions to innovation.
There is evidence that this is one o f the factors contributing to the pick-up of US growth
performance” (OECD 2000: 77).

R eferen ces

A r r o w , K. (1962),Economic Welfare and the Allocation o f Resources for Invention,” in

R, R. Nelson (ed.), The Rate and Direction of Inventive Activity; Princeton: Princeton
University Press.
* B e n -D a v id , J. (1968), Fundamental Research and the Universities, Paris: OECD.
— (1971), The Scientist s Role in Society^New York: Prentice-Hall.
B u r n , B. B, A l i b a c h , R G,, K e r r , C., and P e r k i n s , J. A. (1971), Higher Education in Nine
Countries, New York: McGraw-Hill.
B u s h , V. (1945), Science. The Endless Frontier^ Washington, DC: US Government Printing
Office.

* Asterisked items are suggestions for further reading.

 U N I V E R S I T I E S IN N A T I O N A L I N N O V A T I O N S Y S T E M S 2 JJ

C a lv e r t , J., and P a t e l , P. (2002), “ University-Industry Research Collaborations in the UK,”

unpublished working paper, Science Policy Research Unit, University o f Sussex, Brighton,
UK.
C o h e n , W., F l o r id a , R., R a n d a z z e s e , L , and W a l s h , J. (1998), “ Industry and the Academy:
Uneasy Partners in the Cause o f Technological Advance,” in R. Noll (ed,), Challenges to the
Research University>Washington, DC: Brookings Institution.
* C o h e n , W. M., N elso n , R. R., and W a lsh , J. R (2002), “ Links and Impacts: The Influence
o f Public Research on Industrial R&D,” Management Science 48(1): 1-23.
*C o lyv a s , J., C row , M., G e l ijn s , A., M a zzo len i , R., N elso n , R. R., R o sen berg , N., and
Sa m p a t , B. N. (2002), “ How Do University Inventions Get into Practice?” Management
Science 48(1): 61-72.
D a sg u pta , P., and D a v id , P. (1994), “ Towards a New Economics o f Science,” Research Policy
23(5): 487-521.
D a v id , R, F o ray , D., and S t ein m u e ller , W. E. (1999), “ The Research Network and the New
Economics o f Science: From Metaphors to Organizational Behaviours,” in F. Malerba and
A. Gambardella (eds.), The Organization of Economic In novation in Europe, Cambridge:
Cambridge University Press.
------ M o w ery , D. C., and St ein m u e ll e r , W. E. (1992), “Analyzing the Economic Payoffs
from Basic Research,” Economics of Innovation and New Technology: 73-90.
E t z k o w it z , H., and L e y t e s d o r f f , L. (1997), Universities in the Global Economy: A Triple
Helix of Academic-Industry-Government Relation. London: Croom Helm.
------ W e b s t e r , A., and H e a l e y , P. (1998), “ Introduction,” in H. Etzkowitz, A. Webster, and
P. Healey (eds.), Capitalizing Knowledge, Albany: State University o f New York Press.
F else n st ein , D. (1994), “ University-Related Science Parks— "Seedbeds" or "Enclaves" o f
Innovation?"" Technovation 14: 93-100,
G e ig e r , R. (1986), To Advance Knowledge: The Growth of American Research Universities,
1900-1940, New York: Oxford University Press.
G ibbo n s , M. et al. (1994), The New Production of Knowledge, London: Sage.
G odin , B., and G in g r a s , Y. (2000), “ The Place o f Universities in the System o f Knowledge
Production,"" Research Policy 29: 273-8.
G overn m en t U n iv e r sit y I n d u stry R esearch R o undtable (GUIRR) (1991), Industrial
Perspectives on Innovation and Interactions with Universities, Washington, DC: National
Academy Press,
G rah am , H. D., and D iam o nd , N. (1997), The Rise of American Research Universities.
Baltimore: Johns Hopkins University Press.
*Heller, M. A., and E is e n b e r g , R. S. (1998), “ Can Patents Deter Innovation? The Anti-
commons in Biomedical Research,” Science 280: 298.
H enderso n , R., and C o c k bu rn , I. (1998), “Absorptive Capacity, Coauthoring Behavior, and
the Organization o f Research in Drug Discovery,"" Journal of Industrial Economics 46(2):
157-82,
------ J a f f e , A. R., and T r a jt e n b e r g , M, (1998), “ Universities as a Source o f Commercial
Technology: A Detailed Analysis o f University Patenting, 1965-88” Review of Economics &
Statistics 80:119-27.
* ---- O rsen ig o , L„ and P isano , G. (1999)» “ The Pharmaceutical Industry and the Revolu­
tion in Molecular Biology: Interactions among Scientific, Institutional and Organizational
Change ” in D, C. Mowery and R. R. Nelson (eds.), Sources of Industrial Leadership, New
York: Cambridge University Press, 267-311.
238 D A V I D C, M O W E R Y A N D B H A V E N N . S A M P A T

H ic k s , D„ B r e it z m a n , T., O l iv a s t r o , D., and H a m il t o n , K. (2001), “ The Changing

Composition o f Innovative Activity in the US. A Portrait Based on Patent Analysis,”
Research Policy 30: 681-703.
_____and H a m il t o n , K. (1999), “ Does University-Industry Collaboration Adversely Affect
University Research?” Issues in Science and Technology Online, http://www.nap.edu/issues/
15,4/realnumbers.htm
_____I s a r d , P. A., and M a r t i n , B. R. (1995), “A M orphology o f Japanese and European
Corporate Research Networks,” Research Policy 2 5 : 359- 78.
K a t z , M. L„ and O r d o v e r , J. A. (1990), “ R&D Competition and Cooperation,” Brookings
Papers on Economic Activity: Microeconomics, 137—92.
* K l in e , S., and R o s e n b e r g , N. (1986), “An Overview o f Innovation ” in R. Landau and N.
Rosenberg (eds.)> The Positive Sum Strategy: Harnessing Technology for Economic Growth,
Washington, DC: National Academy Press, xiv, 640.
L a r e d o , R, and M u s t a r , P. (2001), Research and Innovation Policies in the New Global
Economy: An International Comparison, Cheltenham: Elgar.
Leslie, S. W. (1993), The Cold War and American Science: The Military-Industrial-Academic
Complex at MIT and Stanford, New York: Columbia University Press.
------ (2000), “ The Biggest Angel o f Them All: The M ilitary and the Making o f Silicon Valley”
in M. Kenney (ed.), Understanding Silicon Valley: The Anatomy o f an Entrepreneurial
Region, Stanford, Calif.: Stanford University Press.
* ---- and K a r g o n , R. H., (1996), “ Selling Silicon Valley : Frederick Termans Model for
Regional Advantage,” Business History Review 70: 435-72,
* L e v i n , R. G , K l e v o r ic k , A., N e l s o n , R. R., and W i n t e r , S. (1987), “Appropriating the
Returns from Industrial Research and Development,” Brookings Papers on Economic
Activity 3: 783-820,
L ie b e s k in d , J. (2001), “ Risky Business: Universities and Intellectual Property,” Academe
87(5): http://www.aaup.org/publications/Academe/oiSO/sooilie.htm
M a n s f ie l d , E, (1991), “Academic Research and Industrial Innovations,” Research Policy
2 0 :1-12.
M a s s e y , D. B., Q u in t a s , P., and W ie l d , D. (1992), High-Tech Fantasies: Science Parks in
Society, Science, and Space. London and New York: Routledge.
M e r g e s , R., and N e l s o n , R, R. (1994), “ On Limiting or Encouraging Rivalry in Technical
Progress: The Effect o f Patent Scope Decisions,” Journal of Economic Behavior and Organ­
ization 25:1-24.
M o w e r y , D. C., N e l s o n , R. R. S a m p a t , B. N., and Z ie d o n is , A. A. (2001), “ The Growth of
Patenting and Licensing by US Universities: An Assessment o f the Effects o f the Bayh-Dole
Act of 1980,” Research Policy 30: 99-119.
* (2004), The Ivory Tower and Industrial Innovation: University-Indus­
try Technology Transfer Before and After the Bayh-Dole Act, Stanford, Calif: Stanford
University Press.
and R o s e n b e r g , N. (1989), Technology and the Pursuit of Economic Growth, New York:
Cambridge University Press.
(1993), The US National Innovation System,” in Nelson 1993:
N a i io n a l S c ie n c e B o a rd (2002), Science and Engineering Indicators: 2002, Washington,
DC: US Government Printing Office.
N e l s o n , R. R. (1959),The Simple Economics o f Basic Scientific Research,” Journal 0)
Political Economy 6y: 297—306.
 U N I V E R S I T I E S IN N A T I O N A L I N N O V A T I O N S Y S T E M S 239

------ (1982), “ The Role o f Knowledge in R and D Efficiency” Quarterly Journal of Economics
97(3): 453“ 70.
------(i993)> National Innovation Systems; A Comparative Analysis. New York: Oxford Uni­
versity Press.
o ecd (2000), A New Economy Paris: OECD.
------ (2001a), “ Education at a Glance. OECD Indicators 2001 Edition,” Paris: OECD.
------ (2001 b), Main Science and Technology Indicators, Paris: OECD.
------ (2001c), Science, Technology, and Industry Scoreboard, Paris: OECD.
------ (20oid), Basic Science and Technology Statistics, Paris: OECD.
------ (2002), Benchmarking Science-Industry Relationships, Paris: OECD.
R o s e n b e r g , N. (1992), “ Scientific Instrumentation and University Research,” Research Policy
21: 381-90.
------ (1999)5 “American Universities as Economic Institutions,” Mimeo.
*------ and N e l s o n , R. R. (1994), “American Universities and Technical Advance in Industry,”
Research Policy 23: 323-48.
S a x e n ia n , A. (1988), “ The Cheshire Cat’s Grin: Innovation and Regional Development in
England,” Technology Review 91: 67-75.
S l a u g h t e r , S., and L e s l i e , L. L. (1997), Academic Capitalism: Politics, Policies, and the
Entrepreneurial University, Baltimore: Johns Hopkins University Press.
S t u r g e o n , T. J. (2000), “ How Silicon Valley Came to Be,” in M. Kenney (ed.), Understanding
Silicon Valley: The Anatomy of an Entrepreneurial Region, Stanford, Calif.: Stanford
University Press, xvi, 285.
T r a jt e n b e r g , M., H e n d e r s o n , R., and J a f f e , A. B. (1997), “ University Versus Corporate
Patents: A Window on the Basicness o f Inventions,” Economics of Innovation and New
Technology 5:19-50,
T ro w , M. (1991), “American Higher Education: ‘Exceptional5 or Just Different,” in В. E.
Shafer (ed.), Is America Different? A New Look at American Exceptionalistn, Oxford: Oxford
University Press.
W a l l s t e n , S, (2001), “ The Role o f Government in Regional Technology Development: The
Effects o f Public Venture Capital and Science Parks,” Stanford University: SIEPR Working
Paper.
 CHAPTER 9

FINANCE AND
INNOVATION

MARY O*SULLIVAN

9. 1 I n t r o d u c t i o n 1

I n n o v a t i o n is an expensive process; significant resources m u st be expended to

initiate, direct and sustain it. It is a process that takes tim e w h ich m eans that the
resources that su pport it m ust be com m itted until the process is com plete. Finally, its
outcom es are uncertain so the returns to in n o vative investm ents are not assured.
The im portance o f resource allocation to in n o vation , as w ell as the com plexity o f its
relationship to that process, m akes its system atic analysis cru cial to a comprehensive
econom ic theory o f innovation.
It is not surprising, therefore, that Joseph Schum peter, w id ely regarded as the
pioneer in the econom ic analysis o f innovation, m ade the stu d y o f resource alloca­
tion, especially the allocation o f financial resources, central to his study o f innov­
ation. In contrast, as I show in Section 9.2, co n tem p o rary econ om ists o f innovation
have largely neglected the relationship betw een finance and innovation. Though
there are a few exceptions to this rule, they are too recent an d too few to suggest that
we are on the brin k o f any system atic change in this regard.
However, som e financial econom ists w o rk in g in the fields o f enterprise finance2
and finance and grow th have begun to explore concerns that are closely related to
those that anim ated Schum peter’s research. In Section 9.3 I discuss research in
these fields that is relevant to the study o f finance and in n o vation . I note that
 FIN AN CE AND IN N O V A T IO N 2/$l

em pirical research has not kept pace w ith theoretical developm ents and the evidence
that does exist, even on basic propositions, is often am biguous. Furtherm ore, in
Section 9.4 , 1 p o in t to serious lim itations associated w ith the dom inant analytical
approaches em ployed in m icro- and m acroecon om ic research on finance for an a­
lyzing the d ynam ics o f econom ic change.
Intellectual exchange between evolu tion ary econom ist and financial econom ists
seems a better route to an im proved understanding o f the relationship between
finance and innovation. H ow ever, m ethodological differences, especially w ith re­
spect to the im portan ce o f h istory in econom ic analysis, are m ajor obstacles to
integration between the fields. Therefore, it m ay be m ore fruitful for econom ists
o f in n ovation to collaborate w ith econom ic and business historians w ho research
the origins and evolu tion o f financial system s. In closing, I draw attention to
som e crucial questions that need to be addressed in new research on finance and
innovation.

9. 2 T h e R o l e o f F i n a n c e i n t h e
E c o n o m i c s of I n n o v a t i o n

Schum peter’s analysis o f the relationship between innovation and resource alloca-
tion, especially the allocation o f financial resources, was central to his study o f the
econom ics o f in n ovation . M an y o f the details o f Schum peter’s econom ics o f in n o v­
ation w ere controversial and/or incom plete and con tem p orary research on the
subject has m ade considerable progress in refining, and expanding on, his analysis.
In particular, questions about the appropriate characterization o f the innovation
process have received considerable scrutiny. H owever, w ith a few notable exceptions,
the im plication s o f characteristics o f innovative activity for resource allocation,
especially the allocation o f financial resources, have been largely overlooked.

9.2.1 The Pioneering Work o f Schumpeter

The role o f innovation as the primary stimulus to the process of economic develop­
ment was the central preoccupation in Joseph Schumpeter’s work as an economist.
Schumpeter focused on two different, but related, units o f analysis in conceptual­
izing the links between innovation and resource allocation. On the one hand, he was
concerned with the implications o f the microeconomic characteristics of innovative
242 MARY O’ SULLIVAN

activity, notably features of entrepreneurial behavior, for resource allocation* In

addition, at a more aggregate level of economic analysis, he studied the interaction
between structural economic change and resource allocation. In both his micro- and
macroeconomic analyses, he paid particular attention to the role of finance in
facilitating economic change though, especially in his microeconomic analysis, his
understanding of that role shifted over time.

9*2x1 The M icroeconom ics o f Innovation

In Schum peter's early w ritings on the m icro econ o m ics o f in n o vatio n , especially the
The Theory of Economic Development (TED) and Business Cycles: A Theoretical
Historical and Statistical Analysis of the Capitalist Process (B C ), the process o f credit
creation featured prom inently. Later in his life, how ever, especially in Capitalism,
Socialism and Democracy ( CSD ), Schum peter d ow n played the role o f credit creation
in facilitating innovation and econom ic developm ent. Instead, he emphasized the
self-financing o f innovative investm ent b y d o m in an t enterprises. The shift in
Schum peter's thinking about the relationship betw een finance and economic
change reflected the w ell-kn ow n tran sform ation in his ch aracterization o f innov­
ation from a process driven b y new, entrepreneurial ventures to one dom inated by
large, persistent enterprises.

The Entrepreneur the Financier and Innovation. B egin n in g w ith an analysis o f the
allocation o f econom ic resources in the absence o f in n o vatio n , The Theory of
Economic Development is Schum peter's initial treatise on the econom ics o f innov­
ation. Schum peter's central claim was that the com petitive capitalist economy
w ould settle into a routine that m ight be in m o tio n b u t w o u ld tend towards a
stationary or equilibrium state. Resources w o u ld flo w aro u n d the econom y, along
routinized paths, in w hat Schum peter described as the “ circu lar flow o f economic
life." T h ey w ould be fully utilized in that perpetual m o tio n an d as such could not
accum ulate into stocks (Schum peter 1996: 45).
Schum peter then posed the question o f how in n o vatio n co u ld occu r under these
conditions. H e defined innovation as the com m ercial or in d u strial application of
som ething new, such as a new product or process, a new type o f organization, a new
source o f supply or product m arket (Schum peter 1996: 66). H e em phasized three
characteristics o f the innovation process that had crucial im p licatio n s for resource
allocation.
First, innovation depended on the investm ent o f resources: “ m ajo r innovations
and also m any m inor ones entail construction o f N ew Plant (and equipm ent)— or
the rebuilding ot old plant requiring nonnegligible tim e and o u tla y ” (Schumpeter
*939 * 68). Second, innovation, as a general rule, was em b od ied in new firm s that are
founded to undertake the new com bination (Sch u m peter 19 6 4 :6 9 ; see also Schump-
eter 1996. 66). Finally, innovation was usually driven b y entrepreneurs who were
 FINANCE AND INNOVATION 243

“ new m en, that is, w ho w ere not already prom inent in business circles (Schum peter
1996: 78; 1964: 70).
T h rou gh the process o f innovation, the econom y's existing productive resources
were to be com bined in new ways but h ow were new m en, operating in new firm s, to
get the control over the resources that they needed in order to innovate (Schum peter
1996; 68)? Since existing resources were already fully utilized in the circular flow,
som ehow they had to be detached from their current uses and m ade available to
entrepreneurs so that they could com bine them in new ways. W hat was the m ech­
anism through w hich resources w ou ld be reallocated from existing to new uses in the
economy?
The solution that Schum peter provided to the puzzle put the financial system at
center stage. Specifically, he argued that innovation was financed through the
creation o f credit. The pu rchasing pow er required by entrepreneurs to detach
resources fro m the circu lar flo w to undertake new com binations w as generated ex
nihilo. C redit creation did not need to be backed b y an existing stock o f m on ey or
goods (Sch um peter 19 9 6 :10 6 ).
Schum peter believed that credit could be created in a variety o f ways but he gave
prom inence to the role o f the com m ercial bank in generating new purchasing pow er
and m aking it available to entrepreneurs. In entrusting an entrepreneur w ith the
productive resources o f society, “ [the banker] m akes possible the carrying out o f
new com bin ation s, authorises people, in the nam e o f society as it were, to form
them. H e is the ep h or o f the exchange econ o m y” (Schum peter 1996: 74).

The Large-Scale Enterprise, Innovation , and the Role o f Finance . O ver the course o f
his scholarly life, Schum peter dram atically altered his characterization o f in n ovation
from a process d om in ated b y new m en and new firm s to one driven b y the activities
o f large-scale indu strial enterprises. D u rin g the period o f alm ost fo rty years between
the w riting o f TED and CSD, Schum peter m oved to the U nited States. H e was struck
b y the m ajo r changes u n d erw ay in the U S econ om y in the first decades o f the
twentieth cen tu ry and he believed that they represented a shift from nineteenth-
century “ com petitive capitalism ” to tw entieth-century “ trustified capitalism .”
W hereas p revio u sly he had argued that the life and pu rpose o f new firm s faded
over tim e, in CSD he claim ed that the “ perfectly bureaucratized giant industrial
un it” had succeeded in ration alizing and routinizing the process o f Innovation to
such an extent that the large-scale enterprise had becom e “ the m ost pow erful
engine” o f econ om ic progress (Schum peter 1942: 106), A s a result, technological
progress had in creasin gly becom e the business not o f in d ivid u al entrepreneurs but
o f “ team s o f train ed specialists w h o turn out w hat is required and m ake it w ork in
predictable w ays” (Sch u m p eter 19 4 2 :13 2 ; see also Schum peter 1949* 71 )-
Sch um p eter’s revised characterization o f the innovation process led to im portant
changes in his analysis o f resource allocation in a dynam ic econ om y and, m
particular, in his analysis o f finance. Specifically, it led him to d ow n play the role
244 MARY О SULLIVAN

o f external finance and the banking system in favor o f an em ph asis on internal

finance for facilitating innovative investm ent. In deed , fo r a scholar who once
described the process o f credit creation as “ the m o n e ta ry co m plem en t o f innov­
ation” (Schum peter 19 6 4 :8 5), the absence o f any reference in C S D to the process of
credit creation, o r the specific functions o f banks, represents a d ram atic change in
perspective.

9.2.1.2 Fin an ce , Innovation , an d Structural E conom ic C hange

A lthough TED is dom inated b y m icro eco n o m ic analysis, tow ard s the end of the
b o o k Schum peter sketched the outlines o f a m a cro eco n o m ic analysis o f innovation
and econom ic developm ent w hich he articulated in greater detail in Business Cycles.
H is m ain concern was w ith the question o f w h y e co n o m ic d evelopm ent adoes not
proceed evenly as a tree grow s, but as it w ere jerkily,” o f w h y there was a business
cycle, “ the w ave-like m ovem ent o f alternating p erio d s o f p ro sp e rity and depression”
w hich he claim ed had characterized the econ om ic system th ro u g h o u t the capitalist
period (Schum peter 1996; 223).
Schum peter's analysis o f the business cycle b u ilt on the m icro eco n o m ic analysis of
innovation that he had laid out in TEDy notab ly in assu m in g that innovation was
em bodied in new firm s established b y new m en. It w ent b eyon d it, however, in
suggesting that the process o f venture creation h ad im p o rta n t characteristics that
could o n ly be identified at the aggregate level. Specifically, he claim ed that innovative
activity was unevenly distributed over tim e and across indu stries.
As a result, econom ic developm ent was an u n even an d je rk y process (Schumpeter
19 6 4:76 ). The effect o f boom s in entrepreneurial a ctiv ity w as to fundam entally alter
the econom ic system that existed beforehand, d istu rb in g its equilibrium and
starting “ an apparently irregular m ovem ent in the e co n o m ic system , which we
conceive as a struggle tow ards a new eq u ilib riu m p o sitio n ” (Schum peter 1996;
233). Therefore, he rejected the notion o f econ o m ic d evelo p m en t as a mere quanti­
tative change in the level o f aggregate econ om ic a ctivity and argued that it was a
process o f qualitative change that could o n ly be u n d ersto o d w ith reference to the
structural com position and evolution o f the econom y.
H aving characterized the process o f econ om ic d evelo p m en t in this way, the
challenge for Schum peter was to explain h o w such qu alitative econom ic change
m ight occur. O nce again, the financial system or, m o re precisely, the banking system
and the credit expansion and contraction that it facilitated, featured as the crucial
m echanism that facilitated the reallocation o f resources necessary to induce dra­
m atic changes in the structure o f econom ic activity. F o r Schum peter, therefore, the
evolution o f a country s financial system w as o f crucial im p o rtan ce for facilitating
the waves o f innovation that he regarded as the m o tive force beh ind its economic
developm ent.
 FINANCE AND INNOVATION 245

9.2.2 The Neglect o f Finance in the Contemporary

Economics o f Innovation
G iven Schum peter's expansive vision o f the econom ics o f innovation, it is h ardly a
surprise that the details o f his argum ents have been subject to significant critical
scrutiny. M o st o f the attention has focused on his characterization o f the innovation
process, Fo r exam ple, his assertion o f a general historical tran sform ation from an
innovation process dom inated b y new ventures (often described as Type i in n o v ­
ation) to one driven b y large industrial firm s (Type 2) has been rejected. C o n tem ­
p o rary scholars have em phasized that both o f these patterns o f innovation coexist in
the econ om y w ith som e industries characterized b y Type 1 and others b y Type 2
innovation (W inter 1984: 295).
H owever, h ard ly any attention has been devoted to the relationship between
innovation and resource allocation even though Schum peter’s analysis o f it is as
controversial and incom plete as his characterization o f innovation. N evertheless,
questions about resource allocation are latent in m uch o f the w ork that has been
done on characteristics o f the innovation at the level o f the enterprise, the in du stry
and the econom y. Like Schum peter, econom ists o f innovation m ust develop an
explicit analysis o f the im plications o f these characteristics for the allocation o f
resources and, in particular, financial resources. O therwise, we m ight well ask
whether co n tem p o rary research on innovation as yet am ounts to an econom ics o f
innovation at least if w e define econom ics, as m ost people do, as centrally concerned
with resource allocation.

9.2.2.1 Fin ance , In novation , a n d the Enterprise

One w ay o f reading the im plications o f Schum peter’s intellectual evolution from
TED to CSD is as a call fo r a th eo ry o f the innovative enterprise. T h ou gh Schum peter
was confident that such an im p o rtan t research agenda w ould not be ignored, the
subject attracted alm ost no attention for m ore than three decades after his dem ise.
One m ajo r exception was a b o o k entitled The Theory o f the Growth o f the Firm
published in 1959 b y Edith Penrose. Penrose’s m ain argum ent was that the basic
foundation for the grow th o f firm s w as a dynam ic process o f organizational learning
that occurred w ith in enterprises.3 In recent decades, there has been a veritable
explosion o f interest in conceptualizing innovation as it occurs at the level o f the
firm , W hether derived d irectly fro m Penrose or not, the n o tio n that firm -level
innovation is based on a process o f organizational learning is an idea that is
com m on , even pervasive, in the literature on the innovative enterprise (fo r a sim ilar
conclusion, see N elson 1991; Teece, Pisano, and Shuen 1997; see also Lazo nick, Ch. 2,
and Lam , C h . 5, this volu m e).
M ost o f the research on the subject o f organizational learning in firm s m akes som e
reference to the process th rou gh w hich resources are allocated to allow them to
г ф MARY O' SULLI VAN

initiate, direct and sustain organizational learning. H ow ever, it is rare to find system­
atic discussions o f the im plications o f o rgan ization al learn in g fo r the allocation of
financial resources to investm ents and returns. O ne b asic qu estio n that needs to be
addressed is whether, and to w hat extent, the process o f o rgan izatio n al learning affects
the scale o f investm ent that m ust be m ade b y an in n o vative firm . Resources have to be
invested in the learning process itself, as w ell as the im p lem en tatio n o f the results of
such learning w ithin the organization. In add ition , fin an cial resources are required to
develop o r acquire the necessary co m plem en tary assets (Teece 1986) to commercialize
innovations based on this organizational learning. H ow ever, v e ry little research
has been conducted on the im plications o f the ch aracteristics o f organizational
learning for the resource requirem ents o f d ifferent types o f enterprises.
I f the organizational character o f the in n o vation process has im plications for the
scale o f enterprise investm ent, it will also affect the o rg an izatio n and governance of
the process o f resource allocation w ithin the firm . W h o are the key decision makers
in allocating resources to innovative investm ents? O n w h at basis do they make
their investm ent decisions? H ow do they coordinate investm ent decisions across
the firm? W hat capabilities do they have for m akin g these decisions? H ow do their
incentives relate to those o f participants in the learn in g process and to organiza­
tional objectives?
These questions only hint at the plethora o f issues that are raised b y a consider­
ation o f the organizational d im ension o f firm -level in n o vatio n . Econom ists have
also highlighted other im portant characteristics o f the in n o vatio n process, such as
its cum ulative character and inherent uncertainty; their im p licatio n s for resource
allocation also require consideration in any com preh ensive econ om ics o f innovative
enterprise. Scholarly analysis o f these issues is, as yet, u n co m m o n . Indeed, studies of
the internal processes that com panies use to allocate fin an cial resources to any type
o f investm ent, not just innovative ones, are rare (fo r exceptions, see Bower 1970;
Burgelm an and Sayles 1986).

9 .2 .2 .2 Finance, Innovation, an d the In dustry

E m pirical studies have confirm ed the im portan ce o f cro ss-in d u stry variation in
certain basic characteristics o f innovative activity. In du stries v a ry not only in the
distribution o f innovative activity between entrants and in cu m b en ts at a point in
tim e but also over tim e as m easured b y the stability o r tu rb u len ce o f these popula­
tions shares o f innovative inputs and outputs (for a su m m ary, see M alerba, Ch. 14*
this volum e). Sectoral differences in innovative activity have im p o rtan t implications
for resource allocation.
To the extent that entrants dom inate innovative activity, fo r exam ple, the ques­
tion of how they get access to the resources that they require to innovate is crucial.
W hile the financing o f de novo entry is relevant here, em p irical studies have shown
that entrants are not necessarily new ventures; en try also occu rs through spin-offs
 F IN A N C E AND I N N O V A T I O N 247

from established firm s as well as diversification by incum bent firm s in other indu s­
tries, As a result, understanding how the resources o f existing firm s are reallocated,
by decision m akers in those firm s o r the em ployees that leave them to establish new
ventures, is also im portant.
I f the heterogeneity o f entrants" origins generates a range o f questions about
resource allocation, so too do other types o f observed variation in sectoral patterns
o f innovation. Industries disp lay im portant differences in com petitive interactions
am ong firm s. I f we focus on the relationship between entrants and incum bents, for
exam ple, w e find that som e entrants com pete directly with incum bent firm s for
custom ers w hereas others build an innovative strategy based on licensing arrange­
ments or jo in t ventures. C an s, H su, and Stern contrast biotechnology, where join t
ventures and alliances between entrants and incum bents are com m on, w ith hard
disk drives, where entrants confronted incum bents in head-to-head com petition in
product m arkets (C an s, H su, and Stern 2 0 0 0 :3). These differences have im plications
for the type and am o u n t o f investm ent that needs to be m ade by innovative firms.
They affect the extent to w hich firm s are involved in different business activities,
such as p ro d u ctio n o r m arketing, and, m ore generally, the costs for different players
o f organizing them selves to innovate.
Industries differ not o n ly in the com petitive interactions am ong firm s but also in
terms o f the relationships between firm s and other industry actors. Com petitive firm s
m ay have relationships w ith their suppliers and custom ers as well as universities and
governm ents (local, regional, national, and international) that exert an im portant
influence on their allocation o f resources (M alerba, Ch. 14, this volum e). In the early
stages o f the developm ent o f the U S software industry, for exam ple, the federal
governm ent played a crucial role in supporting the education and training o f large
num bers o f software engineers (M ow ery and Langlois 1996). As a result, m any start-up
com panies were able to enter the industry as viable com petitors w ithout having access
to the resources that w ould have been necessary if they had undertaken this effort on
their ow n. Sim ilarly, in the U S biotechnology industry the citation patterns for patents
show a m ajor reliance by private com panies on public science in their innovative
activities w ith sim ilar im plications for their capacity to econom ize on the resources
that they needed to invest to becom e players (M cM illan, N arin, and Deeds 2000),
In add ition to intersectoral variation , the relationship between finance and
innovation m ay v a ry over tim e w ithin a given industry. Industries in w hich entrants
drive in n o vatio n in one era m ay change over tim e to a structure in w hich innovation
becom es m ore increm ental in character and is dom inated b y incum bent firm s, as in
the au tom ob ile industry. Cases o f the opposite trend also exist, as incum bent-
dom inated indu stries face waves o f innovation-led entry— the developm ent o f
bio tech n o logy w ith in pharm aceuticals is a good exam ple. The characteristics o f
the enterprises that require finance, the sources o f finance on w hich they rely, and the
im plications o f these finan cial arrangem ents for innovation, are likely to differ
significantly as a result o f these evo lu tio n ary changes.
248 M A R Y 0 *S U L L I V A N

W c m u st also co n sid er the p o ssib ility th at tb e d ire c tio n o f c a u sa tio n goes in the
o th e r d irectio n , fro m fin an ce to innovation» A s fa r as the d istrib u tio n o f innovative
a c tiv ity b etw een en tran ts an d in cu m b e n ts is c o n c ern e d , fo r e x a m p le , w e could ask
w h eth er in cu m b e n t firm s d o m in a te b ecau se th e y are m o re in n o v a tiv e o r because
en tran ts are to o fin a n cia lly co n strain ed to co m p e te w ith th e m . C o n versely, when
en tran ts d o m in a te is it b ecau se th ey are m o re in n o v a tiv e th a n in cu m b en ts or
b ecau se a ready, an d even excessive, a v a ila b ility o f fin a n c ia l re so u rc es allow s them
to d o so?
T h e latter q u estio n is the su b ject o f tw o related stu d ie s o f th e h a rd d isk drive
in d u s tr y In an article called “ C a p ita l M a rk e t M y o p ia ,” W illia m S ah lm an and
H o w a rd S teven so n (1985) arg u e th at v en tu re ca p italists a n d th e sto ck market
m assiv ely o ver-in vested in the in d u stry d u rin g the p e rio d fr o m 19 7 7 to 1984 with
negative co n seq u en ces fo r fin an cial retu rn s an d fo r in n o v a tio n . H o w ever, another
stu d y o f the d isk d rive in d u stry claim s that th at the d ia g n o sis o f c a p ital market
m y o p ia is su spect in lo n g -te rm p ersp ective. B y g ra ve , L a n g e, R o e d e l, a n d W u accept
that m an y players in the d isk d rive in d u stry failed b u t th e y a rg u e th at the survivors
u ltim ately en jo yed su fficien t co m m ercial success to ju s tify th e fin a n c ia l bets that
w ere m ad e o n the in d u stry (B ygrave, L an ge, R o ed el, a n d W u 2 0 0 0 :17 ) .
T h ese tw o stu d ies h in t at the p o ten tial o f an alyses o f the jo in t in flu e n c e o f finance
an d in n o v a tio n o n in d u stry evo lu tio n b u t th e y are e x c e p tio n s (see also Carpenter,
L azo n ick , an d O 'S u lliv a n 2003). T h e exten t to w h ic h la b o u r is reallo cated from
in cu m b en t firm s to en tran ts th ro u g h s p in -o ff a c tiv ity h as attra cted so m e recent
interest (see e.g. K le p p e r 2 0 0 1). To date, h ow ever, the re la tio n sh ip betw een the
a llo catio n o f fin a n c ia l reso u rces an d in d u stria l p a ttern s o f in n o v a tiv e activ ity has
b een largely ign o red .

9.2.2.3 F in a n c e , In n o va tio n , a n d the E co n o m y

C o n te m p o ra ry e co n o m ists have d evelo p ed several a p p ro a c h e s to th in k in g about the
e co n o m ics o f in n o v a tio n at the level o f the eco n o m y, th at is, to the relationship
betw een in n o va tive a c tiv ity an d e co n o m ic d ev elo p m e n t. O n e p ersp e ctiv e, that o f
te c h n o -e c o n o m ic p a ra d ig m s, b u ild s d irectly o n S c h u m p e te r’s w o rk o n business
cycles (fo r a su m m ary , see Freem an an d Lou ^a, 2 0 0 1). A se c o n d a p p ro a c h — the
n atio n a l system s o f in n o va tio n fra m e w o rk — has been d e v e lo p e d o v e r the last 25
years to exp lain c o m p a rative —h isto rical p attern s in e c o n o m ic d e v e lo p m e n t (N elson
t 993 )*
B o th ap p ro ach es share a co m m itm en t to the S c h u m p e te ria n e m p h a sis on the
stru ctu ral co m p o sitio n and evo lu tio n o f the e c o n o m y in an alyses o f econom ic
d evelo p m en t. The tw o ap p ro ach es differ, h ow ever, in th eir a tte n tio n to resource
allo catio n . In the n atio n al system s o f in n o v a tio n lite ratu re, th e p ro cess through
w h ich la b o r and capital are allocated to in n o v a tio n p ro cesses is ra re ly d iscu ssed even
 FINANCE AND INNOVATION 249

when its significance is recognized (N elson 1993: 13), In contrast, the literature on
techno-econom ic paradigm s has long been concerned with the interaction between
innovative activity and the allocation of labou r (Freem an 1977). W ith the recent
publication o f Technological Revolutions and Financial Capital: The Dynamics of
Bubbles and Golden Ages b y C arlota Perez (2002), the relationship between finance
and tech n o-econ om ic paradigm s has also received system atic treatm ent.
Perez follow s Schu m peter in placing technological revolutions, that is, “ clusters o f
radical in n ovation fo rm in g successive and distinct revolutions that m odernize the
whole productive structure,” at the heart o f her theory. H ow ever, her concept o f a
revolution places m uch greater em phasis on the diffusion o f innovation than is
found in Sch um p eter’s w o rk .4 She also em phasizes, to a m uch greater degree than
Schum peter d id in his w o rk on business cycles, that the effects o f technological
revolutions go far beyond their econom ic im pact to include “ a tran sform ation o f the
institutions o f governance, o f society and even o f ideologies and culture” (Perez
2002: 2 4 -5 ).5
Perez’s analysis o f the interaction between the financial and productive system
is also m ore com prehensive than that o f Schum peter. W hile the latter focused
prim arily on the role o f the financial system in fu nding an initial burst o f innovative
investm ent, Perez characterizes the ways in w hich the financial system m ay be
involved in the p rodu ctive system throughout the life cycle o f technological revolu­
tions. She argues that the relationship between the financial and productive sectors
alters as the eco n o m y m oves fro m one stage o f the life cycle to another.
P erezs b o o k represents an im portan t contribution to our understanding o f
finance and in n o vation . N ot surprisingly, given its originality, it suffers from certain
lim itations that w ill need to be addressed in future w ork. In particular, system atic
em pirical su p p o rt fo r som e o f the key argum ents that she m akes about the role o f
finance in fu n d in g tech nological revolution s is not pro vid ed .6
H ow ever, a sim ilar Schum peterian analysis o f finance and technology is found in
several recent em p irical articles b y B oyan Jovan o vic and his co-au th ors on the
relationship betw een the developm ent o f the stock m arket and technological revo­
lutions in the U S eco n o m y fro m the late nineteenth century until the present
(Jo van o vic and G reen w ood 1999; Jo van o vic and R ousseau 2001; H ob ijn and Jo v a ­
novic 2001). F o r exam ple, Jo van o vic and R ousseau (2001) com pare and contrast the
im plications o f the IT revolu tion w ith that o f the “ electricity-era” technological
revolution fo r the U S stock m arket. T h ey sh ow that the transition periods from
founding to listin g on the stock m arket, from incorp oration to listing and from first
product o r process in n o vatio n to listing were shorter at the beginning and end o f the
twentieth cen tu ry than in the intervening years, and argue that the explanation for
this pattern can be fo u n d in the characteristics o f technological change (Jo van o vic
and R ousseau 2 0 0 1:3 3 6 ).
250 M A R Y o ’ S U L L I V A N ___________

9 .3 In n o v a t io n a n d t h e
E co n o m ics of F in a n c e

In the previous section, I em phasized that econ o m ists o f in n o vatio n , as a general

rule, have not been greatly inspired b y S ch u m p eter’s co n cern w ith finance and
innovation. For a long tim e, there w as a general neglect o f the interaction between
finance and the real econ om y in other bran ches o f eco n o m ics. H ow ever, in recent
years that has begun to change.
Theoretical developm ents in the field o f co rp o rate finan ce have stim ulated inter­
est in the interaction between enterprise finance an d investm ent. In the literature on
econom ic grow th, the influence o f endogen ou s gro w th th e o ry h as led economists to
consider h o w and to w hat extent the fin an cial system m ig h t affect the rate and
process o f econom ic grow th. As a result, som e o f the issues that are now being
explored b y econom ists, such as the ch aracteristics an d im p o rtan ce o f venture
capital as a source o f finance, as well as the fin an cin g o f R & D -in te n siv e firms and
new ventures, are o f direct interest to those con cern ed w ith the economics of
innovation. H owever, em pirical research has lagged theoretical developm ent and
the evidence that does exist is often am b ig u ou s even o n so m e o f the basic theoretical
proposition s that have been advanced. M oreover, there are serio u s lim its to the
dom inant conceptual approaches em p lo yed b y fin a n cia l econ o m ists for under­
standing the relationship between finance and in n o vatio n .

9.3.1 The Microeconomics of Enterprise Finance

From the late 1970s, there was a m ajor shift in theoretical research on corporate finance
based on the econom ics o f inform ation. A w hole new set o f theories o f corporate
finance emerged that took as their starting assu m ption the im portance o f “infor­
mation asym m etry” 7 as a determ inant o f enterprise finance. O ne im pact o f the
grow ing influence of inform ation econom ics on corp orate finance w as a transform­
ation in the w ay financial econom ists thought about alternative sources o f finance.
There was a m ove away from the rather sim ple cost-benefit analyses o f different
sources o f finance that had been inspired by the w o rk o f M o d ig lian i and M iller (195$)*
Compaiisons o f alternative sources o f finance increasingly considered their implica-
tions for the inform ation, incentives and control o f different econom ic actors.
These developm ents stim ulated fresh treatm ents o f old d istin ction s am ong alter­
native sources o f finance; trad ition al co m p arison s betw een debt and equity and
bank and m arket-based finance, for exam ple, w ere con cep tu alized in new ways
based on the econom ics o f in form ation (A llen and G ale 2000). M oreover, sources of
finance that had previou sly received little attention fro m finan cial economists
 FINANCE AND INNOVATION 251

attracted interest given w hat appeared to be their distinctive characteristics from an

in form ation -econ om ics perspective; the best exam ple is the burgeoning literature
on venture capital.
A n article b y W illiam Sahlm an (1990) pioneered in using the logic o f asym m etric
inform ation as the fo u n d atio n for a theory o f venture finance. Sahlm an identified
several different m echanism s, co m m o n ly em ployed in the process o f venture capital
investing, that he claim ed venture capitalists used to overcom e the problem s o f
financing ventures in the presence o f asym m etric inform ation. In subsequent w ork,
other scholars delved fu rther into the roles played by these m echanism s, such as
staging, com pensation, m o n itorin g and control o f investee com panies, and exit, in
the relationship betw een venture capitalists and their investee com panies (for a
sum m ary o f this research, see G om pers and Lerner 1999).
Im portant theoretical advances have been m ade in analyzing the econom ics o f
alternative sources o f enterprise finance em pirical analysis has lagged behind. Even
basic evidence on the relative im portance o f different sources o f enterprise finance,
such as internal and external sources o f funds, as well as alternative sources o f
external funds like banks, stock and bon d m arkets, and venture capitalists is m odest
(see B o x 9.1 for a su m m ary h isto ry o f the U S venture capital industry). M ore

Box 9.1 Venture capital in comparative-historical perspective

The United States is the country most closely associated with a professional venture
capital industry. Although individuals and families in the United States, as in most
industrial economies, long have used their private fortunes to fund new ventures, the
US venture capital “industry” began with the founding in 1946 o f American Research
and Development (ARD) for the finance o f new ventures. ARD was founded in Boston
by members o f the local investment community, as well as professors and adminis­
trators from MIT. With the expansion o f federally funded research at M IT during
World War II, the principal sources o f ARD’s deal were the federally funded laborator­
ies at Harvard and MIT. By far its most successful deal was its 1957 investment in a new
firm founded by Kenneth Olsen o f MIT's Lincoln Laboratory, the Digital Equipment
Corporation (DEC) (Rosegrant and Lampe 1992: 72, 110- 14).
The enormous success ofA R D 's investment in DEC— when the computer firm went
public in 1966 ARD's original investment o f $ 70,000 was valued at $37 m illio n -
generated huge interest in venture capital in the financial community. By that time,
a new breed of financiers had emerged out of Silicon Valley's high-technology enter­
prises. Successful entrepreneurs like Fairchild's Eugene Kleiner and Don Valentine
reinvested the capital that they accumulated in promising local start-ups and brought
cash and technical skill, operating experience, and networks of industry contacts to the
ventures they funded. Silicon Valley's venture capitalists were heavily involved with
their ventures, advising entrepreneurs on business plans and strategies, helping find
: -investors, recruiting key managers, and serving on boards o f directors (Saxenian
252 MARY O SULLIVAN

Box 9*1 (continued)

Several regulatory changes contributed to the further expansion o f the US

capita! industry in the 1980s and 1990s, Particularly important were a serife #
legislative initiatives in the late 1970s that made venture capital a much more attractive
investment option, the capital gains tax rate was reduced from 49.5 to 20 per cent and it
became much easier for pension funds to invest in venture capital partnerships. There
was a major increase in the flow o f money into the US venture capital from then on (see
fig л л in Gompers and Lerner 1999: 7). Historical trends in investments and disburse­
ments by the industry displayed considerable volatility, with m ajor boom -bust cycles
recorded in the 1960s, the 1980s and again in the late 1990s/early 2000s.
Investments by the US venture capital industry were highly concentrated by sector. For
the period from 1965 to 1992, four industries— office and computing machinery, com­
munications and electronics, pharmaceuticals, and scientific instruments— accounted
for 81 per cent o f all investments by the US venture capital industry. The relative import­
ance o f these sectors changed considerably over time; office and computing machines
became less important over time whereas the drugs sector increased its share o f venture
capital investment. In the 1990s, especially in the second half o f the decade, Internet-related
investments were estimated to account for 70 per cent and 75 per cent o f all venture capital
investments in 1999 and 2000 respectively (Venture Economics).
Measured as a share o f gross economic output, the US venture capital industry is the
largest in the world. The venture capital industries o f other countries enjoyed consider­
able growth in the 1990s; for the EU as a whole, for example, venture capital investment
increased from 0.04 to 0.12 per cent o f GDP for the period from 1989 to 1999 with
Belgium, Sweden and the Netherlands recording particularly strong expansion
(Table 9.1). Nevertheless, even more rapid growth in the US venture capital industry
during this period meant that it actually increased its lead over other countries,
particularly in early-stage investment activity by venture capitalists.

challenging questions, such as w hether enterprises ch oo se alternative sources of

finance for the reasons posited by financial econ om ists o r w h eth er their use o f these
sources has the im plications for perform an ce that theories o f enterprise finance
suggest, rem ain largely unexplored (R ajan and Z in gales 19 9 5 :14 2 1).
1 he grow ing influence o f in form ation econ o m ics stim u lated another body of
theoretical analysis o f enterprise finance, en cou ragin g interest in the long-neglected
interaction between the financing and investm ent b eh a vio r o f enterprises. Since
inform ation asym m etries between enterprises and fin an ciers m ad e external finance
m ore expensive for com panies than internal finance, ceteris p aribu s, enterprises with
access to substantial liqu id ity from internal sources w o u ld invest m ore than enter­
prises that have to resort to external finance,
There is now a substantial b od y o f em pirical evidence that suggests that “ liquidity
constraints’ m atter to capital investm ent (for a su m m a ry see H u b bard 1998:199)*
H owever, serious m ethodological concerns have been raised ab ou t these studies.
 Table 9.1 Venture capital and early-stage investment as a percentage of GDP

Country V C Investm ent as percentage Early-Stage Investm ent as

of GDP percentage o f G DP

1989 1999 1989 1999

Austria 0.01 0.03 0.006 0.007

Belgium 0.05 0.26 0.015 0.093
Denm ark 0.01 0.05 0.009 0.019
Finland 0.01 0.11 0.003 0.057
France 0.05 0.12 0.009 0.039
Germ any 0.01 0.13 0.004 0.051
Greece n.a. 0.06 n.a. 0.017
Ireland 0.05 0.09 0.002 0.048
Italy 0.02 0.05 0.002 0.014
The Netherlands 0.05 0,25 0.006 0.096
Portugal 0.02 0.05 0.004 0.008
Spain 0.02 0.09 0.009 0.018
Sweden 0.02 0.19 0.004 0.113
United Kingdom 0.13 0.20 0.023 0.020
EU 0.04 0.12 0.008 0.036
us 0.11 0.59 0.027 0.056

Source: Christofidis and Debande, 2001, p, 20, p. 23.

Particularly controversial has been the com m on practice o f im pu tin g the im p o rt­
ance o f liq u id ity constraints from the sensitivity o f a com pany's investm ent to its
cash flow. Som e scholars claim that “ investm ent-cash flow sensitivities provide no
evidence o f the presence o f financing constraints" thus calling into question the
valid ity o f the em pirical results reported in this literature (K aplan and Zingales 2000;
Fazzari, H ubbard, and Petersen 2000).
Theories that posited that the influence m ight go in the other direction, that the
characteristics o f enterprise investm ent m ight influence enterprise finance, were also
developed. O ne stream o f literature posited a relationship between the types o f
activities in w h ich enterprises invested and their financial behavior. A n oth er sug­
gested that characteristics o f the investing enterprise, for exam ple, its stage o f
developm ent, m attered to the w ay it w as financed.
In analyses o f the influence o f enterprises' activities on their financing, R M >
investm ents attracted p articu lar attention since they were deem ed to create acute
in form ation asym m etries between corporate m anagers and financiers (Bah and
D um on tier 2001: 675; H im m elb erg and Petersen 1994; H all 2002). The m ain im p li­
cation that has been d raw n fro m such analyses is that the gap between the costs ot
financing R & D investm ent fro m internal and external sources should be greater than
254 MARY O SULLIVAN

for other form s o f investment* Therefore, R & D -in te n siv e firm s sh ou ld be more
inclined than other firm s to rely on internal fu n d s to fin an ce their investments*
M oreover, financing constraints arising fro m im p erfectio n s in capital markets
should have a m uch greater im pact o n R & D , than other, investm ents (H all 2002;
C arpenter and Petersen 2002: F55). Recently, so m e sch olars have suggested that these
argum ents ap p ly not just to R & D investm ents b u t to all investm ents in high-
technology industries (see Bank o f England 2001: A n n e x, 81—5).
E m p irical analyses o f the relationship betw een finan ce an d R & D are primarily
analyses o f the links between cash flo w and R & D exp en d itu res. T h e com m on finding
o f these studies is that R & D investm ent is indeed p o sitiv ely correlated w ith cash flow
(for a su m m ary see H all 2002). H ow ever, these studies are subject to the same
m ethodological criticism s as the em pirical w o rk o n liq u id ity constraints and capital
investm ent to w hich we have already referred, and until these issues are resolved we
cannot be confident o f their findings.
Financial econom ists have also begun to analyse w h eth er the characteristics of
investing firm s m ight m atter to their fin an cin g beh avior. P articu lar attention has
been paid to a com pany's stage o f d evelopm ent an d the co n cep t o f a “ financing
grow th cycle" is n o w w idely used to characterize the challenges fo r firm s as they
evolve from new venture to going concern h ig h lig h tin g once again the extent of
inform ational asym m etry involved (Berger and U dell 1998: 622). Perhaps the most
straightforw ard im plication o f this type o f analysis is that firm s at earlier stages o f the
cycle, such as start-up com panies, are likely to have d ifficu lties raising external
finance. A s a result, they should be m ore h eavily d ep en d en t o n insider finance
than firm s at later stages o f developm ent.
Once again, however, em pirical research has n ot kept pace w ith theoretical
developm ents. The evidence that has been com piled , m oreover, does not provide
clear su p port for som e o f the m ost basic p ro p o sitio n s ad van ced in the theoretical
literature. In this regard, Berger and U dell h igh light tw o fin d in gs fro m their empir­
ical analysis w hich seem particu larly surprising. T h e first is that the funds provided
by the principal ow ner are m ore im portan t as the firm gets o ld e r th an at early stages.
Second, their evidence suggests that finance fro m in siders never outw eighs that
provided b y outsiders even for the youngest firm s (B erger and U dell 1998: 625).
T h e tw o stream s o f literature on the influence o f investm ent characteristics on
enterprise finance the one on characteristics o f the investm ents being made and
the other on the characteristics o f investing firm s— have recently been brought
together in research on the financing o f sm all firm s m ak in g R & D , high-technology,
or technology-based investm ents, often referred to as tech n o logy-b ased small firms
(T B SFs). Theoretical m odels from the “ n e w " co rp o rate finan ce predict that these
firm s will be m uch m ore tightly constrained b y th eir o w n internal resources in
financing their investm ent than other firm s. A s yet, how ever, the ju r y is out on
whether even this basic proposition is b orn e out b y em p irical evidence. As a recent
report by the Bank ot England concluded: “ the evidence fro m such studies is
 FINANCE AND INNOVATION 255

conflicting o n the key issue o f w hether T B SF s face greater difficulties in accessing

finance than SM E s [sm all and m edium enterprises] in general” (Bank o f England
2001: 83)*

9.3.2 Financial Systems and Economic Growth

C o n tem porary grow th theory, in recognizing the im portance o f technological
change for econ om ic grow th, echoes Schum peter’s em phasis on the im portance o f
innovation as the p rim a ry im petus for the process o f econom ic developm ent.
M oreover, w hereas Schum peter em phasized the im portance o f financial system s in
fueling in n ovation and, therefore, econom ic developm ent, the idea that the devel­
opm ent and structure o f a co u n try ’s financial system m ight have im plications for its
econom ic grow th was treated w ith skepticism or indifference by m ost m acroecon o ­
mists in the second h a lf o f the twentieth century. In recent years, however, there has
been a m ajor increase in interest in the relationship between financial developm ent
and econom ic grow th, a trend that is usually attributed to the influence o f endogen­
ous grow th th eo ry in m acroecon om ics. A variety o f theoretical articles m odel the
m echanism s th rou gh w h ich the financial system m ight affect lon g-ru n growth.
Em pirical studies have also been undertaken, p rim arily based on large-scale cross­
country regressions, to analyze the relationship between financial system s and
econom ic grow th.
M ost o f these studies o f finance and grow th seek to relate the level o f developm ent
and structural characteristics o f the financial system to aggregate econom ic activity.
Consistent w ith the d o m in an t approach taken in neoclassical grow th theory, eco­
nom ic d evelopm ent is u nderstood as an undifferentiated quantity generated b y an
aggregate p ro d u ctio n fu nction. A s a result, con tem p orary analyses o f finance and
growth m ake no reference to the structural com position and evolution o f the
econom y and, therefore, ign ore w hat Schum peter regarded as the essential charac­
teristic o f the process o f econ o m ic developm ent, that is, its lum piness over tim e and
across sector.
H owever, there are a few exceptions to the general rule. Rajan and Zingales
(1998) differentiated am o n g industries in term s o f their investm ent and financing
behavior, argu in g that their financial requirem ents are technologically deter­
mined:

there is a technological reason why some industries depend more on external finance than
others. To the extent that the initial project scale, the gestation period, the cash harvest
period, and the requirement for continuing investment differ substantially between indus­
tries, this is indeed plausible. (Rajan and Zingales 1998: 563)
256 MARY O’ S ULLIV A N

T h eir m ain hypothesis is that industries that are m o re dep en d en t on external

finance sh ould grow faster in countries w ith m o re d evelo p ed finan cial markets.
Several other studies have follow ed R ajan an d Z in g ales in d iscrim in atin g among
industries in term s o f their dem and for finance. Som e o f them go fu rth er to consider
w hether financial structure— notably w hether a fin an cial system is m arket- or bank-
based— m atters to the developm ent o f d ifferent in d u stries (B eck an d Levine 2002;
D em irginpK unt an d M ak sim o vic 2002; C arlin an d M ayer 2003).
The central im plication o f this disaggregated a p p ro ach to the relationship be­
tween finance and grow th is that the econ om ic im p act o f fin an cial systems maybe
reflected not o n ly in aggregate rates o f econ o m ic grow th b u t also in the differential
developm ent o f particu lar industries. D ifferences in the grow th trajectories of
particular industries w ill in turn be reflected in v ariatio n s across country in the
co m p o sitio n o f econom ic grow th; sectors favou red b y a n atio n s financial system
w ill becom e m ore and m ore pro m in en t in the e c o n o m y o ver tim e w hile other, less-
favored sectors w ill languish o r fail to develop. H ow ever, as yet, the task o f analyzing
the relationship between financial system s and the stru ctu ral evolution o f the
econ om y has not been treated in any detailed w ay in these studies.

9 .4 A N ew A gend a for R esear ch

on F in a n c e and In n o v a t io n

C o n tem porary research in econom ics has a lo n g w ay to go before it can help us to

understand the relationship between finance and in n ovation , In the economics of
innovation, that relationship has been largely neglected though questions about
resource allocation are latent in existing research. In finan cial econom ics, some
scholars have begun to explore concerns that are directly relevant to the study of
finance and innovation, but financial econom ists have been m ore effective at general-
ing new theoretical argum ents than adducing em pirical evidence to support them.
In principle, there is a good case for integration o f the tw o fields as a route to new
insights on the relationship between finance and in n o va tio n . H ow ever, there are
serious lim itations o f the dom in an t theoretical ap p ro ach es that finan cial economists
em ploy for analyzing the process o f econom ic change. M o reo ver, barriers o f mutual
ignorance and, m ore fundam entally, m eth o d o logical d ifference, m ake such integra­
tion unlikely. C ollaboration between econom ists o f in n o va tio n and historians of
finance is a m uch m ore prom isin g path to a better u n d erstan d in g o f the relationship
bet ween finance and innovation. These scholars share a co m m itm en t to history as a
technique o f econom ic analysis which is crucial given that the essential processes
 FINANCE AND INNOVATION 257

that need to be u nderstood in analyzing the interaction between finance and

innovation are h istorical processes.

9.4.1 Finance Theory and the Dynamics of Economic Change

Im portant weaknesses o f the research o f financial econom ists becom e evident when
we bring the literature on the econom ics o f finance into contact with what we know
about the dynam ics o f innovation. As a result, it is doubtful that, left to their ow n
devices, financial econom ists w ill ever develop a satisfying understanding o f the role
o f finance in the process o f innovation. In the m icroeconom ic literature, the m ain
problem is the centrality o f the concept o f asym m etric inform ation in theories o f
enterprise finance. It is true that analytical space for the “ real” econom y was created
by incorporating this concept; so long as one could argue that a real phenom enon led
to im portant asym m etries o f in form ation , financial econom ists were w illing to take
it seriously in theories o f enterprise finance. H ow ever, the versatility o f the concept
hints at its lim itations.
One problem is that it has been invoked in a rather casual m anner w ithout m uch
in the w ay o f p r o o f o f the extent o f asym m etries o f in fo rm atio n or their im portance
in influencing econ om ic relationships. Fo r the case o f R & D investm ents, for
example, one typ ically finds an assertion that they are subject to acute problem s o f
asym m etric in fo rm atio n . Yet, w hy shou ld we assum e that investors have less in fo r­
m ation about the likely success o f pharm aceutical com panies' R8dD efforts than the
factors that determ ine the p ro d u ctivity o f a new autom obile plant?
M ore fundam entally, it is not d e a r that privileged access to inform ation by som e
econom ic actors is the m ajo r determ inant o f the challenges for enterprises in
financing such investm ents. W hen m aking innovative investm ents, a m ore im p o rt­
ant challenge than asym m etric in fo rm atio n is the fundam ental uncertainty that
characterizes the relationship between investm ents and their outcom es. In an
environm ent characterized b y fundam ental uncertainty, the crucial problem is not
that one person kn ow s som eth ing w hereas another does not; rather the challenge to
decision m aking is ignorance, the fact that n o b od y really know s anything.
U ncertainty in this sense is different from the concept as it is used in neoclassical
econom ics (A rro w and D ebreu 1954; A rro w 1974) where the m ain concern is w ith
param etric uncertainty.8 T h e environ m en t in w hich econom ic decisions are m ade is
characterized as a set o f m u tu ally exclusive but collectively exhaustive possible states
o f the w orld. In such a w orld , one that is closed and determ inistic, rational decisions
based o n p ro b ab ilistic estim ates are a reasonable basis for action.
W hen u n certain ty is fu ndam ental, as it is when in n ovation occurs, econom ic
agents are uncertain n o t ju st about w hich possible state w ill obtain but about w hich
ones are even possible. In making innovative investments, therefore, there are really
258 MARY О*SU LLIV A N

no objective guidelines for m aking decisions o r fo r reso lv in g disputes. So bow docs

anyone act under these circum stances? H o w do in n o vative investm ents ever get
m ade given the uncertainty that su rrou n d s them ?
Q uestions about decision m aking u n d er fu n d am en tal u n certain ty are further
com plicated by the fact that the process o f in n o vatio n reveals new, possible states of
the w orld (K line and R osenberg 1986: 29 7-8 ). In oth er w o rd s, the uncertainty
inherent in innovation unfolds as the process evolves. A s a result, the future state
o f the w orld cannot be defined until it is discovered th ro u g h the process o f innov­
ation (Rosenberg 1994: 53-4 ). T h rou gh in volvem en t in that process, or close famil­
iarity w ith it, decision m akers learn and, as th ey d o so, their perceptions of the
possibilities and problem s o f innovative investm ent ch an ge. T h ere is no reason to
assum e that, in altering their perceptions based on n ew in form ation , decision
m akers necessarily m ove closer to som e kin d o f tru th . To the contrary, they are
likely to m ake m an y m istakes in interpreting w h at the in fo rm a tio n that they have
im plies for the challenges and o p p ortu n ities o f in n o vatio n .
Econom ists w ho take seriously the fu n d am en tal u n certain ty that surrounds
innovative investm ent have tended to em phasize the im p o rtan ce o f subjective
judgem ents, based on perceptions and b e lie f system s, fo r d ecision making. They
have also suggested that, in reacting to the u n fo ld in g o f the u n certain ty inherent in
innovation, decision m aking is experien tial as w ell as interpretative. Financial
econom ists tend to overlook the basic cognitive ch aracteristics— ignorance, learning
and error— o f resource allocation and the m ech an ism s that allo w enterprises to
m ake com m itm ents o f resources to in n o vative a c tiv ity n o tw ith stan d in g the chal­
lenges o f d oin g so. To date, in analyzing enterprise finan ce, fin an cial economists
have em phasized rationality, indeed a rather lim ited an d static concept o f rationality,
to the exclusion o f the subjective and experien tial d im en sio n s o f decision making.
W hile rational analysis m ay well feed into the process o f in n o vative resource alloca­
tion, it cannot dom inate it.
I f financial theory is o f lim ited use for u n d erstan d in g the cognitive challenges
associated w ith innovative investm ent, it p ro vid es no help at all in dealing with other
im portant features o f firm -level in n ovation such as its o rgan ization al character.
N otw ithstan ding the im portan t progress that has been m ad e in m icroeconom ics in
conceptualizing the econom ics o f the firm , theorists o f co rp o rate finance have failed
to em bed their analyses in a substantive th eo ry o f the firm (Z in gales 2000). In fact,
m any financial econom ists rem ain w edded to the idea o f the firm as a nexus of
contracts and reject the n otion o f firm s as o rgan izatio n s that have a logic that is
distinct from that o f m arkets.
As far as the m acroecon om ic literature is con cern ed, its m ain problem for
understanding the relationship between finance and grow th is that it conceptualizes
it in a rather m echanical way. M ost o f the m odels used in this literature treat
econom ic growth as if it was generic across the e co n o m y an d over tim e. Typically
they m ake no allow ance for variation and change in the organizational and
 FINANCE AND INNOVATION 25 9

institutional contexts in w h ich financial resources are allocated and em ployed to

facilitate econom ic grow th. Instead, the relationship between finance and grow th is
understood in term s o f the influence o f the quantity o f finance provided and, to a
lesser extent, the price at w hich it is supplied, on the “ am o u n t” o f econom ic grow th.
In contrast, research on innovation and technological change suggests the central
im portance o f the context in w hich these processes occur to their im pact on
productivity and econ om ic developm ent. It is for this reason that G avin W right, a
prom inent econom ic historian, has recently exhorted econom ists to em brace a [a]
conception o f tech nology that is historically-contingent and institutionally-
specific” (W right 19 9 7 :15 6 2 ). From this perspective, research on finance and in n o v­
ation m ust devote greater attention to the role that contextual factors play in shaping
financial relationships. W ho gets financial resources, when they get them , h ow they
use them and other factors that can o n ly be identified by m aking qualitative
distinctions am o n g enterprises, tim e periods, and investm ents, are likely to be
more im portan t than the overall quantity o f financial resources that is invested in
an industry.
Recent studies that in corporate sectoral differences in their analyses o f finance
and grow th tend to o verlook the im portance o f these contextual factors. They
assume that given characteristics o f technologies m atter to the scale and tim ing o f
enterprise investm ent and, as result, to enterprise dem and for finance. However,
even if w e take technological characteristics as given, a range o f variables, such as the
balance between entrants and incum bents and, m ore generally, the com petitive
structure o f the indu stry, defines the context that shapes the relationship between
technology, investm ent requirem ents and financing needs. M oreover, the relation­
ship between tech n o logy and finance is likely to change within industries over tim e,
in part because tech nological characteristics o f industries evolve, but also since the
intervening variables that determ ine their relationship to finance also change.

94.2 The Possibilities and Problems of Integration

between Fields
Some of the w eaknesses o f financial econom ics fo r dealing w ith the relationship
between finance and the dynam ics o f econom ic change could be overcom e, at least in
principle, through closer integration with research on innovation. There is some
recognition of this fact within financial economics. One prominent scholar in the
field, Luigi Zingales, has been quite explicit in his call for unew foundations, notably
a more sophisticated analysis of the firm, fo r the study of enterprise finance.
Particularly interesting from my perspective is the fact that he emphasizes the
importance o f aa theory of entrepreneurship” for understanding how firm s innovate
(Zingales 2000).
2бО MARY O ’ SULLIV A N

Econ om ists o f innovation, in turn, can learn fro m research iti finance. In part,
their having greater contact w ith financial econ o m ists m ig h t stim ulate a general
interest in finance am on g them but the specific details o f existin g research can also be
a source o f useful insights. For exam ple, the recent atten tion b y fin an cial economists
to the relationship between stages o f an enterprise s d evelo p m en t and its financing
requirem ents could be fru itfu lly in corp orated b y eco n o m ists o f innovation.
I f the o n ly b arrier to intellectual integration w as lack o f fam iliarity with each
other's w ork then it w ould be relatively easy to o vercom e. H ow ever, ignorance of
another's field o f in q u iry typically conceals deeper barriers. In add ition to theore­
tical differences described above, m eth o d o logical d ifferences pose a formidable
barrier to the integration o f research o n finance and in n o vatio n .
Perhaps the m ost im portan t m eth odological obstacle to cross-fertilization be­
tween the fields o f in n ovation and finance is the different p rio rities that they assign
to h isto ry in econom ic analysis. In recognition o f the im p o rta n ce o f change over
tim e in the process that they study, econom ists o f in n o va tio n o ften describe them­
selves as "ev o lu tio n ary econom ists” and they typ ically assign considerable import­
ance to historical research in their econ om ic analysis. In th eir em pirical work, they
tend to em phasize the im portan ce o f qualitative variatio n s across the econom y and
over tim e in innovation and technological change. T h e y are often skeptical of
theories o f innovation and econom ic d evelopm en t that are abstracted from the
historical contexts and em phasize the im p o rtan ce o f “ reasoned h isto ry ” and “ his­
to ry-frien d ly m odels” (Freem an and L o u $a 2001; L azo n ick 19 9 1; M alerb a et a l 1999).
In contrast, m ost financial econom ists neglect the fact that capitalism is an
evo lu tio n ary process. E q u ilib riu m analysis o verw h elm in g ly d om in ates theoretical
research to the neglect o f the historical origin s and evo lu tio n o f econom ic behavior.
In em pirical research, quantitative analysis is the n o rm an d attention to qualitative
variation s w ithin the econom y and over tim e is m odest. It is h ard ly surprising,
therefore, that lim ited attention is paid in theoretical o r em p irical research to the
historical developm ent o f the financial system , the stru ctu ral evolution o f the
econ o m y o r change over tim e in the relationship betw een the finan cial system and
the real econom y.
It seems unlikely that such m ethodological differences can be overcom e in the
foreseeable future. A m ore p rom isin g avenue to intellectu al progress on the rela­
tionship between innovation and finance, therefore, is fo r econ o m ists o f innovation
to collaborate with financial historians. E vo lu tio n ary econ o m ists have already had
direct experience o f the value o f a shared m eth o d o logical co m m i tm ent to historical
analysis in facilitating intellectual exchange th ro u gh their co llab o ration in research
on technological change w ith historians o f technology, labor, and business. Similarly
fruitful exchanges are likely to result fro m interaction w ith eco n o m ic and business
historians who study the h isto ry o f financial system s. In deed, there are already
som e prom isin g signs o f research in this d irection (see e.g. L a m o reau x and Sokoloff
2004).
 FINANCE AND INNOVATION l 6l

9.5 C o nclusio n

What are the priorities in new research on finance and innovation? It is w idely
recognized that the p au city of, and flaws in, existing em pirical research on the
financial system are m ajo r barriers to econom ic analyses o f its role. The dearth o f
evidence is a p rob lem even i f w e focus on con tem p orary patterns o f financial
dem and and supply. W hen we lo o k to historical data, the holes in our understanding
o f these patterns are even m ore gaping (Levine 2003; O 'Sullivan 2004a, b; Zingales
2003).
It is true that patterns o f financial dem and and supply by enterprises, firm s, and
econom ies serve o n ly as backgrou n d evidence for m ore specific em pirical questions
about the relationship betw een finance and innovation. In an ideal w orld, scholars
who are p rim arily interested in researching that relationship could com pile these
data from readily available sources. H ow ever, the current, rather dire, state o f
em pirical research on patterns o f financial dem and and supply m ean that this option
is not available.
It is hard to see, therefore, h o w substantial progress in our understanding o f
finance and in n o vation can be m ade unless scholars w ho are interested in the subject
are w illing to contribu te to the com pilation o f these data so that basic em pirical
questions can be addressed. A t a m in im u m , we need evidence that allow s us to
identify variation and change in the historical patterns o f financing b y firm s, in d u s­
tries, and nations. We also need to understand long-term trends in the supply o f
resources by different finan cial institutions. For exam ple, how im portant has the
stock m arket been as a source o f finance for corporate investm ent in different
econom ies at vario u s tim es? W hat types o f industries and firm s did it finance during
different periods? A re there m arked differences in the characteristics o f firm s and
industries fu n ded b y b on d m arkets and banks? (For the weaknesses o f existing
em pirical research on these questions, see O 'Su llivan 2004a, b.)
O nly when w e have identified these basic patterns in financial dem and and supply
can we hope to understand h ow they are related to the dynam ics o f econom ic
change. For exam ple, in the case o f a p articu lar in d u stry we m ight ask how patterns
o f financial dem and and su p p ly are related to the balance between incum bents and
entrants in in n o vative activity. F o r an econom y, the relationship between financial
patterns and changes in the structure o f the econom y, such as the shift from
m anufacturing to services, are clearly relevant. For understanding these types o f
interactions betw een finance and the dynam ics o f econom ic change, em pirical
studies that treat the d em an d for, and su p p ly of, finance as characterized prim arily
by quantity and price can o n ly take us so far. D etailed case studies are needed to
generate insights o n the qu alitative dim ensions o f these interactions that could then
be used to illustrate, enrich and perhaps confront som e o f the im plications that can
be draw n from q u an titative studies.
262 M A R Y o ’ s U L L l VAN

ОС course» progress in the analysis of finance end innovation cannot be made

based on empirical research alone. At a minimum» empiricel studies need to be
theoretically informed in terms of the questions that they pose. Therefore» a certain
amount of conceptual development on the relationship between finance and
innovation will be necessary to facilitate empirical studies on the subject. One
promising route to new ideas is the further development of the implications
for resource allocation of observed characteristics of innovation at the level of
the firm» the industry and the economy along the lines discussed above. However»
theoretical research needs to go farther than an analysis of the implications of the
characteristics of innovation for finance to consider how the structural characteris­
tics and evolution of the financial system influence innovative activity in the
real economy.

N o t e s

1. The author gratefully acknowledges the helpful critiques and suggestions that she received
from the participants in the various workshops organized to produce this book. She
would like to thank Jan Fagerberg» Bronwyn Hall» Bill Lazonick» David Mowery, and
Richard Nelson for the particularly detailed comments that they provided on earlier drafts
o f this chapter as well as Ron Adner and Bruce Kogut, her colleagues at INSEAD» for their
helpful suggestions.
2. The term “ enterprise finance” is used herein to refer to research on corporate and venture
finance.
3. The other major exception is Alfred D. Chandler. From a theoretical perspective» much of
his work can be seen as contributing to a similar concept o f the firm as that o f Penrose.
Indeed» in recent years» in reflecting on the general implications o f his research» Chandler
has tended to use the language o f organizational learning (see e.g. Chandler 1992).
4. This is true of most of the contemporary literature on techno-economic paradigms
(Freeman and Lou^a 2001:149)
5. In CSDy however» Schumpeter was centrally preoccupied with the social and political» as
well as the economic, implications o f changes in the characteristics o f the innovation
process.

6. Indeed, Perez explicitly portrays her book as “ a Think-piece»' the spelling out of an
interpretation, with enough illustrations to strengthen the case and stimulate discussion
(Perez 2002: xix).
/, In economics, the term information asymmetry” is used to refer to a situation in which
one economic agent has more information than another about a task that affects both
agents' welfare.
8. For general discussions o f the difference between structural or radical uncertainty and
param etric uncertainty, see Loasby 1976; CTOnscoH and Rizzo 1985; Langlois 1986; Shackle
1992.
 FINANCE AND INNOVATION 263

References

Comparing Financial Systems, Cambridge, Mass.: MIT Press.

A l l e n , F., and G a l e , D . (2000),
A rro w , K. (1974), The Limits of Organization, New York: Norton.
------ and D e b r e u , G. (1954), “ Existence o f an Equilibrium for a Competitive Economy,”
Econometrics 22: 265-90.
B ah , R.> and D u m o n tier , P. (2001), R 8dD Intensity and Corporate Financial Policy: Some
International Evidence, Jourttcil of Business Finance and Accounting 28(5/6): 671—92.
Bank o f England (2001), Financing of Technology-Based Small Firms, London: Domestic
Finance Division, Bank o f England, February.
B eck , X , and L e v in e , R. (2002), “ Industry Growth and Capital Allocation: Does Having a
Market- or Bank-Based System Matter?” Journal of Financial Economics 64(2): 147-S0.
B erg er , A „ and U d ell , G. (1998), “ The Economics of Small Business Finance: The Roles o f
Private Equity and Debt Markets in the Financial Growth Cycle,” Journal of Banking &
Finance 22(6-8): 613-73.
B o w e r , J. (1970), Managing the Resource Allocation Process: A Study of Corporate Planning
and Investment.t Cambridge, Mass.: Division o f Research, Graduate School o f Business,
Harvard University.
B urgelm an , R., and S a y l e s , L. (1986), Inside Corporate Innovation: Strategyy Structure, and
Managerial Skills, New York: Free Press.
*B yg ra v e , W., La n g e , J., R oedel , J. R., and Wu, G. (2000), “ Capital Market Excesses and
Competitive Strength: The Case o f the Hard Disk Drive Industry, 1984-2000,” Journal of
Applied Corporate Finance, 13(2): 8-19.
C a r lin , W., and M a y e r , C. (2003), “ Finance, Investment and Growth,” Journal of Financial
Economics 69:191-226.
^Ca r p en t er , M., L a zo n ic k , W„ and O’S u lliv a n , M. (2003), “ The Stock Market and
Innovative Capability in the New Economy: The Optical Networking Industry,” Industrial
and Corporate Change 12(5): 963-1034.
^C a r p e n t e r , R., and P e t e r s e n , B. (2002), Capital Market Imperfections, High-Tech Invest­
ment, and New Equity Financing,” Economic Journal 112: F54-F72.
C h a n d l e r , A. (1992), “ Organisational Capabilities and the Economic History o f the Indus­
trial Enterprise,” Journal of Economic Perspectives 6(3): 79-100.
C h risto fid is , C , and D e b a n d e , O. (2001), “ Financing Innovative Firms Through Venture
Capital,” EIB Sector Papers.
D emirguc -K u n t , A., and M a k sim o v ic , V. (2002), “ Funding Growth in Bank-Based and
Market-Based Financial Systems: Evidence from Firm-Level Data,” Journal of Financial
Economics, 65: 337-63.
Fa zza ri , S., H u b b a r d , G., and P e t e r se n , B. (2000), “ Investment-Cash Flow Sensitivities
are Useful: A Comment on Kaplan and Zingales,” Quarterly Journal of Economics 115: 695-
705.
F r e e m a n , C. (1977), “ The Kondratiev Long Waves, Technical Change and Unemployment,”
in Structural Determinants of Employment, vol. 2, Paris: OECD, 181-96.
------ and LougA, E (2001), As Time Goes By: From the Industrial Revolutions to the Infor­
mation Revolution, Oxford: Oxford University Press.
G o m p e r s , P., and L e r n e r , J. (1999)* The Venture Capital Cycle, Cambridge, Mass.: M IT Press,

* Asterisked items are suggestions for further reading.

264 M A R Y O S U L L I V AN

C an s . J., H su , D., and Stern , S. (2000), “ When Does Start-Up Innovation Spur the Gale of
Creative Destruction?” NBER Working Paper 7851. ^
♦ H all, B. (2002), “The Financing of Research and Development,” NBER, Working Paper
8773.
H im m elb er g , C., and P etersen , B. (i994)> R&D Internal Finance. A Panel Study of Small
Firms in High-Tech Industries,” Review of Economics and Statistics 76(1): 38-51.
H o b ijn , B., and J o vano vic , B. (2001), “ The Information-Technology Revolution and the
Stock Market: Evidence ,” American Economic Review 91(5): 1203-20.
H u bba r d , G. (1998), “ Capital-Market Imperfections and Investment” Journal of Economic
Literature 36(1): 193-225.
J o vano vic , B., and G reenw ood , J. (1999)> “ The Information-Technology Revolution and
the Stock Market,” American Economic Review 89(2): 116-22.
* ---- an<3 R o u sseau , P. (2001), “ Why Wait? A Century o f Life before IPO,” American
Economic Review 91(2): 336-41.
K a pla n , S., and Z in g a les , L. (2000), “ Investment-Cash Flow Sensitivities are not Valid
Measures o f Financing Constraints,” Quarterly Journal of Economics 115: 707-12,
K l e p p e r , S. (2001), “ Employee Startups in High-Tech Industries,” Industrial and Corporate
Change 10: 639-74.
K l in e , S., and R o sen berg , N. (1986), “An Overview o f Innovation ” in R. Landau and N.
Rosenberg (eds.), The Positive Sum Strategy: Harnessing Technology for Economic Growth)
Washington, DC: National Academic Press, 275-305.
* Lam o reau x , N., and S okoloff , K. (eds.) (2004), The Financing of Innovation in Historical
Perspective) Cambridge, Mass.: M IT Press.
Lan g lo is , R. (1986), “ Rationality, Institutions and Explanation,” in id., Economics as a
Process: Essays in the New Institutional Economics, Cambridge and New York: Cambridge
University Press, 225-55.
Lazo n ick , W. (1991), Business Organization and the Myth of the Market EconomyyNew York:
Cambridge University Press.
L e v in e , R. (2003), “ More on Finance and Growth: More Finance, More Growth?” Federal
Reserve Bank o f St. Louis, July-August, 31-46.
L o asby , B. (1976), Choice, Complexity; and Ignorance: An Inquiry into Economic Theory;
Cambridge and New York: Cambridge University Press.
M a l e r b a , R, N elso n , R., O rsen igo , L., and W in t e r , S. (1999), “ ‘ History-friendly’ Models
o f Industry Evolution: The Computer Industry,” Industrial and Corporate Change 8:3-40*
M c M illa n , G. S., N a r in , R, and D eed s , D. (2000), “An Analysis o f the Critical Role of
Public Science in Innovation: The Case o f Biotechnology,” Research Policy 29(1): 1-8.
M o d ig lia n i , F. and M il l e r , M. H, (1958) “ The Cost o f Capital, Corporation Finance and the
Theory of Investment,” American Economic Review 48(3): 261-97.
M o w ery , D., and La n g lo is , R. (1996), “ Spinning off and Spinning on (?): The Federal
Government Role in the Development o f the US Computer Software Industry,” Research
Policy!5(6): 947-66.
N elson , R. (1991), “ Why do Firms Differ and how does it Matter?” Strategic Management
Journal 12: 61-74,
(2993)> National Innovation Systems: A Comparative Analysis, Oxford: Oxford Univer­
sity Press.
О D risco ll , G., and Rizzo, M. (1985), The Economics of Time and Ignorance, Oxford and
New York: Basil Blackwell.
 FIN AN CE AND INNOVATION 265

О S u l l iv a n , M. (2004a), The Financing Role o f the US Stock Market in the 20th Century,”
Working Paper, INSEAD.
------ (2004b), H istorical Patterns o f C orporate Finance at General Electric and Westing-
house Electric” Working Paper, INSEAD.
* P erez , C. (2002), Technological Revolutions and Financial Capital: The Dynamics of Bubbles
and Golden Ages>Cheltenham, UK and Northampton, Mass.: Edward Elgar.
R a ja n , R., and Z in g a l e s , L. (1995), “ W hat do we know about Capital Structure: Som e
Evidence from International Data,” Journal of Finance 50(5): 1421-60.
* -----------(1998b Financial Dependence and Grow th,” American Economic Review 88(3):
559-86.
R o s e g r a n t , S., and L a m p e , D. (1992), Route 128: Lessons from Bostons High-Tech Commu­
nity New York: Basic Books.
R osenberg , N. (1994b Exploring the Black Box: Technology Economics, and History Cam­
bridge and New York: Cambridge University Press.
Sahlm an , W. (1990b “ The Structure and Governance o f Venture-Capital Organizations,”
Journal of Financial Economics 27(2): 473-521.
* ----and S tev en so n , H. (1985), “ Capital Market M yopia” Journal of Business Venturing
1(1): 7“ 3°*
Sax en ia n , A. (1994 ) уRegional Advantage: Culture and Competition in Silicon Valley and Route
128, Cambridge, Mass, and London: Harvard University Press.
Sch u m peter , J. (1939), Business Cycles, vol. 1, New York: McGraw Hill.
------(1942), Capitalism, Socialism, and Democracyy New York and London: Harper 8c
Brothers.
------(1949b “ Economic Theory and Entrepreneurial History,” in Research Center in Entre­
preneurial History, Harvard University, Change and the Entrepreneur, Cambridge, Mass.:
Harvard University Press.
------(1954 b History of Economic Analysis, Oxford: Oxford University Press.
—-— (1964), Business Cycles: A Theoretical, Historical and Statistical Analysis of the Capitalist
Process, abridged, with an introd., by Rendigs Eels, New York: McGraw-Hill.
------(1975), Capitalism, Socialism and Democracy, New York: Harper Torchbooks.
------(1996), The Theory of Economic Development>New Brunswick: Transaction Publishers.
Sh ackle , G. L. S. (1992), Epistemics and Economics: A Critique of Economic Doctrines, New
Brunswick, Transaction Publishers.
T eece , D. (1986), “ Profiting from Technological Innovation,” Research Policy 15(6): 285-305.
------P isano , G „ and S h u e n , A. (1997), “ Dynamic Capabilities and Strategic Management,”
Strategic Management Journal 18(7): 524-6.
W in t er , S. (1984), “ Schumpeterian Competition in Alternative Technological Regimes,”
Journal of Economic Behavior and Organization 5: 287-320.
W rig h t , G. (1997), “ Towards a More Historical Approach to Technological Change” The
Economic Journal 107(444): 1560-6.
* Z in g a les , L. (2000), “ In Search o f New Foundations,” Journal of Finance 55(4): 1623-53*
(2003), “ Commentary,” Federal Reserve Bank o f St. Louis, July-August, 47-52.
 C H A P T E R 10

INNOVATION AND
INTELLECTUAL
DDA Dl i D T V D Т Г І І Т Г
F KUFBKl I KlvjrllS

OVE G R A N S T R A N D

l o. i I n t r o d u c t i o n 1

T h e use o f property-like rights to induce innovations o f v ario u s kinds is perhaps the

oldest institutional arrangem ent that is particular to in n o vatio n as a social phenom­
enon. It is now custom ary to refer to these rights as intellectual p ro p erty rights (IPRs),
com prising old types o f rights such as patents fo r inventions (judged as sufficiently
novel* non-obvious and useful)* trade secrets* copyrights, tradem arks, and design
rights, together with newer ones such as breeding rights an d database rights.2 The
various IPRs usually have long legal and econom ic histories, often w ith concomitant
controversies. Nonetheless, despite their lo n g history, u n til recently IPRs did not
occupy a central place in debates over econom ic policy, national competitiveness, or
social welfare. In the last quarter o f the twentieth century, however, a new era—
dubbed the pro-patent or pro-IP era— em erged, first in the U S and then diffused
globally This change was em bedded in a deeper, m ore b road -b ased and m uch slower
flow o f events towards a m ore inform ation- (know ledge-) intensive and innovation-
based economy. (This type o f econom y has in recent years been dubbed the new
econ om y” som ewhat m isleadingly, as if the entire econ o m y has suddenly changed
 INNOVATION AND IN TELLECTU AL PROPERTY RIGHTS 267

into som ething new, replacing the old.) These changes provided policy makers in both
developed and developing countries with new challenges.

10. 2 H i s t o r y of t h e IPR S y s t e m

The b rie f historical account below w ill focus p rim arily on patents, being in general
the m ost im portan t and representative IP R , and will be divided into eras, su m m ar­
ized in Table 10 л .

10.2.1 The Non-Patent and Pre-Patent Era

Ancient cultures, as in Babylonia, Egypt, Greece, and the R om an E m p ire are not
know n to have had any patent-like institutions for technical inventions, but there are
clear indications o f other form s o f IP in these cultures. It was not until late m edieval
times that patent-like institutions started to appear, m ostly in the fo rm o f privileges
granted b y rulers to special individ u als or professions.

10.2.2 The National Patent Era

Concepts o f IP becam e m ore elaborate and closely linked to political institutions as
trade and tech n ology developed in the M id d le Ages. In 1474 Venice prom ulgated the
first form al patent code. Inventions show n (at least b y a m odel) to be w orkable and
useful received ten years o f protection fro m im itation, subject to certain co m p u lso ry
licensing p ro vision s. T h e 1474 patent code constituted a p olicy for Venice to attract
engineers fro m the outside and stim ulate orderly technical progress. These laws
signified the em ergence o f a n ew era, w h ich w e refer to as the “ national patent era,”
since patent system s typical o f this period were national (or local) phenom ena
pertaining o n ly to single city-states or nations.
The gran tin g o f patent-like privileges b y governm ents or rulers was not confined
to Venice an d the practice spread w ithin Europe. A s nation states w ith m ore
absolutist govern m en ts em erged, controversies also em erged between governm ents
and rulers regardin g the conditions for gran tin g patents and m o n o p o ly privileges.3
The practice o f gran tin g patents also spread in E ngland and France d u rin g the
sixteenth cen tu ry as part o f national m ercantilist policies. T h u s patents becam e
268 OVE G R A N S T R A N D

Table 10.1 Eras in the history o f patents and IP

Characteristics

1. Non-patent era Emergence of science separated from technology

(Ancient cultures: Egypt, Emergence of cultural and industrial arts
Greece, etc.) Secrecy and symbols emerging as recognized №
No patent-like rights or institutions for technical
inventions

2. Pre-patent era Emergence of universities

(Middle Ages to Renaissance)
Secrecy, copyright and symbols (artisan/trade mark$| ■
>
>
names) as dominant IP, also collectively organized
Emerging schemes to grant privileges and remunerate
disclosure
Extensions of mining laws to inventions

3. National patent era Breakthrough of natural sciences

(Late 15th—late 18th cent)
Local codifications of laws for patents (Venice 1474,
England 1623, etc.), copyrights (Venice 1544, England
1709, etc.), etc
Regulation of privileges
Conscious stimulation of technical progress at national
level, linked to economic policies (e.g. mercantilistic)

4. Multinational patent era Emergence of modern nation states

(Late 18th-late 19th cent)
5, International patent era
(Late 19th-late 20th cent)
6. The pro-patent/pro-IP era
(Late 20th cent-?)
Industrialization
Continued international diffusion of the patent system
Local anti-patent movements
Emerging international patent relations (e.g. disputes)
Emerging industrial and military REtD
International coordination of the patent system (Paris
Convention 1883, WIPO, PCT, EPO, etc.)
Separate IP regimes in socialist countries and LDCs
Intellectual capital surpasses physical capital for many
entities
Intensified international competition
Global activism for IP from industrialized countries,
especially from the US (leading to TRIPS and the WO)
Almost worldwide adoption of the patent system
Increased international patenting

Note:

Discerning eras, epochs or stages in a historical stream of events may be a useful sorting device but it always
involves some arbitrariness, even if good criteria are used. (Here the degrees of codification and geographical
diffusion of the patent system are used as primary criteria for distinguishing different eras.) Also, beneath the
events that surface in an era is often an undercurrent of events that lead up to a later era.
 IN N O V AT IO N AND IN T E L L E C T U A L PROPERTY RIGHTS 269

linked to trade policies, a link that has been im portant as well as controversial
ever since.
An im portant event in the early diffusion o f the patent system was the passage in
1623 o f the Statute o f M o n o po lies b y the English Parliam ent, which gave a clear
recognition o f the underlying ideas and specific form o f a patent system .4 This later
came to serve as a m odel, fo r exam ple, for British colonies in N orth A m erica, which
started to adopt sim ilar patent laws in the seventeenth century. An interesting
feature o f the statute w as that although the patent granted m o n o p o ly privileges to
the true and first inventor, the invention had to be new in England. This provision
was intended to stim ulate dom estic technical progress (e.g. by attracting foreign
engineers and entrepreneurs to England) and reflected concern b y England's p o lit­
ical leadership that the nation had fallen back in som e technical areas and needed to
catch up. T h e statute established a 14-year lifetim e for a patent (twice the tim e
needed for a m aster to train a generation o f apprentices). A third interesting feature
o f the statute was its explicit shift o f the granting authority from a royal ruler o r
sovereign to a govern m en t o r its bureaucracy. The governm ent was considered the
source o f patent rights, in contrast to the view s that patent rights derived from
sovereigns o r were n atural rights o f the individual. The latter view underlined the
French patent law at the tim e o f the French Revolution in 1791 and lived on in
nineteenth-century France.
A nother im p o rtan t event was the U S enactm ent o f a federal patent law in 1790.
The im portance attached to patents and in divid u al IP R s in the new ly created U SA is
clear from the fact that the A m erican C on stitu tion stated that Congress had the
power “ to p ro m o te the progress o f science and useful arts, by securing for lim ited
times to authors and inventors the exclusive right to their respective w ritings and
discoveries”.
Thom as Jefferson played a key role in the early days o f the U S patent system. As
Secretary o f State he was responsible for adm inistering the patent laws, and as a head
o f a new ly created “ Patent B o a rd ”, he personally exam ined patent applications. He
was noted fo r his o p p o sitio n to m on opolies but believed in the value o f lim ited
m onopolies fo r au thors and inventors. The new U S patent system had a slow start,
just as it had had in Venice three centuries earlier and w ou ld have in Jap an a century
later. The A ct o f 1793 m ade substantial changes, om itting the requirem ent that a
patentable invention had to be “ sufficiently useful and im p o rtan t”. The exam ination
o f applications fo r n o velty and usefulness w as replaced by m ere registration, m aking
the issuing o f patents m ore o r less a clerical matter, and the Patent B oard w as
abolished. T h e Patent A ct o f 1836 in essence reestablished the exam ination system
that had been in place until 1793 and created an executive Patent Office as a separate
bureau w ith in the D ep artm en t o f State. T h e U S Patent O ffice created by the 1836 Act
was adm inistered b y a C o m m issio n er o f Patents, appointed by the President upon
approval b y the Senate. T h e present U S system for review ing and adm inistering the
patents is largely based on the principles set forth in the Act o f 1836.
270 OVE G R A N S T R A N D

Box 10.1 International IP conventions

■- — '— ' 1 ....................... .................................. 1 , " " 1 1 . ^ 7 ' ■ " >

H e ^ a rfa lirte m a tiA ti С о ! Ш і ш : й г the Protection

(the Convention” ) covering patents, trademarics, and designs and the accom­
panying Berne Convention For the Protection o f Literary and Artistic Works of 1886
covering copyright and some related rights were the results of complex interplay
between different interests during preceding decades. Switzerland without a patent
system at the time but with an increasingly pro-patent watch industry was one of the
countries active in pulling through the Paris Convention (plus housing the Berne
Convention) and was entrusted with secretariats to administer and supervise these
conventions. The secretariats shortly thereafter merged into a bureau (“ В Т О Р ). In
1967 it was reorganized as the World Intellectual Property Organization ( W f 6 |
Largely through the diplomatic efforts o f the former US representative to the M ir
Union and the Berne Union (respectively consisting o f the signatories to the Paris
Convention and the Berne Convention), WIPO later became a United Nations agency
in 1974.
The Paris Convention is based on two major principles: (a) Foreigners and foreign
patent applications should receive the same treatment in a member state as domestic
applicants and applications (non-discrimination); (b) a priority claim established in
one member state should be recognized by all others, i.e. once an application for a
patent is filed in a member state, the applicant can within twelve months file a patent
application for the same invention in any other member state, which must regard the
latter applications as being filed on the date o f the original first application.

10.2.3 The Multinational Patent Era

The period fro m the late eighteenth to the late nineteenth cen tu ry w as characterized
by the d iffu sion o f the patent system th ro u gh o u t the in d u strial and industrializing
econom ies, though at an uneven rate and n o t w ith o u t setbacks. A n anti-patent
m ovem ent em erged in G erm an y and so m ew h at later in H o llan d , w here patent laws
were repealed in 1869, and Sw itzerland rejected several patent law proposals. Even
England considered a p roposal to w eaken sign ifican tly h er patent laws, and France
had already weakened patent protection at the tim e o f the French Revolution.
The anti-patent m ovem ent was a consequence o f free trad e an d anti-monopoly
m ovem ents which considered patents to be associated w ith m ercan tilist policies and
m o n o p o ly privileges. H owever, interest g ro u p s in em erg in g indu stries and in some
strong-patent nations created pro -p aten t lo b b yin g g ro u p s that gradually gained
influence. Finally, the w orldw ide depression in the 1870s revived protectionism
and the anti-patent era by and large ended in the 1870s.
I he case o f Switzerland provides an interesting illu stratio n o f the forces affecting
the international diffu sion of the patent system . A fter p o p u la r referendum s in 1866
and 1882 had rejected proposals to in trodu ce patent law s, Switzerland finally
 I N N O V A T I O N AND INT ELLEC T UAL PROPERTY RIGHTS

approved such laws in a referendum in 1887, m ainly because its im portant watch
industry was under pressure from foreign im itations. H owever, the 1887 law lim ited
patent protection w ithin Swiss borders to m echanical inventions, since firm s in the
em erging Swiss chem ical indu stry w anted to im itate and catch up with the m ore
advanced G erm an chem ical industry. A fter G erm any threatened Swiss chem ical
firm s with retaliatory tariffs, Sw itzerland extended its patent coverage to include
chem ical process (but not product) inventions in 1907 (see Penrose 1951 and
Kaufer 1989).

10.2.4 The International Patent Era

Eventually the patent system was w idely adopted, concom itant w ith the grow th o f
international trade and com petition in industrial goods. N ation states adopted
various policies for p ro m o tin g their industries, policies that often discrim inated
against foreign in d ivid u als or firm s, in turn creating a need fo r international
cooperation in patent m atters. The Paris C onven tion o f 1883 w as the first m ilestone
in this respect, follow ed b y several other treaties and agreem ents, such as the Berne
C onvention fo r copyrigh ts in 1886, covering a w ide range o f IP R s (see B o x 10 .1).
The em ergence o f indu strial R8cD in the twentieth century transform ed the
m odes and settings for innovative w ork. The in divid u al inventor, who w as the
original target for patent laws, gradually becam e less im portant. Inventions increas­
ingly required large resources, and industrial firm s becam e the prim e m overs o f
technology in b oth the East and the West. E con om ic and industrial differences
between variou s categories o f nations increased and becam e alarm ingly large,
creating tension am o n g institutions, including national IP regim es in developed
and developing nations. Science and technology progressed and accum ulated tre­
m endously at an increasing pace. N onetheless, the IP system and its essential ideas
survived and contin ued to spread internationally, not least after the dow nfall o f the
Soviet U n io n and other planned econom ies.
International h arm o n ization o f the w orld's patent system s received a new im petus
after the end o f W orld W ar II, as part o f a broader set o f efforts to establish or
strengthen in tern ation al organizations. A convention establishing the W orld In tel­
lectual P ro p erty O rganization— W IP O — was prom ulgated in 1967 by fifty-one
governm ents, m o stly fro m developed nations. W IP O join ed the U N system in
1974 a^ d thereby cam e under m u ch stronger influence fro m developing countries.
A lthough W IP O w as established to adm inister and supervise variou s international
IP treaties such as the Paris C on ven tion , the organization also becam e involved in
teaching, arb itration , and consultancy, and the processing o f patent applications
w ithin the fram ew o rk o f the Patent C o op eratio n Treaty (P C T ), signed in 1970
but not effective until 1978. T his treaty w as an im portant step in the process o f
27 2 OVE G R A N S T R A N D

international h arm onization, since it established an in tern atio n al clearing house

enabling a patent application to take effect in som e o r all o f the P C X m em ber states
(103 in 1999) at the applicant's choosing.
In Europe the E uropean Patent C o n ven tio n (E P C ), w h ich w as signed in 1973 and
becam e effective in 1978, began a process for the ad ju stm en t o f n ation al patent laws
o f sign atory nations (thirteen in 1986) to a E u rop ean stan d ard . T h e European Patent
O ffice (EPO ) was established in 1977 in M u n ich to process paten t applications for the
protection o f an invention in som e o r all sign atory n ation s. H ow ever, a patent issued
b y EPO is on ly a bundle o f national patent rights that are enforceable according to
the local law and court system in each national ju risd ic tio n . A E u rop ean Commu­
nity Patent C on ven tion was signed in 1975 to establish a u n ified E u rop ean patent that
w ould be valid in all m em ber states, but this go al has n o t been achieved as o f 2003.
A key issue in the full h arm on ization o f the E u ro p ean IP R system Is the design of a
court system and court procedures for en forcem en t, p o ten tially including the
creation o f a single European C o u rt o f A p peal fo r IP R d isp u tes sim ilar to the US
C o u rt o f A ppeals for the Federal C ircu it (C A F C ), discu ssed below .
The case o f Jap an provides an interesting illu stration o f the creation o f a patent
system fo r the pu rpose o f catching up w ith ad van ced n ation s. The visit by US
C o m m o d o re Perry to Jap an in 1852 dem onstrated to Jap an ese leaders the power of
m odern m ilitary innovations and forced Jap an to reopen the c o u n try to foreigners.
The M eiji Restoration and its broad p ro g ram o f in d u strial m odernization and
“ catch -up” led to Ja p a n s first patent law in 1871. In the fo llo w in g decades, new
laws were enacted for variou s IP R s (patents, trad em ark s, u tility models, and
designs), each o f them m odeled on vario u s E u ro p ean an d U S laws. A Japanese
Patent O ffice was established in 1885, w ith K. Takahashi, w h o subsequently served
as Japan's Prim e M inister, as its first D irector G eneral. T h e Jap an ese patent system
evolved over the years into an im po rtan t vehicle fo r catch in g up and promoting
national interests. In the beginning, foreigners w ere b arred altogether from
obtaining patent rights but becam e eligible w hen Jap an in 1899 becam e a member
o f the Paris C onvention.
The postw ar IP system in Jap an was but one co m p o n en t o f a b road er complex of
policies for trade, industry, and tech n ology that fo cu sed o n reconstruction and
catch-up w ith the West, especially the U nited States. Law s w ere passed in 1950 for
regulating foreign investm ent, exchange, and trade, in au g u ratin g a period o f sub­
stantial tech nology im ports from the U S and E u rop e. Jap an ese govern m en t agencies
and firm s collected and analyzed technical in fo rm a tio n , in clu d in g information
disclosed in dom estic and foreign patent d ocu m en ts, to evaluate technological
developm ents abroad and within Japan. The Jap an ese req u irem en t for publication
o f patent applications w ithin eighteen m o n th s o f the filin g o f an application (a
p olicy sim ilar to that o f m any E uropean patent system s that w as adopted by the
United States only in 1999) supported d om estic as w ell as in tern atio n al diffusion of
technical Inform ation.
 IN N O V A T IO N AND I N T E L L E C T U A L PROPERTY RIGHTS 273

The Japanese patent system lim ited both the num ber and scope o f patent claim s.
M any Japanese firm s acquired large portfolios o f relatively narrow dom estic patents
and participated in dense patent netw orks (also in their foreign patenting* see
Granstrand 1999)* If* disputes were avoided and cross-licensing and diffusion o f
technical in fo rm atio n w ere prom oted by special features o f Japanese patent laws and
practices (see O rdover 1991). Use o f patents (both foreign and dom estic) by Japanese
firm s for technological catch -u p” purposes was facilitated by the often lax enforce­
ment by W estern firm s o f their IPRs* as w ell as the lim ited attention paid to dynam ic
com petition and IP m atters b y W estern nations until the 1980s.
Nonetheless* Jap an su p po rted international harm onization efforts. In 1978 Japan
acceded to the PCT. The Japanese Patent Office ( JP O )* together w ith a sm all num ber
(around ten) o f other patent offices (in other P C T m em ber states)* was entrusted
with au th o rity to p erfo rm international searches for prior art to assess whether the
novelty criterion for patentability was met. Japan subsequently becam e active in
trilateral patent office cooperation am ong the EPO* JPO , and U SPT O (U nited States
Patent and T radem ark Office)* another vehicle for international coordination and
harm onization am o n g the industrial nations.
By 19 9 9 ,155 nations had adopted the Paris Convention* which in 1883 had been
signed b y ten. In 1994* another large step tow ards international h arm onization was
taken w ith the signing o f the U S-in spired T R IP S agreem ent (see B o x 10.2)* co n ­
sidered b y m ost experts to be the m ost im portant international IP R agreem ent since
the Paris C on ven tion . The T R IP S agreem ent has been criticized for favoring de­
veloped nations and im p ed in g econom ic developm ent in developing ones. The least
developed nations in particular* m an y o f w hich lack the capabilities to enter a
virtuous catch-u p developm ent circle b y themselves* m ay be hindered b y the
T R IPS agreem ent, although m ost such nations have a considerable period o f tim e
to adhere to all o f the p rovision s o f T R IP S (see e.g. the collection o f articles in
M ansfield and M an sfield 2000).
D espite lo n g-stan d in g efforts to coordinate and harm onize the national patent
laws, m an y im p o rtan t differences remain* and a global patent system , w ith inter­
national o r global patents* seem s far away.

10.2.5 The Pro-Patent Era

Towards the end o f the twentieth century a new era— the pro-patent era— emerged*
characterized b y stron ger enforcem ent o f a broader array o f IP R -holder rights and by
additional efforts at in tern ation al coordination and harm onization. The dow nfall o f
the Soviet em p ire and U S d iplom atic pressure contributed to a higher rate o f
convergence o f IP regim es in the world* exem plified by the T R IP S agreem ent and
the creation o f the W T O (see below ).
2 74 OVE G R A N S T R A N D

Four developments in the United States led to the “pro-patent era” (see Table 10.2
for an overview). The first concerned the creation in 1982 of the Court of Appeals for
the Federal Circuit (CAFC) to hear patent appeals in lieu of the other circuit courts
of appeal.5This type of specialized court had been discussed for a long time in patent
circles.6As the complexity of patent disputes grew, the pressures within pro-patent
circles in law and industry for a specialized court of appeals mounted and finally
resulted in the creation of the CAFC. As many of its proponents had hoped, the
CAFC began to act in a pro-patent manner, in contrast to what US courts had done
previously. The validity of patents was upheld far more often (as if they were “bom
valid” ), and patent damages were increased. The effect of the C A F C s creation and its
decisions was to increase the economic value of patentholder rights.
A second factor behind the em ergence o f the p ro -p ate n t era w as linked to a change
o f attitude w ithin the A n titrust D ivisio n o f the U S D ep artm en t o f Justice in the early
1980s under Assistant A ttorn ey G eneral W illiam Baxter. S in ce the late 1930s, the
A ntitrust D ivisio n had been hostile to IP legislation an d IP licensin g, interpreting
patents as m on opolies h arm in g com petition. B axter w as in stru m en tal in shifting the
Justice D epartm ent s enforcem ent p o licy to em ph asize the role o f patents in pro-
m otin g in n ovation , em phasizing the d yn am ic benefits rath er th an the static costs,
T h is change in attitude could be traced b ack to ideas an d perspectives emerging in
the 1960s am ong econom ists, especially w ith in the em erg in g field o f law and
econom ics.7 The shift in antitrust p o licy in the early 1980s in the U SA is a good
(albeit rare) exam ple o f h ow changes in sch o larly th in k in g h ave had a direct impact
on policies.
The third stream of events contributing to the rise of the pro-patent era came from
large U S corp orations that pressed for stron ger IP p ro te ctio n and enforcement
against infringers and counterfeiters d om estically an d ab ro ad . U S industry also
pressed for a “ trade-based ap p ro ach ” to im p ro ve IP p ro tectio n b y including IP
m atters in U S trade negotiations and in the G A T T fra m e w o rk o f international trade
negotiations, resulting in a n u m ber o f “ trade related aspects o f IP R s” (TRIPS)
subjected to negotiations (see B o x 10.2). These initiatives, w h ich w ere spearheaded
by U S pharm aceuticals, entertainm ent, and electronics firm s, w ere part o f a larger
upsurge in political concern over the com petitiveness o f U S In d u stry and a growing
b elief that technology was a key asset that had to be protected . Individual US
corporations such as Texas Instrum ents and M o to ro la b ecam e aggressive litigators
against both dom estic and foreign, especially Jap an ese, in frin g ers in the mid-1980s,
M ost o f the largest awards o f dam ages in patent in frin g em en t cases, however,
occurred in litigation am ong U S firm s, A lan d m ark case in this connection was
Polaroid Corp. v. Eastman Kodak , w hich in 1991 resulted in a dam ages award to
Polaroid o f alm ost $U S9oo m illio n .8 Cases like these an d the fin an cial success of the
litigation strategy o f Texas In strum ents w ere w id ely p u b licized and drew the atten­
tion o f top corporate m anagem ent to IP m atters an d the e co n o m ic value o f strong
patent po rtfolio s and w ell-conceived IP strategies ^
 IN N O V A T IO N AND IN T E L L E C T U A L PROPERTY RIGHTS 275

Box 10.2 Trade-Related aspects of Intellectual Property Rights—TRIPS

The idea of linking IP policies to trade policies can be traced far back in history (for
instance, in the national patent era IPRs were often used in a mercantilistic way). The
acronym TRIPS refers to a US initiative of the 1 9 8 0 s that sought to link more stringent,
internationally harmonized IP policies to international trade policy. The US strategy
was to move IPR issues from the auspices of WIPO (seen by US as too weak and
narrowly focused) into the GATT Uruguay Round of multilateral trade negotiations in
which the US had more influence. The outcome was a success for the US and its allies,
but developing nations were frustrated. When the World Trade Organization (WTO)
came into being in 1 9 9 5 as a successor to GATT, the TRIPS agreement was one of its
founding components. The agreement consisted of seven parts and seventy-three
articles covering all aspects of IPRs, their enforcement and institutional arrangements.
It provided general obligations regarding national nondiscrimination and transpar­
ency, it stipulated substantive minimum standards in almost all IPR areas (patents,
copyrights, trademarks etc.) plus standards for effective enforcement of IPRs (also
involving dispute settlement mechanisms at the WTO). It further set up a TRIPS
Council for monitoring the operations of the agreement. Finally, transition periods
were stipulated, giving one year for developed nations from entry into the WTO to
comply with all TRIPS requirements and eleven years for least-developed nations (i.e.
until 1 January 2 0 0 6 ) with an option to request extensions. The TRIPS agreement
implied particularly significant changes in the coverage of patents (forcing many
nations to extend patent protection to chemical, pharmaceutical, and biotechnological
inventions), requirements for protection of plant varieties, protection of computer
software and effective measures to protect trademarks and trade secrets (see e.g.
Maskus 2 0 0 0 for details).
The TRIPS agreement of 1 9 9 4 has been characterized as the most significant
international harmonization effort of IPRs in history, certainly on par with the Paris
Patent Convention of 1 8 8 3 . It also appears to become the most controversial one,
perhaps creating an anti-IP movement of much larger international proportions than
the anti-patent movement in Europe in the 1 8 5 0 s to 1 8 7 0 s. Particularly controversial
issues concern developing nations’ access to new technologies, especially drugs, and the
effects of stronger IPRs on the efforts of these nations to catch-up economically (see
e.g. Scherer 2 0 0 4 , Scherer and Watal 2002, and the chapters by Anawalt, Barton, and
Verspagen in Granstrand 2 0 0 3 ).

A fourth force beh ind the em ergence o f the pro-patent era was the U S G o vern ­
ment, especially the R eagan adm inistration. This “ political stream was also related
to the grow ing dom estic concern o f the 1980s for U S industrial com petitiveness,
which included the w idespread perception that a num ber o f Asian econom ies were
“ free-riding” on U S tech n o lo gy as they m ade significant inroads into U S m arkets. In
addition, U S in d u strially funded R & D spending was grow ing slow ly during the early
1980s w ith little o r no increase in patenting. M eanw hile foreign corporations.
2j 6 OYE G R A N ST R A N D

especially Japanese firm s, increased their p aten tin g in the U S A .10 O ne com ponent of
a broader po licy response to the perceived decline in U S com petitiveness was
legislative action to strengthen IP R s an d o th er incentives to invest in R & D (such
as R & D tax credits and favorable co n d itio n s fo r the creatio n o f R & D consortia), and
to encourage patenting o f the results o f fed erally fu n d ed R & D to facilitate interin-
stitutional R & D collaboration and tech n o lo gy transfer. T h e B a y h -D o le Act o f 1980
sim plified the procedures under w hich U S u n iversities co u ld patent and license the
results o f federally fu nded R & D (see Ch. 8 b y M o w e ry an d S am p at in this volume).

Table 10.2 Chronological overview of major events in US post-war IPR

development (through 2000)

Year Event

1949 Patents so frequently declared invalid when litigated that Supreme Court Justice Jackson
remarks, “the only patent that is valid is one which this Court has not been able to get its
hands on". (Jungerson v. Ostby Et Barton Co.)

1952 The present (as of 2003) US Patent Law is passed. Revisions have occurred continually.

1976 US Copyright Act enacted.

1979 US Senate and President Carter desire to strengthen domestic patent enforcement

1980 US Supreme Court declares man-made microorganisms to be patentable and states in a

dictum that “anything under the sun that is made by man" can be patented. Bayh-Dole
Act enacted, facilitating for universities to patent inventions from federally funded
research.

1981 The US Justice Department revises its antitrust enforcement activity to make it easier for
patents not to violate antitrust statutes. US Supreme Court decision in the Diehrcase
leads through its USPTO interpretation to patentability of certain computer software.

1982 CAFC is established. In quick order, the court changes the validity of litigated patents
from 30°/o to 89%, thus initiating an era in which patents are of much greater interest to
industry.

1983 Patent Commissioners' trilateral conference started.

1985 WIPO Harmonization conference. USITC litigations increased. The Young Report
delivered to President Reagan by the Commission on Industrial Competitiveness (headed
by Hewlett-Packard's John Young),

1986 Tl semiconductor patent litigation initiated at USITC. GATT TRIPS negotiations started,

1988 US Trade Act (Special 301). US Tariff Act 337 amended.

1989 The Structural Impediments Initiative (SII) talks initiated between the USA and Japan
remove structural impediments to trade between the two nations, and include
intellectual property protection. Japan on Watch List of Special 301,

1992 US Patent Law reform report Honeywell won patent litigation against Minolta.
 IN N O V A T IO N AND I N T E L L E C T U A L PROPERTY RIGHT S 2/7

1993 GATT TRIPS n e g o tia tio n s co m p le te d .

1994 W o rld s in d u s trialize d n a tio n s agree to h arm o n ize aspects o f th e ir in te lle c tu a l p roperty
protection under the auspices of GATT, known as the TRIPS agreement
U S -J a p a n Patent C om m issioners’ U n d erstan d in g signed. After years of favorable court
decisions, all software is now clearly patentable.

1995 GATT-related TRIPS agreement causes USA (and other nations) to amend its patent laws
to expand the patent term to 20 years from filing date (from previous 17 years from
issuing date, thus giving mixed effects depending upon the application processing time
at the USPTO), allow inventive activity abroad to be considered by the patent office, and
permit the filing of provisional patent applications.

1998 The CAFC declares inventions of so-called business methods to be patentable (which
include e.g. financial inventions, teaching methods, and e-commercial methods) in Sta te
S tre e t Bank and Trust v. Signature Financial Group by stating that "since the 1952 Patent
A ct, business m eth o d s have been, and should have been, subject to th e sam e legal
re q u ire m e n ts for p a te n ta b ility as ap p lied to a n y o th e r process or m e th o d ". The Digital
M ille n n iu m C o p yrig h t A c t enacted.

The CAFC and the change in antitrust policies paved the way for effective
domestic enforcement of existing US IP laws. The trade-based approach to IP
legislation* however* focused primarily on international standards and enforcement
of intellectual property protection. This effort was largely successful (from the US
point of view)* in part because the US Congress created leverage for US trade
negotiators through a number of changes in US trade laws.11 However* the pro­
patent era* set in motion by the actions of US corporations and policy makers*
gained ground internationally for other reasons as well. Technology-based MNCs*
not only in the USA but also in Europe and especially in Japan* shared an interest in
stronger international protection for intellectual property.
There is an ongoing debate, fueled by the bursting of the “ IT b u b b le ” as to
whether in fact a new type of economy has emerged and what characterizes such a
“New Economy”. Although much of the “ New Economy” rhetoric is now dis­
credited, many scholars believe a new type of economy has emerged* albeit gradually*
in which intellectual capital has surpassed physical capital in importance. “ Intellec­
tual capitalism”, then, refers to a capitalist economic system with a dominance of
intellectual capital (see Granstrand 1999). What role did the IP system and the pro-
patent era play in the emergence of “ Intellectual capitalism” ? A definitive answer to
this question is difficult at this stage, but a few observations may still be in order.
ICTs are generally recognized as a key technological contributor to the em ergence
o f “ intellectual capitalism ” as w ell as the “ N ew Econom y. It is natural to ask,
therefore, h o w im p o rtan t the IP system was for the em ergence o f I d s ; let us
consider som e w ell-k n ow n cases. The transistor w as patented at Bell Labs but
278 OVE G R A N STR A N D

licensed liberally (in p srt because o f antitrust litig atio n <md pressu re from the US
Justice D epartm ent). The subsequent em ergence o f the sem ico n d u cto r industry was
significantly spurred by pu blic procu rem ent and a la x IP regim e (M o w e ry 1996). The
sam e could be said abou t the em ergence o f In ternet u n d er the D efense Advanced
Research Projects A gency (D A R PA ). T h e softw are in d u stry also em erged under alax
IP regim e (Sam uelson 1993)- The telecom in d u stry w as larg ely operated by national
m o n opolies until the 1980s and 1990®» and IP R s played little role in the rapid advance
o f tech n ology in this industry. M obile teleph o n y also em erged u n til the late 1980s
under a lax IP regim e (G ran stran d 1999). T h e co n clu sio n seem s to be that the IP
system has not been o f m ajo r im portan ce for the em ergen ce o f IC T s (at least not in
the early stages). In fact it m ay even be argued that lax IP regim es w ere instrumental
fo r the em ergence o f several IC T industries.
The strengthening o f the IP regim e m a y have rein forced so m e features o f intellec­
tual capitalism , but it appears that the p ro -p aten t era w as as m u ch a consequence of
intellectual capitalism as a cause o f it, and that it w as n o t a n ecessary condition for
the em ergence o f the industries and technologies that fostered it.

10.3 R o le of IPR s in In n o v a t io n Systems

10.3.1 Perspectives on IPRs

The IP system has over the years spawned a series o f legal and economic controver­
sies. Among the legal controversies is the nature o f IPRs: are they rights in the first
place? Couldn't a liability approach do better? Do they have to be exclusive and/or
temporary rights? And if a right, what kind of right? Is it an individual natural (or
moral) right or a right conferred to the individual by society, justified on the grounds
that its consequences are beneficial to the society? These types of questions are
primarily addressed by legal scholars in jurisprudence and, although important, will
not be discussed further here.
As for economic issues, there has been a continuing discussion (with varying
intensity) over the centuries about the pros and cons of the patent system. One key
question is whether the system can correct for (or lead to) over- or underinvestment
in R&D and innovation (from a societal point o f view). Another issue is whether the
system distorts, redirects, or blocks technological progress. Still another (but re­
lated) topic concerns how the patent system affects static and dynamic efficiency
through its impact on competition and trade. For a classic review o f these issues, see
Machlup (1958); for a more recent contribution, see Mazzoleni and Nelson (1998)*
 IN N O V A T IO N AND IN T E L L E C T U A L PROPERTY RIGHTS 2 /9

Surprisingly, fo r m uch o f the twentieth century, econom ists devoted little atten­
tion to the patent system , and even less to other IPRs. From the 1960s onw ards,
however, the literature on the econom ics o f IPRs, and patents in particular, has
grown significantly, A sem inal w ork in this category is A rrow (1962) who argued
that, from society s poin t o f view, private firm s will underinvest in R & D because o f
their inability to app ropriate sufficient returns o f their R & D investm ents.12
Follow ing this view , patent protection can be justified as one o f several alternative
means, such as contracts, prizes, subsidies, and research consortia, to deal w ith this
market failure. A lternative m eans to address m arket failures in the innovation
process are also discussed by, am ong others, W right (1983) and D avid (1993)
em phasizing conditions and factors (such as uncertainty and elasticity o f research
supply) determ ining their relative advantages.
The opposite view , that capitalist econom ies m ay overinvest in R & D and in n o v­
ation, also has its adherents. A n um ber o f recent theoretical contributions in in d u s­
trial econom ics, grow th theory, and behavioral finance exam ine this possibility,
highlighting the ways in w hich com petitive races affect incentives for innovation and
can result in overinvestm ent in R & D . The intuition behind such m odels is that
patent races m ay result in substantial duplication o f R & D investm ents b y the
com peting firm s and lead to in d u stry-w id e levels o f R & D investm ent for which
the social returns m ay be less than the cost o f the overall investm ent.13 Such over­
investment is m ost likely in races in w hich rew ards are skewed heavily to the early
finishers. In this case agents face strong incentives to accelerate both the start and the
com pletion o f R & D projects (at least until rents are dissipated).
Thus ap p ro p riab ility problem s m ay lead to underinvestm ent in innovation, e.g.
in “ w aiting gam es”, o n the one hand, or prospects o f a quick success m ay lead to
“ patent races” that result in overinvestm ent. W hich type o f gam e is actually observed
will depend on m an y circum stances. M ost o f this literature is theoretical, and few
em pirical studies have addressed these theories and m odels.
A nother im p o rtan t issue that attracts considerable interest concerns the length o f
the time for w hich an inventor should be awarded a patent. For instance, N ordhaus
(1969) argued that increasing the length o f patent protection increases the incentives
for investm ent in process in n ovation (and hence “ dynam ic efficiency” ) but at the
expense o f “ static efficien cy” (since increased protection m eans less com petition,
higher prices, and slow er d iffu sio n ). A n “ optim al patent length”, N ordh aus pointed
out, involves a tra d e -o ff between these two effects, and w ill depend on the nature ot
com petition, the price elasticity o f dem and and the R & D elasticity o f process cost
reduction.
More recent research focuses on the optimal breadth or scope of a patent (see Jaffe
2000 for an overview) as well as optimal combinations of length and breadth
(Klemperer 1990). The scope of a patent defines the range of its industrial applica­
tions by delineating the set of technological designs that the claims in the issued
patent give p rotection to (i.e., exclusion o f im itators). In m ost indu strial-econom y
280 ove g r a n str a n d

patent system s, patent scope is initially determ in ed b y n ego tiatio n s between the
applicant and a patent exam iner; ultim ately, how ever, the sco p e fo r “ im portant”
patents (those deem ed b y patentholder or co m p etito rs as especially valuable) is
likely to be determ ined in the courts as a result o f private law suits initiated by the
patentholder o r others.
The scope o f a patent is far m ore d ifficult to param etrize th an its duration. Thus,
the “ o p tim al” scope for patents is a v ery co m p lex issue, as sh o w n b y Merges and
N elson (1990). The scope o f a patent affects the private as w ell as the social rates of
return fro m patented indu strial in n ovation s (just as the tim e d u ra tio n o f a patent
does), and these returns w ill v a ry am on g indu stries an d tech nologies, Accordingly, it
is difficult for legislators to design an overall “ o p tim a l” paten t system that fits equally
well everyw here. For exam ple, patent offices often issue b ro ad patents to applicants
pioneerin g in a new technological field characterized b y little patent-based "prior
art” and a great deal o f uncertainty. (Such patents are o ften regarded as being too
broad b y ex post observers.)
In o rd er to obtain a patent for an invention, the in ven to r has to disclose infor­
m ation. The resulting disclosure o f patent in fo rm a tio n accelerates the diffusion of
patented technical in form ation , and m ay reduce d u p licate R8cD, indu ce substitute
technologies (through “ inventing aro u n d ” an im p o rta n t p aten t), stim ulate new
ideas, direct RScD efforts to o p p o rtu n ity-rich areas o r b ottlen eck problem s, provide
a basis fo r b en ch-m arking and com petitive intelligence, and stim u late technology
exchange and co o p eratio n .14 Thus, the d isclosure req u irem en t is one o f the key
features o f the patent system and a rationale behind it (see O rd o ver 1991). It should
be kept in m ind, however, that there are several oth er ch an nels fo r dissemination
than just disclosure o f patent in fo rm atio n , and that, in general, in fo rm atio n about
new technologies leaks out fairly q u ickly (see M an sfield 1985).
In sum m ary, IP R s, particu larly patents, play several im p o rta n t roles in innovation
system s— to encourage in n ovation and investm ent in in n o va tio n , and to encourage
d issem ination (diffusion) o f in fo rm atio n ab ou t the p rin cip les an d sources o f innov­
ation throughout the econom y. H ow ever, as w e shall see, the im portan ce o f these
roles varies across sectors (industries) and cou ntries, and o ver tim e.

10.3.2 Evidence on IPRs

M ansfield (1986), in an em pirical study o f U S firm s, shed light on the impact of a
possible abolition o f the patent system on the rate o f invention and innovation, and
concluded that the effects o f abolition w ould be sm all in m ost industries. The excep­
tions were pharmaceuticals and chemicals, for which the patent system was shown to
be essential In spite o f this evidence on the relatively m odest effects o f patents on
innovation, however, the firm s in Mansfield’s study patented extensively.
 I N N O V A T IO N AND IN T E L L E C T U A L PROPERTY RIGHTS 281

The p ro pen sity to utilize the patent system (instead o f alternative m eans) is the
subject o f a large literature: see e.g. Scherer (1983) and A rundel and Kabla (1998). The
Yale study b y Levin, N elson and others (Levin et aL 1987) investigated, through a
survey o f hundreds o f R8cD m anagers in the U S in m ore than a hundred industries,
sector-specific variation s in ap p ro p riab ility conditions and the role o f patents. Their
study also concluded that innovations w ould continue to appear in the absence o f
patent protection, and that in general patents were not sufficient to appropriate or
capture all benefits from innovation (once again, a significant exception to this
general finding was pharm aceuticals, where patent protection was deem ed to be
especially valuable). The Yale study was follow ed up b y an expanded international
study (the C a rn e g ie -M ello n study), which revealed substantial nation-and sector-
specific differences in the use o f patents, secrecy, lead tim es and other m eans for
appropriation o f the returns from innovation (see e.g. C ohen et a l 2003). For
instance, the latter study indicated that lead tim e and patents were the m ost
im portant ap p ro p riation m echanism s for Japanese firm s, w hile lead tim e and
secrecy were m ost im p o rtan t for U S firm s.
A com parative study o f Japanese and Swedish corporations, accounting for m ore
than 50 per cent o f industrial R8cD in Japan and over 90 per cent in Sweden
(Granstrand 1999), also shed light on this issue. Table 10.3 com pares the results from
this study w ith the results in Levin et al. (1987). As in the C arnegie-M ellon study,
Japanese firm s stand out b y assigning greater im portance to patent protection than

Table 10.3 Means for commercializing new product technologies

(Scale: No importance = 0,1,2,3.4 — Major importance)

Means Japan1 Sweden1 us2

Taking out patents to deter imitators (or to 3.3 1.9 2.0

collect royalties)
Exercising secrecy 2,4 2.0 1.7
Creating market lead times 2.7 2.4 2,9
Creating production cost reductions 2,9 2,7 2.7

C it in g superior marketing 2.7 3.0 3,1

Й Ш іп д switching costs at user end 1.9 1.7 n.a.

Atotes: '■
1 Sample of 24 large corporations. Perceptions for 1992.
2 % reported m Levin et al. (1007). Perceptions for mid-1980s, rescaled to the scale used in Granstrand (1999).
fpiTO; .Qranstrand (1999).
282 OVE G R A N S T R A N D

firms from other countries. A 1988 study of Swiss firms, however, concluded that lead
time was the most important mechanism for appropriating the returns from innov­
ation, while patents ranked lowest (Harabi 1995)* One has to interpret these results
with care, though. First, the different mechanisms are to some extent complementary
(both patents and secrecy may be seen as means to create lead time by increasing speed
to market relative to competitors). Second, the attitudes of firms towards IPRs have
changed recently, following the emergence of the pro-patent era. Particularly
European industry in general was slow to adapt to this new environment.

10.3.3 Differences in IPRs across Sectors

A s show n in m ost i f not all studies o f the role o f IP R s, especially patents, differences
across industries o r sectors are strikingly large (see e.g. the studies b y Mansfield,
Scherer, Levin et al., and C oh en et al. cited above and also M alerb a, Ch. 14 in this
v o lu m e ).15 Table 10.4 provides an illustration o f this fo r the sam p le o f large Japanese
com panies m entioned above. Sim ilar findings have also been rep o rted for US and
U K firm s (M ansfield 1986, Taylor and Silberston 1973). Several explanations have
been set forth for these in ter-in d u stry differences, in clu d in g in d u stry and market
structure (com petitive conditions, size and d iversificatio n o f firm s, barriers to entry
m arket grow th, R & D intensity etc.), the nature o f the tech n o lo g y (technological
opp ortu n ities, codifiability, capital inten sity etc. ) and the n atu re o f IP R s (patents for
technology, copyrigh t for softw are and creative in d u stries, trad em arks in mass
consum er m arkets, etc.). But these largely static cro ss-in d u stry com parison s rarely
have incorporated considerations o f the stage o f in d u stries5 e v o lu tio n .16

Table 10.4 Sensitivity of the R&D investments of large Japanese corporations to

length of term (1992)

What would the effect be on your Chemical Electrical Mechanical Total

company’s total R&D budget (as a (n =9) (n =10) (n = 5 ) (n=24)
rough percentage), if the maximum
length of patent protection was:

(a) Increased by 3 years +8.5 + 2 .8 + 0 .3 +4.8

(b) Decreased to 10 years —21.2 -3 ,7 -0 .3 —10,7
(c) Decreased to 0 years
(i.e. patent protection ceases) -5 9 .2 -4 0 .0 —5.5 -3 8 .2

So u rce : Granstrand (1999).

 IN N O V A T IO N AND IN T E L L E C T U A L PROPERTY RIGHTS 283

As m entioned above, the role o f a strong IP regim e in em erging industries is

am biguous. There is som e evidence that several leading-edge U S industries based on
ICTs developed after W orld War II under a fairly lax IP regim e (see above). O n the
other hand, there also exist cases in w hich a strong IP regim e has fostered the
em ergence o f new leading-edge industries— pharm aceuticals and chem icals are
am ong the best-know n exam ples o f such industries.
In general, patents are m ost likely to su p po rt the grow th o f know ledge-intensive
industries in fields characterized b y low ratios o f im itation to innovation costs. Such
low ratios are likely in areas w ith large-scale R8cD projects, especially i f the R & D
results in h igh ly codified know ledge, as in chem icals (see Table 10.4), and reverse
engineering is ch eap.17 In such industries other institutional m eans to induce
innovation than patents are also co m m on ly em ployed, e.g. procurem ent contracts,
consortia, o r natural m onopolies. H owever, m any em erging industries are charac­
terized b y relatively low innovation costs and strong “ first-m over advantages'1,
m aking firm s in such sectors less sensitive to free-rider problem s and “ w aiting
gam es" and, hence, reducin g the im portance o f patents.
In the later stages o f indu stry evolution, the R & D scale is often high, and barriers to
entry tend to be built up by incum bents, especially against small firm s (see e.g.
Granstrand and Sjolander 1990, and A rora et al. 2001). The use o f various patent
portfolio strategies such as blanketing or “ evergreening" together with litigation
threats by large firm s (both incum bents and diversifying entrants) m ay serve this
purpose (see G ranstrand 19 9 9 ,2 0 0 4 ).18 This m ay result in a division o f R8cD labor, in
which small firm s specialize in early-stage R8cD, and license their new technologies to
established firm s specializing in later stages o f the innovation process, and/or seek to
be acquired by established firm s (rather than investing in production and m arketing).
These intersectoral differences in the im portance o f IP R s have led several scholars
to criticize the “ on e-size-fits-а ІГ design o f the patent system. R eform proposals have
suggested a m o re differentiated in d u stry tailored system , e.g. regarding patent
duration o r special n ew (sui generis) types o f IP R s for certain industries such as
software (see e.g. T h u ro w 1997 and R eichm an 1994)* C ounterargum ents to industry-
specific schem es fo r IP R s em phasize the resulting high transaction costs, including
IP adm in istration costs, and the fact that a certain am ount o f indu stry tailoring
already exists in the laws and practices o f patent offices and courts. Perhaps because
o f these com plexities, the debate over “ in d u stry-sp ecific" patent system s has raged
for m any years w ith little o r no action b y in du strial-econ om y governm ents.

10.3.4 Differences in IPRs across Nations

There is widespread consensus today that technical progress, the promotion of
which is one of the purposes o f the patent system, is the major determinant behind
284 OVE G R A N S T R A N D

econ om ic p ro gress.19 It is therefore logical to tu rn to e co n o m ic h isto ry for evidence

on the role o f IP R s. H ow ever, a strong patent system has n ot been necessary for
co u n tries’ indu strialization and econ o m ic grow th. A lth o u g h m a n y countries, in­
clud in g Japan , successfully industrialized in the presence o f a p aten t system (see e.g.
D utton 1984), other countries such as G erm an y, H o llan d , an d Sw itzerland did not
(see e.g. K aufer 1989). S c h iff (19 71). stu d yin g H o llan d and Sw itzerland , found no
evidence that industrialization in these cou ntries w as h am p ered b y the absence o f a
patent system . M o ser (2003), stu d yin g tw o n in eteen th -cen tu ry w o rld fairs (in
L o n d o n 1851 and Philadelphia 1876), fo u n d no evidence that stron g patent laws
increased national levels o f innovative activities b u t co n clu d ed that they influenced
the intersectoral d istrib u tion o f in n ovative activities. L ern er (2000), exam ining 177
p o licy changes across sixty countries o ver the past 150 years, fo u n d that changes
increasing patent protection had m uch stron ger effect o n in w ard patenting by
foreigners than on patenting b y dom estic entities.
T h u s the IP R system in general, and the patent system in p articu lar, have been
neither necessary nor sufficient for h istorically sign ifican t tech nical and/or eco­
n o m ic progress at national and co m p an y le v e l A lth o u g h h ard ly su rprisin g, this is
an im p o rtan t conclusion. There seem s to be so m e con sen su s in the scholarly
literature that the patent system has m ade p o sitive co n trib u tio n s to technical
progress, b ut these contribu tions are secon d ary an d co m p le m e n tary to other
factors, p articu larly other institutional d evelopm ents such as a general property
rights system (see N orth 1981 and N elson 1993).
M oreover, current research provides little gu idan ce on the p o ten tial contributions
o f an internationally strong patent system to the p rospects f o r “ c a tch -u p ” by the less
developed countries in the co n tem p o rary w orld. In deed, a certain am o u n t o f “ free­
rid in g” under a w eak IP regim e, elem ents o f w h ich are ap p aren t in the nineteenth-
cen tu ry U nited States or the Japanese eco n o m y o f the 19 5 0 -8 0 perio d , m ay aid in
successful catch-up. In this context T R IP S m ay be seen as an attem pt by leading
countries and com panies to increase the econ o m ic payo ffs o f their R & D , m aking it
m ore costly fo r d eveloping countries to catch up.

10.4 S u m m a r y a n d C o n c l u s i o n s

Instrum ents, strategies, and policies for p rotection o f IP, ran gin g fro m trade secrets
and tradem arks to copyrigh ts and patents, have a lo n g h isto ry. F ro m this perspec­
tive, TRIPS is o n ly the latest expression o f the links betw een trade and intellectual
p ro p erty protection. D espite this lo n g history, su rp risin g ly little sch o larly attention
 IN NO V A TIO N AND IN T E L L E C T U A L PROPERTY RIGHTS 285

has been devoted to the study of intellectual property rights and innovation. One
impetus for research on IPRs has been a series of controversies in various national
jurisdictions over their effects. These controversies reveal many basic similarities
across nations and centuries, although they often concern different types of IPRs
(patent, copyrights, trademarks, design rights, etc.).
It seems fair to say that until recently the role o f IP R s in national, sectoral and
corporate in n o vation system s has been relatively m odest (although not w ithout
exceptions). H ow ever, a U S-originated pro-patent or p ro -IP era has em erged since
the 1980s, perhaps as a consequence rather than a cause o f a m uch broader and
gradual transition into a new type o f capitalist econom y that is based m ore heavily
on know ledge (and in fo rm atio n ), innovation and intellectual capital. The conse­
quences o f the p ro -IP era have been far-reaching and seem likely to broaden, since
there are no signs o f a reversal o f these global trends. IPR s have been applied to m ore
form s o f intellectual p ro p erty through extensions o f old and creation o f new types o f
rights; they have becom e m ore econom ically valuable and have assum ed a m ore
strategic role in national, sectoral and corporate innovation systems.
M ore research on the com plex relations between IP and innovation is clearly
needed, especially since the em ergence o f the pro-patent era has led to significant
change and created new challenges (e.g. the so-called “ an ticom m on s” problem that
stems from a pro liferatio n o f interdependent IPR s: see H eller and Eisenberg 1998 for
a pioneering w o rk ). Future research in this area needs to focus o n the interplay
between econom ic, technical, and legal dynam ics and, in a better w ay than before,
link theoretical w o rk w ith em pirical longitudinal studies.

N o t e s

1. Helpful comments from the editors and participants in the TEARI-project and the
assistance o f Thomas Ewing are gratefully acknowledged.
2. Breeding rights refer to exclusive time limited rights to commercialize certain cultivated
plant or animal varieties. Database rights refer to exclusive time limited rights to com­
mercialize certain collections o f data or content material, including literary, artistic, and
musical works or collections o f material such as texts, sound, images, numbers, and facts,
However, the term “ database” should in this context not be taken to extend to computer
programs used in the making or operation o f a database. Database rights exist in Europe
since the mid-1990s (although with some older precedents) but not in the US, and has so
far created a great deal o f controversies.
3. The history o f the term “ patent” is interesting in this context, in English the term was
short for “ letters patent,” in turn deriving from “ litterae p a te n tesw h ich in medieval
Europe referred to sealed but open royal letters, granting the holders certain rights,
privileges, tides, or offices. The term derives from Latin patere>meaning “ to be open.”
4. In fact, the patent monopoly rights became an exemption in the Statute o f Monopolies,
which generally limited monopoly privileges. The handing out o f such privileges by the
286 OVE G R A N S T R A N D

royalty had degenerated, and the English Parliament wanted to put an end to it,
apparently recognizing the exceptional importance o f technical progress.
5. See further the Federal Courts Im provem ent Act o f 19S2 and Dreyfuss (1989)*
6. A proposal can e.g. be found in recommendations o f the US Senate Committee TNEC
from the 1940s, see Folk (1942: 281-95).
7. Prof. William Baxter, personal communications.
8. The total cost to Kodak was much more, however, since Kodak paid damages to
customers and legal fees, had to shut down a plant and fire about 700 people, and lost
investments and goodwill (see Granstrand 1999 &ftd Rivette and Klein 1999)* The original
Polaroid damages claim was over $US 5 billion, which if awarded could bankrupt even a
large corporation such as Eastman Kodak.
9. A number of law suits against (alleged) Japanese infringers were brought by US firms,
and a number of out-of-court settlements also were reached among these parties.
Royalty rates for licenses were also raised. In general, these events signified the outbreak
of the so-called patent war between USA and Japan. (See Warchofsky (1994) and
Granstrand (1999).) The increased management attention paid to patenting also con­
tributed to the surge of patenting in the US (see Kortum and Lerner 1999), as well as in
several other developed nations,
10. In fact, the share o f foreigners’ patenting in the USA rose from 22 per cent in 1967 to 40
per cent in 1980 (Evenson in Griliches 1984, p. 92).
11. Important new trade legislation included the Trade and Tariff Act of 1984, which
included Section 301 (authorizing US government to take retaliatory action against
countries judged to give an inadequate IP protection) and Section 501 (authorizing the
President to judge the adequacy of IP protection in granting tariff preferences to a
country), combining to a stick and carrot approach. The Omnibus Trade and Competi­
tiveness Act of 1988 moved further along these lines with a “ Special 301” that required the
US Trade Representative to watch, identify and investigate foreign states denying
adequate IP protection to US firms.
12. Empirical studies have followed this up by collecting data on imitation costs and times in
relation to innovation costs and times for a number o f sectors. Patents do increase
imitation costs, especially in pharmaceuticals, but apart from that industry, patents have
not been essential for the rate o f innovations, at least not before the pro-patent era. (See
in particular Mansfield et al. 1981. See also below.)
13. Central references in this literature include Scherer (1966,1967), Barzel (1968), Dasgupta
and Stiglitz (1980), Fudenberg et al. (1983), and Aghion and Howitt (1992). For over­
views, see Baldwin and Scott (1987), Tirole (1988), Romer (1996) and Aghion and Howitt
(1998).
14. There is in fact a whole service industry emerging around the processing of patent
information for various purposes. Japan is a nation that has significantly fostered and
benefitted from patent information analysis through methods subsumed under “patent
mapping , Disclosure of patent information imposes a disadvantage upon patentees but
its perceived advantages to patentees were found to be significantly higher in Japanese
firms (see Granstrand 1999).
15. Cross-national differences m ight be larger than in dustry differences, however, as was the
case between Japan and Sweden in G ranstrand (1999). One has also to keep in mind that
firms are diversified and m ore so regarding their patent p o rtfo lio than their product
portfolio (see Pavitt 1999).
 IN N O V A T IO N AND IN T E L L E C T U A L PROPERTY RIGHTS 28 7

16. An exception to this statement is the longitudinal study of patenting behaviour in an

industry by Hall and Ziedonis (20m).
17** Note that both secrecy building and secrecy breaking in itself is technology dependent
and thus results in cost changes as new technologies appear (e.g. in cryptography and
chemical analysis).
18. Evergreening refers to a patent strategy aimed at prolonging the effective duration of
patent protection in a business area through continually patenting related inventions,
often incremental ones covering product and process improvements and new applica­
tions» as well as through patenting new, emerging technologies for technology transi­
tions into new product generations in the area.
19. Note that a patent is granted to a technical invention based on whether it is novel to the
world and non-obvious (i.e. its technical advance has a certain size, i.e. it fulfills a
minimum inventive step requirement) but not based on its economic merits (apart
from a general requirement o f industrial applicability or usefulness of the invention),
although the underlying assumption is that by so doing, economic progress will be
stimulated.

R eferen ces

A g h io n , R, and H o w i t t , P. (1992), “A Model of Growth through Creative Destruction,”

Econometrics 60: 323-51.
-------------(1998), Endogenous Growth Theory, Cambridge, Mass., London, England: The
MIT Press.
A naw alt , H. (2003), “ Intellectural Property Scope: International Intellectual Property,
Progress, and the Rule o f Law,” In Granstrand (2003: 55-76).
*A rora , A., F o sfu ri , A., and G a m b a r d e l l a , A. (2001), Markets for Technology: The Eco­
nomics of Innovation and Corporate Strategy>Cambridge, Mass.: MIT Press.
* A r r o w , K. J. (1962), “ Economic Welfare and the Allocation o f Resources for Invention,” in
R. R. Nelson (ed.), The Rate and Direction of Inventive Activity: Economic and Social Factors,
Princeton: Princeton University Press for the National Bureau o f Economic Research,
609-26.
A rundel , A., and K a b l a , I. (1998), “ What percentage o f innovations are patented? Empirical
estimates for European firms,” Research Policy 27(2): 127-41.
^Baldw in , W. L., and S c o t t , J. T. (1987), Market Structure and Technological Change, New
York: Harwood.
B a r t o n , J. H. (2003), “ New International Arrangements in Intellectual Property and Com ­
petition Law”, in Granstrand (2003:105-22).
Barzel , Y. (1968), “ Optimal Timing o f Innovations,” Review of Economics and Statistics 50:
248-355-
B r a n d i - D o h r n , M. (1994), “ The Unduly Broad Claim,” International Review of Industrial
Property and Copyright Law>25(5): 648-57.
C h a n d l e r , Jr. A. D . (1990), Scale and Scope— The Dynamics of Industrial Capitalism,
Cambridge, Mass., and London: Belknap Press o f Harvard University Press.

* Asterisked items are suggestions for further reading.

288 OVE G R A N ST R A N D

C o h en , W. M., G oto , A., A k iy a , A., N elso n , R. R., and W a l s h , J, R (2003), "RM > Infor­
mation Flows and Patenting in Japan and the United States, in Granstrand (200311123—54)•
D asg u pta , P, and S t ig l it z , }. (1980), “ Industrial Structure and the Nature of Innovative
Activity” The Economic Journal 90: 265-93.
D a v id , P. A. (1993), "Intellectual Property Institutions and the Panda’s Thumb: Patents,
Copyrights, and Trade Secrets in Economic Theory and History,” in M. B. Wallerstein, M.
E. Mogee, and R. A. Schoen, (eds.), Global Dimensions of Intellectual Property Rights in
Science and Technology, Washington, DC: National Academy Press, 19-61.
D r e y f u ss , R. C. (1989), "The Federal Circuit: A Case Study in Specialized Courts,” New York
University Law Review 64(1): 1-77.
D u tto n , H. I. (1984), The Patent System and Inventive Activity during the Industrial Revolu­
tion 1750-1852, Manchester: Manchester University Press.
E v en so n , R. E. (1984), "International Invention: Implications for Technology Market An­
alysis,” in Griliches (1984: 89-126).
F olk , G. E, (1942), Patents and Industrial Progress: A Summary, Analysis and Evaluation of the
Record on Patents of the Temporary National Economic Committee, New York: Harper 8c
Brothers.
F u d e n b e r g , D., G il b e r t , R., S t i g l i t z , J., and T iro le , J. (1983), "Preemption, Leapfrogging,
and Competition in Patent Races,” European Economic Review 22: 3-31.
G ra n str a n d , O. (1982), Technology, M anagem ent an d M arkets: A n Investigation ofR&Dand
Innovation in Industrial Organizations, London: Pinter.
* ---- (1999), The Economics and Management of Intellectual Property, Cheltenham: Edward
Elgar.
* ---- (ed.) (2003), Economics, Law and Intellectual Property, Dordrecht: K luw er Academic
Publishers.
------ (2004), “ The Economics and Management o f Technology Trade— Towards a Pro­
licensing Era?” International Journal of Technology Management 27(2/3): 209—40.
------ and S j o l a n d e r , S. (1990), "The Acquisition o f Technology and Small Firms by Large
Firms,” Journal of Economic Behavior and Organization 13: 367-86.
G r il ic h e s , Z. (1984), R&D, Patents, and Productivity, Chicago: University o f Chicago Press.
H a l l , В. H., and Z ied o n is , R. H. (2001), "The Patent Paradox Revisited: An Empirical Study
o f Patenting in the U.S. Semiconductor Industry, 1979-1995” RAND Journal of Economics
32(1): 101-28.
H a r a b i , N. (1995), Appropriability o f Technical Innovations: An Empirical Analysis,”
Research Policy 24: 981-92.
J a ff e , A. (2000), The U.S. Patent System in Transition: Policy Innovation and the Innov­
ation Process,” Research Policy 29; 531-57.
K a u f e r , E. (1989), The Economics of the Patent System, New York: Harwood Academic
Publishers.
K lem perer, P. (1990), “ How Broad should the Scope of Patent Protection be?” RAND
Journal of Economics 21(1): 113-30.
K o r t u m , S., and L e r n e r , J. (1999), “ What is behind the Recent Surge in Patenting?” Research
Policy 28: 1-22.
L e r n e r J, (2000), 150 Years o f Patent Protection,” NBER , Working paper No. 7478.
,

L e v i n , R. C „ K l e v o r i c k , A. K., N e l s o n , R. R., and W in t e r , S. G. (1987), "Appropriating the

Returns from Industrial Research and Development,” Brookings Papers on Economic

Activity 3: 783-831.
 IN N O V A T IO N A N D I N T E L L E C T U A L P R O P ER T Y RIGHTS 289

Machlup , F. (1958), An Economic Review of the Patent System, Study No 15 of the Subcommit­
tee on Patents, Trademarks, and Copyrights of the Committee on the Judiciary, US Senate,
Washington, DC: US Government Printing Office.
M a n s f ie l d , E. (1985)» 'How Rapidly Does New Industrial Technology Leak Out?; The
Journal of Industrial Economics, XXXIV, No. 2. 217-23
—— (1986), Patents and Innovation: An Empirical Study”, Management Science 32(2):
173-81.
------Rapo po rt , J., R omeo , A., W a g n er , S., and B ea r d sley , G. (1977), “ Social and private
rate o f return from industrial innovations,” Quarterly Journal of Economics 71 (May):
221-40.
—— Sch w artz , M., and W a g n e r , S. (1981), “ Imitation Costs and Patents: An Empirical
Study,” Economic Journal 91 (Dec.): 907-18.
♦ M a n s f i e l d , E, D., and M a n sfie ld , E. (guest eds.) (2000), “ Intellectual Property Protection
and Economic Development,” International Journal of Technology Management, Special
issue, 19 (1/2).
♦ M a s k u s , K. E. (2000), Intellectual Property Rights in the Global Economy; Washington, DC:
Institute for International Economics.
*M azzo leni , R., and N elso n , R. R. (1998), “ The Benefits and Costs o f Strong Patent
Protection: A Contribution to the Current Debate,” Research Policy 27(f): 273-84.
M erges , R. R, and N elso n , R. R. (1990), “ On the Complex Economics o f Patent Scope,”
Columbia Law Review 90(4): 839-916.
M oser , R (2003 ), “ How Do Patent Laws Influence Innovation? Evidence from 19th-Century
World Fairs,” National Bureau o f Economic Research, Working Paper No. 9909.
M owery , D. (1996), The International Computer Software Industry New York: Oxford
University Press.
N elson , R. R. (ed.) (1993), National Innovation Systems: A Comparative Analysis, New York
and Oxford: Oxford University Press.
N ordhaus , W. D. (1969), Invention, Growth and Welfare, Cambridge, Mass.: MIT Press.
N orth , D. C. (1981), Structure and Change in Economic History; New York: W.W. Norton.
O rdover , J. A. (1991), “A Patent System for Both Diffusion and Exclusion,” Journal of
Economic Perspectives 5(1): 43-60.
Pa v itt , K. (1999), Technology»Management and Systems of Innovation, Cheltenham: Edward
Elgar.
P enrose , E. T. (1951), The Economics of the International Patent System, Baltimore: Johns
Hopkins University Press.
R eichm an , J. H. (1994), “ Legal Hybrids between the Patent and Copyright Paradigms,”
Columbia Law Review 94(8): 2432-558.
R iv ette , K. G., and K l in e , D. (1999), Rembrandts in the Attic: Unlocking the Hidden Value of
Patents, Boston: Harvard Business School Press.
R o m e r , D. (1996), Advanced Macroeconomics, Berkeley, University o f California: McGraw-
Hill.
Sam uelson , P. (1993), “A Case Study on Computer Programs,” in Wallerstein et ai. (1993:
284-318).
Sch erer , E M. (1966), “ Time-Cost Tradeoffs in Uncertain Empirical Research Projects,”
Naval Research Logistics Quarterly 13: 71-82.
---- -(1967), “ Research and Development Resource Allocation under Rivalry, Quarterly
Journal of Economics 81:359-94*
290 OVE G R A N S T R A N D

Scherer , F. M. (1983), “The Propensity to Patent,” International Journal o f Industrial

Organization 1(1): 107-28.
----- (2004), “ Global Welfare in Pharmaceutical Patenting” The World Economy (forthcom­
ing).
*------ and Ross, D. (1990), Industrial Market Structure and Economic Performance, 3rd edn,,
Boston: Houghton Mifflin.
------ and W a t a l , J. (2002), “ Post-trips Options for Access to Patented Medicines in De­
veloping Nations”, Journal of International Economic Law 3(4): 913-39.
Sc h if f , E. (1971), Industrialization without Patents, Princeton: Princeton University Press,
T aylo r , С T., and S ilber sto n , Z. A. (1973), The Economic Impact of the Patent System,
A Study of the British Experience, Cambridge: Cambridge University Press.
T hurow , L. C (1997), “ Needed: A New System o f Intellectual Property Rights” Harvard
Business Review (Sept-O ct.): 95-103.
T iro le , J. (1988), The Theory of Industrial Organization, Cambridge, Mass.: MIT Press.
V e r s p a g e n , B. (2003), “ Intellectual Property Rights in the World Econom y” in Granstrand
(2003: 489-518).
*W a l le r s t e in , M. B., M ogee , M. E., and S ch o en , R. A (eds.) (1993), Global Dimensions of
Intellectual Property Rights in Science and Technology, Washington DC: National Academy
Press.
W a r sh o fsk y , F. (1994), Patent Wars, Chichester: John Wiley 8c Sons,
W r ig h t , B. D. (1983), “ The Economics o f Invention Incentives: Patents, Prizes, and Research
Contracts,” American Economic Review 73(4): 691-707.
 C H A P T E R 11

THE GEOGRAPHY
OF I N N O V A T I O N
REGIONAL
INNOVATION SYSTEMS

B J 0 R N T. A S H E I M
M E R I C S. G E R T L E R

li.i In tr o d u c tio n

There are two paradoxical characteristics of the contemporary global economy.

First, innovative activity is not uniformly or randomly distributed across the
geographical landscape. Indeed, the more knowledge-intensive the economic activ­
ity, the more geographically clustered it tends to be. The best examples include
industries such as biotechnology or financial services, which have become ever more
tightly clustered in a small number o f major centers, despite the attempts o f many
other places to attract or generate their own activities in these sectors. Second, this
tendency toward spatial concentration has become more marked over time, not less
(Leyshon and Thrift 19971 Feldman 2001; Cortright and Mayer 2002). This reality
contradicts longstanding predictions that the increasing use o f information and
communication technologies would lead to the dispersal o f innovative activity over
292 B J 0 R N T. AS HEI M AND MERI C S. G E R T L E R

time. Given these rather striking stylized facts, it would appear that the process of
knowledge production exhibits a very distinctive geography.
We argue in this chapter that this geography is fundamental, not incidental, to the
innovation process itself: that one simply cannot understand innovation properly if
one does not appreciate the central role o f spatial proxim ity and concentration in
this process. Our goal is to demonstrate why this is true, and to examine how
innovation systems at the subnational scale play a key part in producing and
reproducing this uneven geography over time.
This chapter addresses four key issues. First, w hy does location “ matter” when it
comes to innovative activity? If one considers the production and circulation o f new
knowledge to be the core o f innovation, then it is important to have a sound
understanding o f the nature o f the different types o f knowledge involved and their
geographical tendencies. Second, what are regional innovation systems, and what
role do they play in generating and circulating new knowledge leading to innov­
ation? Third, what is the relationship between regional systems o f innovation and
institutional frameworks at the national level? Finally, what is the relationship
between local and global knowledge flows, and is there any evidence that the global
nature o f today’s economy has weakened or altered the influence o f proximity on the
geography o f innovation?

11.2 T y p e s of K n o w l e d g e a n d
th eir G eographies

A grow in g b o d y o f thought argues that in a com petitive era in w h ich success depends
increasingly upon the ability to produce new o r im p ro ve d p ro d u cts and processes,
tacit know ledge constitutes the m ost im p o rtan t basis fo r in n o vation -b ased value
creation (Pavitt 2002), A s M askell and M alm b erg (1999; 172) have put it, when
everyone has relatively easy access to exp licit/cod ified kn ow ledge, the creation of
unique capabilities and products depends on the p ro d u c tio n and use o f tacit
know ledge:

1 hough often overlooked, a logical and interesting consequence of the present development
towards a global economy is that the more easily codifiable (tradable) knowledge can be
accessed, the more crucial does tacit knowledge become for sustaining or enhancing the
competitive position of the firm.. . . In other words, one effect of the ongoing globalisation is
t at many previously localised capabilities and production factors become ubiquities. What
is not ubiquified, however, is the non-tradable/non-codified result of knowledge creation—
the embedded tacit knowledge that at a given time can only be produced in practice. The
 T H E G E O G R A P H Y OF I N N O V A T I O N 293

fundamental exchange inability of this type of knowledge increases its importance as the
internationalisation ot markets proceeds.

Im plicit in the above quote is a fundam entally spatial argum ent: tacit knowledge
is a key determ inant o f the geography o f innovative activity. There are two closely
related elements to this argum ent. First, because it defies easy articulation or
codification (Polanyi i 95$> 1966), tacit know ledge is difficult to exchange over long
distances. It is h eavily im bued w ith m eaning arising from the social and institutional
context in which it is produced, and this context-specific nature makes it spatially
sticky (Gertler 2003). T h e second relates to the changing nature o f the innovation
process itself and, in particular, the grow ing im portance o f socially organized
learning processes. T h e argum ent here is that innovation has com e to be based
increasingly on the interactions and know ledge flows between econom ic entities
such as firm s (custom ers, suppliers, com petitors), research organizations (univer­
sities, other pu blic and private research institutions), and public agencies (technol­
ogy transfer centers, developm ent agencies). T his is fundam ental to Lundvall and
Johnsons (1994) learning econ om y thesis, and is especially well reflected in their
concept o f “ learning through interacting.” W hen one com bines these two features o f
the innovation process— the centrality o f “ sticky,” context-laden tacit know ledge
and the grow ing im portan ce o f social interaction— it becom es apparent w hy geog­
raphy now “ m atters” so m uch.
The recent literature on learning regions further explores the character and
geographical consequences o f tacit know ledge (see Lundvall and Johnson 1994;
Florida 1995; A sheim 1996, 2001; M organ 1997; C ooke and M organ 1998; Lundvall
and M askell 2000). It argues that tacit know ledge does not “ travel” easily because its
transm ission is best shared through face-to-face interaction between partners who
already share som e basic com m onalities: the sam e language; com m on “ codes” o f
com m unication and shared conventions and norm s that have been fostered b y a
shared institutional environ m en t; and personal know ledge o f each other based on a
past h istory o f successful collaboration or inform al interaction. These co m m on al­
ities are said to serve the vital p u rp ose o f bu ilding trust between partners, which in
turn facilitates the local flo w o f tacit (and codified) know ledge between partners.
This approach adopts the learning-by-interactin g m odel as the cornerstone o f its
conceptual fram ew ork, and argues that the production o f tacit know ledge occurs
sim ultaneously w ith the act o f tran sm ission— p rim arily through the m echanism o f
user-producer interaction (Lu n dvall 1988; G ertler 1995)* A ccording to this perspec­
tive, know ledge does not flo w u n id irection ally from technology producers to users.
Instead, users p ro vid e tacit and proprietary, codifiable know ledge to producers in
order to enable the latter to devise innovative solutions to users" practical problem s.
But at the sam e tim e, b y su p p lyin g users w ith innovative technologies, producers are
also sharing their tacit and other p ro p rietary know ledge w ith their custom ers. The
end product arisin g fro m this close interaction benefits both users and producers,
294 B J 0 R N T. A S H E I M A N D M E R I C S . G E R T L E R

and em bodies w ithin it new know ledge that co u ld not h ave been p ro d u ced b y either
p arty w o rk in g in isolation. This, in effect, describes a so cial process o f jo in t innov­
ation and know ledge produ ction .
Lam (1998, 2000) points out that the skills req u ired fo r effective knowledge
transfer w ithin collective learning processes are h ig h ly tim e- and space-specific.
Interactive, collective learning is based on co m p atib le intra- o r interorganizational
routines, tacit norm s and conventions regulatin g collective actio n as well as tacit
m echanism s for the ab sorp tion o f codified know ledge. T h is requires that the actors
in question have a shared understandin g o f alocal codes,” o n w h ich collective tacit as
w ell as d isem bodied codified know ledge is based (A sh eim 1999; Lu ndvall 1996).
T h us, the ab ility to interpret local codes in consisten t w ays w ill be critical for the
integration o f the operations o f a firm w ith in a lo cal in terfirm learn in g network.
Since spatial p ro xim ity is key to the effective p ro d u c tio n an d transmission/
sharing o f tacit know ledge, this reinforces the im p o rtan ce o f in n o vative clusters,
districts, and regions. M oreover, as M askell an d M a lm b e rg (1999) p o in t out, these
regions also benefit from the presence o f localized capab ilities an d intangible assets
that further strengthen their centripetal pu ll (D o si 1988; S to rp e r 1997). M an y o f these
are social assets— i.e. they exist between rather than w ith in firm s. A lth ou gh they are
therefore not fu lly appropriable b y in d ivid u al firm s, o n ly lo cal firm s can enjoy their
benefits. These assets include the region's u n iq u e in stitu tio n al endow m ent, which
can act to su p p o rt and reinforce local advantage. Because such assets evolve slowly
over tim e, exh ibiting strong tendencies o f p ath -d ep en d en t d evelopm ent (David
1994; Z ysm an 1994), they m ay prove to be v ery d ifficu lt to em ulate b y would-be
im itators in other regions, thereby preservin g the in itial ad van tage o f “ first mover”
regions. M askell and M alm b erg argue (19 9 9 :18 1):

It is the region s distinct institutional endowment that embeds knowledge and allows for
knowledge creation which—through interaction with available physical and human re­
sources constitutes its capabilities and enhances or abates the competitiveness of the
firms in the region. The path-dependent nature of such localised capabilities makes them
difficult to imitate and they thereby establish the basis of sustainable competitive advantage.

We discuss the precise nature o f this “ d istinct in stitu tio n al en d o w m en t” in the

follow in g section o f this chapter. Before d oin g so, how ever, it is im p o rtan t to explore
further the different types o f know ledge base in the econ om y, since the precise roles
o f tacit and codified (or codifiable) know ledge tend to d iffer accordingly.

11.2.1 Industrial Knowledge Bases

W hen one considers the actual know ledge base o f v ario u s in d u stries and sectors of
the econom y, it is clear that know ledge and in n o vatio n have b eco m e increasingly
com plex in recent years. There is a larger variety o f kn ow led ge sources and inputs to
 T H E G E O G R A P H Y OF I N N O V A T I O N

be used by organizations and firm s, and there is m ore interdependence and a finer
division o f lab ou r am on g actors: individuals, com panies, and other organizations
(Cowan et a l 2000). N on aka and Takeuchi (1995) and Lundvall and Borras (1999)
have pointed out that the process o f know ledge generation and exploitation requires
a dynam ic interplay betw een, and transform ation of, tacit and codified form s o f
knowledge as well as a strong interaction o f people w ithin organizations and
between them. Thus, these know ledge processes have becom e increasingly inserted
into various form s o f netw orks and innovation system s— at regional, national and
international levels (see Ch. 3 b y Powell and G rodal in this volum e for a discussion o f
the role o f netw orks in in n ovation ).
Despite the general trend tow ards increased diversity and interdependence in the
knowledge process, Pavitt (1984) and others have argued that the innovation process
o f firms is also strongly shaped by their specific know ledge base, which tends to vary
systematically b y indu strial sector (see also Ch. 1 b y Fagerberg and Ch. 15 by von
Tunzelmann and A cha, in this volu m e). For the purposes o f this chapter, we
distinguish between tw o types o f know ledge base: “ analytical” and “ synthetic”
(Laestadius 1998). These types entail different m ixes o f tacit and codified knowledge,
as well as different codification possibilities and lim its. T h ey also im ply different
qualifications and skills, reliance on different organizations and institutions, as well
as contrasting in n o vation challenges and pressures.1
A synthetic know ledge base prevails in industrial settings where innovation takes
place m ainly throu gh the app lication o r novel com bination o f existing knowledge.
Often this occurs in response to the need to solve specific problem s arising in the
interaction w ith clients and suppliers. In d u stry exam ples include specialized indu s­
trial m achinery, plant engineering, and shipbuilding. R & D is in general less im p o rt­
ant than in other sectors o f the econom y. W hen it occurs, it tends to take the fo rm o f
applied research, but m ore often it involves increm ental product or process devel­
opment related to the solu tion o f specific problem s presented by custom ers (von
Hippel 1988). U n iversity 4 n d u stry links are relevant, but they are clearly m ore
significant in the realm o f app lied research and developm ent than in basic research.
Knowledge is created less in a deductive process or through abstraction than through
an inductive process o f testing, experim entation, com puter-based sim ulation, or
practical work. K n ow ledge em bodied in the respective technical solution or engin­
eering w ork is at least p artially codified. H ow ever, tacit know ledge seems to be m ore
im portant than in oth er types o f activity, due to the fact that know ledge often results
from experience gained at the w orkplace, and through learning by doing, using, and
interacting. C o m p ared to the second know ledge type (“ analytical ) described
below, m ore concrete kn ow -h ow , craft and practical skill is required in the k n o w ­
ledge pro d u ction and circu latio n process. These form s o f know ledge are often
provided b y pro fession al and polytech n ical schools, or b y on-th e-job training.
The in n o vation process fo r industries w ith a synthetic know ledge base tends to be
oriented tow ards the efficien cy and reliability o f new solutions, or the practical
296 B J 0 R N Т. А S Н Е I М AND MERI C S. G E R T L E R

u tility and user-friendliness o f produ cts fro m the perspective o f the customers.
Inn о vat ion “related activities are d om in ated b y the m o d ifica tio n o f existin g products
and processes. Since these types o f in n o vation are less d isru p tive to existing routines
and organizations, m ost o f them take place in existin g firm s, m ak in g spin-offs and
new firm fo rm atio n for the developm ent and exp lo itatio n o f n ew synthetic know­
ledge relatively infrequent.
In contrast, an analytical know ledge base d om in ates e co n o m ic activities where
scientific know ledge is h igh ly im portan t, and w h ere k n o w led ge creation is often
based on fo rm al m odels, codified science and ration al processes. P rim e exam ples are
b io tech n o logy and in fo rm atio n technology. B o th basic and ap p lied research, as well
as the system atic developm ent o f produ cts and processes, are central activities in this
form o f know ledge prod u ction . C o m p an ies typ ically have th eir o w n in-house R&D
departm ents but they also rely on the research o u tp u t o f u n iversities and other
research organizations in their in n ovation process. U n iv e rsity -in d u stry links and
netw orks are thus im portan t, and this type o f in teraction is m o re frequent than in
the synthetic type o f know ledge base. K now ledge in pu ts and o u tp u ts in this type o f
know ledge base are m ore often codified (o r read ily codifiable) than in the case o f
synthetic know ledge. T h is does not im p ly that tacit kn ow led ge is irrelevant, since
both kinds o f know ledge are always involved in the process o f know ledge creation
and in n ovation (N onaka et al. 2000, Joh n son , Lorenz, an d L u n d vall 2002).
T h e im portance o f codification in analytic kn ow ledge reflects several factors:
know ledge inputs are often based on review s o f existin g (cod ified ) studies, know­
ledge generation is based on the app lication o f w id ely sh ared and understood
scientific principles and m ethods, know ledge processes are m o re fo rm a lly organized
(e.g. in R & D departm ents), and ou tcom es tend to be d o cu m en ted in reports,
electronic files, or patent descriptions. K n ow led ge ap p licatio n takes the form o f
new products or processes, which are m ore likely to constitu te rad ical innovations
than in those industries for w hich synthetic know ledge constitu tes the principal
know ledge base. N ew firm s and sp in -o ff com panies (i.e. n ew m arket entrants rather
than existing firm s) are an im portan t co n d u it fo r the a p p licatio n o f knowledge
em bodied in these radically new inventions or p ro d u cts.2
H o w is the im portan ce o f tacit, as opposed to cod ified , know ledge, as well as the
geography o f in n ovation , affected b y this d ifferential im p o rtan ce o f synthetic and
analytical know ledge bases across industries and technologies? Clearly, the 'learn in g
through interacting scenario at the core o f the learn in g e c o n o m y and learn-
ing regions thesis seems to be based im p licitly o n activities fo r w h ich synthetic
form s o f know ledge are central. For instance, m an y o f Lu ndvalFs (1988) original
exam ples com e from the realm o l m echanical engineerin g and specialized industrial
m achinery, where non-linear, iterative interaction betw een users and producers
represents the p rim ary m ode o f innovation. For such e co n o m ic activities, the spatial
concentration ot interacting firm s sharing a co m m o n social and institutional
context is an o bviou s prerequisite to socially organized , interactive learning
 T H E G E O G R A P H Y OF I N N O V A T I O N

processes (Gertler 2004)» But what about those sectors for which analytical know­
ledge is pre-em inent? G iven the greater prom inence o f codified and eodifiable
knowledge in the in n ovation process, m ight we not expect innovation processes
within analytically based industries to be m ore w idely distributed spatially?
A pparently not. F o r starters, econom ists have produced m uch striking evidence
about the highly uneven geography o f innovation in analytically based activities.
One im portant approach proceeds by m easuring knowledge spillovers through the
use o f indicators such as patent citations.3 For exam ple, in their classic study, Jaffe et
al. (1993) find evidence that patent applicants in analytically based industries cite
other patents o rigin atin g in the sam e city m ore frequently than they cite patents
originating non-locally, Fu rtherm ore, they find that patent citations are m ore likely
to be localized in the first year following the establishment of the patent, with the
effect fading over time, as the knowledge diffuses more widely.
A related app roach tracks know ledge spillovers in analytically based industries
such as b iotech nology and pharm aceuticals through the analysis o f “ star scientists.57
Zucker, Darby, and colleagues have tracked the location o f these highly productive
scientists and their impact on innovation in the local economy, demonstrating that
the rates o f start-up o f new biotech firms are significantly higher in those regions in
which these key scientists live and work (Zucker and Darby 1996; Zucker, Darby, and
Armstrong 1998; Zucker, Darby, and Brewer 1998). Moreover, firms that have
established working relationships with star scientists outperform firms that do not
enjoy this kind o f access, in terms o f productivity growth, new product develop­
ment, and employment growth.
Both o f these sets o f findings strongly suggest that in fact the innovation process in
industries based on analytical forms o f knowledge is no less spatially concentrated
than those forms o f innovative economic activity based on synthetic types of
knowledge. Indeed, if anything, there is compelling evidence to suggest that the
former may exhibit an even higher degree o f geographical concentration than the
latter (Cortright and M ayer 20 0 2).4
How can one explain this counterintuitive finding? What are the processes
underlying innovation in analytically based industries that explain their distinctive
and highly uneven geography? There are three principal forces at work here. First, it
is clear that, despite the importance o f eodifiable knowledge in analytically oriented
sectors, the circulation o f new knowledge remains highly localized, as the economic
literature on knowledge spillovers (reviewed above) attests. This is because these
spillovers occur first, fastest and most readily within established local social net­
works o f scientists— often b y word o f mouth, well before formal results are p u b ­
lished in widely accessible outlets. Some forms o f valuable knowledge are almost
never transmitted non-locally. For example, knowledge concerning failures in
scientific experiments is rarely, if ever, published. Yet, the knowledge that a particular
research strategy failed to yield expected results can save research teams considerable
time an d expense i f it prevents them fro m p u rsu in g unproductive lines o f inquiry
298 B J 0 R N T. A S H E I M A N D M E R I C S. G E R T L E R

(Enrigh t 2003). The existence o f this type o f lo calized kn o w led ge c ir c u la t io n -

und erp in ned b y co m m o n ly shared fram es o f experien ce an d understandings— has
been highlighted in the recent w o rk o f Storper an d V enables (2003), w h o have coined
the term “ buzz” to capture this ph en om en on .
Second, the central im portance o f h igh ly educated (and potentially footloose)
w orkers in the production o f innovations in analytically based industries means that
those places that offer the m ost attractive em ploym ent opp ortu n ities w ill be favored
over others. O nly a relatively sm all n um ber o f places o ffer a local lab or m arket that is
sufficiently rich and deep to prom ise not just one but a series o f challenging employ­
m ent opportunities in w hich these people can w o rk at the cuttin g edge for well-known
firm s o r research institutes (Florida 20 0 2a). In other w ords, these w orkers are attracted
to those places that offer this kind o f career-based “ buzz,” and w here they can also find
a critical m ass o f people w orking in the sam e o r sim ilar occu patio n al categories. Once
a particular place becom es recognized b y such w orkers fo r its p o rtfo lio o f attractive
em ploym ent prospects, as well as b y em ployers for its deep p o o l o f h igh ly skilled labor,
increasing-returns dynam ics will generate a p ow erfu l v irtu o u s circle o f long-term
grow th and dynam ism for analytically based sectors.
Third, those locations that offer a high quality o f life in add ition to attractive career
opportunities w ill have an even m ore m arked advantage in the “ battle for talent”
These highly talented w orkers can live in m any places, b u t they tend to choose to live in
those cities that offer a high quality o f place, defined b y a particu lar social character.
A ccording to Florida (2002h), such places are im bued w ith a critical m ass o f creative
activity and workers, strong social diversity (m easured in term s o f ethnic or national
origin) and tolerance (best indicated by, for exam ple, a large gay popu lation ). Florida
argues that such places have low barriers to entry fo r talented new com ers from diverse
social backgrounds, m aking it easy for them to gain en try to local social networks and
labor markets. They are also likely to offer colorful, attractive neighborhoods and
cultural am enities that further enhance the attractive po w er o f such places. The more
highly educated (and creative) the worker, the stronger this effect w ill be. Hence, in
those industries with the m ost know ledge-intensive w orkforce, we ought to find the
strongest degrees o f geographical concentration.

11.3 R egional In n o v a tio n System s

and Localized L earn in g

Having presented the most important arguments to explain the consistent tendency
towards the geographical concentration o f innovative activities, we turn now to
 T H E G E O G R A P H Y OF I N N O V A T I O N 299

consider the role o f in n ovation system s at the subnational level in fostering and
prom oting this process.
The concept o f a regional in n ovation system (R IS) is a relatively new one, having
first appeared in the early 1990s (A sheim 1995, A sheim and Isaksen 1997; C ooke 1992,
1998,2001), follow in g F reem an s use o f the innovation system concept in his analysis
o f Japan s econ om y (Freem an 1987b and at approxim ately the sam e tim e that the
idea o f the national in n ovation system was exam ined in books by Lundvall (1992)
and Nelson (1993)* A s this ch ro n o logy suggests, the regional innovation system
concept was inspired b y the national innovation system concept, and it is based on
a similar rationale that em phasizes territorially based innovation system s.5
One such rationale stem s fro m the existence o f technological trajectories that are
based on “ sticky” know ledge and localized learning w ithin the region. These can
become m ore innovative and com petitive b y prom oting stronger system ic relation­
ships between firm s and the re g io n s know ledge infrastructure. A second rationale
stems from the presence o f know ledge creation organizations whose output can be
exploited for econ om ically useful pu rposes b y supporting newly em erging econom ic
activity. The em ergence o f the concept o f a regional innovation system coincides
with the success o f regional clusters and industrial districts in the post-Fordist era
(Asheim 2000; A sh eim and C o ok e 1999; Piore and Sabel 1984; Porter 1990,1998), and
the elaboration o f the concept represents an attem pt b y students o f the geographical
econom y to understand better the central role o f institutions and organizations in
prom oting in n ovation -b ased regional grow th (A sheim et al. 2003; G ertler and W olfe
2004).6
The regional in n o vation system can be thought o f as the institutional infrastruc­
ture supporting in n o vation w ith in the produ ction structure o f a region. Taking each
element o f the term in tu rn (A sheim and C o oke 1999), the concept o f region
highlights an im p o rtan t level o f governance o f econom ic processes between the
national level and the level o f the in d ivid u al cluster or firm . Regions are im portant
bases o f econom ic co ord in atio n at the m eso-level: athe region is increasingly the
level at w hich in n o vation is produced through regional netw orks o f innovators, local
clusters and the cross-fertilizin g effects o f research institutions” (Lundvall and
Borras 1999:39). In v ary in g degrees, regional governance is expressed in both private
representative organizations such as branches o f indu stry associations and ch am ­
bers o f com m erce, and pu blic organizations such as regional agencies w ith pow ers
devolved fro m the national (or, w ithin the E uropean U nion, supranational) level to
prom ote enterprise and in n o vation su p port (A sheim et al. 2003; C o oke et al. 2000).
The system ic d im en sio n o f the R IS derives in part fro m the team -like character
associated w ith in n o vatio n in netw orks. A lth ou gh an innovation system is a set o f
relationships betw een entities o r nodal points involved in innovation (see Lundvall
1992 for m ore d iscu ssion ), it is m uch m ore than this. Such relationships, to be
systemic, m ust involve som e degree o f interdependence, though to varyin g degrees.
Likewise, not all such system ic relations need be regionally contained, but m any are.'
ЗОО В J 0 R N Т. A S H E I M A N D M E R I C S. G E R T L E R

A s the interactive m ode o f in n o vation grow s in im p o rtan ce, these relations are more
likely to becom e regionally contained, especially in the case o f specialized suppliers
w ith a specific tech n ology o r know ledge base. Su ch su p p liers o ften depend on tacit
know ledge, face-to-face interaction and tru st-b ased relation s and, thus, benefit from
co op eration w ith custom ers in region al clusters, w h ile cap a city subcontractors are
increasingly sourced globally.8 Fu rther rein forcin g the system ic character o f the RI$
is the prevalence o f a set o f attitudes, values, n o rm s, rou tin es, and expectations—
described b y som e as a distinctive “ regional cu ltu re” — that influ ences the practices
o f firm s in the region. A s n oted earlier, it is this co m m o n region al culture— itself the
pro d u ct o f co m m o n ly experienced in stitu tion al forces— that shapes the w ay that
firm s interact w ith one anoth er in the region al econom y.

11.3.1 Varieties of Regional Innovation Systems

The “ in n ovation system ” concept can be u n d ersto o d in b o th a n arro w as well as a
b road sense (see C h. 7 b y E dquist, in this vo lu m e). A n a rro w d efin ition o f the
in n o vation system p rim arily in corp orates the R8dD fu n ctio n s o f universities, public
and private research institutes and co rp oratio n s, reflectin g a to p -d o w n , linear model
o f in n o vation as exem plified b y the triple h elix ap p ro ach (E tzkow itz and Leydesdorff
2000). A b road er conception o f the in n o vatio n system s in clu d es “ all parts and
aspects o f the econom ic structure and the in stitu tion al set-u p affecting learning as
w ell as searching and e xp lo rin g ” (Lun dvall 19 9 2 :12 ). T h is b ro ad d efin ition incorpor­
ates the elem ents o f a b otto m -u p , interactive in n o vatio n m o d el o f the sort described
in o u r earlier discussion o f the “ learning region s” concept.
In o rd er to reflect the conceptual variety and em p irical rich ness o f the relation­
ships linking the pro d u ction structure to the “ in stitu tion al se t-u p ” in a region,
A sh eim (1998) distinguishes am o n g three types o f R ISs (see also C o o k e 1998; Asheim
and Isaksen 2002). The first type m ay be denoted as territorially embedded regional
innovation systems, where firm s (p rim arily those em p lo yin g synthetic knowledge)
base their innovation activity m ain ly on localized learn in g processes stim ulated by
geographical, social and cultural proxim ity, w ith o u t m u ch direct interaction with
know ledge organizations. T h is type is sim ilar to w h at C o o k e (1998) calls “ grassroots
R IS, and im plies the b roader d efinition o f in n o vatio n system s described by Lund­
vall (1992) above.
1 he best exam ples o f territorially embedded regional innovation systems are net­
w orks o f SM E s in industrial districts. T h u s in Italy's E m ilia -R o m a g n a , for example,
the innovation system can be described as territo rially em b ed ded in spatial struc­
tures o f social relations w ithin that particu lar region (G ran o vetter 1985). These
territorially em bedded system s provide b o tto m -u p , n etw ork-based support
through, for exam ple, technology centers, in n o vation netw orks, o r in d u stry centers
 T H E G E O G R A P H Y OF I N N O V A T I O N

providing market research and intelligence services, to promote the “ adaptive

technological and organizational learning in territorial context" (Storper and
Scott 1995: 513)**
Another type of RIS is the r e g io n a lly n etw o rk ed in n o v a tio n system . The firms and
organizations are still embedded in a specific region and characterized by localized,
interactive learning. However, policy interventions lend these systems a more
planned character through the intentional strengthening of the region’s institutional
infrastructure— for example, through a stronger, more developed role for regionally
based R&D institutes, vocational training organizations, and other local organiza­
tions involved in firms’ innovation processes. The networked system is commonly
regarded as the ideal type of RIS; a regional cluster of firms surrounded by a regional
“supporting" institutional infrastructure. Cooke (1998) also calls this type “ network
RIS." The network approach is most typical of Germany, Austria, and the Nordic
countries.

Box 11.1 Baden-Wurttemberg's regionally networked innovation system

The German state of Baden-Wiirttemberg is one of the country's most prosperous

regions. It is home to some of Germany's most important mechanical engineering
firms, including Daimler-Chrysler, Porsche, and Robert Bosch. These firms are well
supported by a highly developed network of small and medium-sized enterprise
specializing in the development, production, and supply of components, machinery,
and systems, within a finely articulated social division of labor. Their most important
local competence is their ability to solve the complex technological problems of their
customers, resulting in custom-designed solutions or improvements to existing prod­
ucts and processes. While analytical knowledge is not irrelevant, synthetic knowledge
predominates throughout this set of industries.
Given the importance of these supplier firms to the competitiveness of the regions
large, flagship firms (and hence, to the overall performance of the regional economy),
the regional innovation system has evolved to produce and diffuse these competencies
in incremental mechanical engineering innovation. The most Important elements of
this regionally networked innovation system are:
* A strong vocational education, apprenticeship, and training system that produces a
highly skilled and versatile work force.
* A well-developed infrastructure for technology transfer, incorporating both basic
research facilities and market-oriented development, with special focus on the needs
of SMEs. The Steinbeis Foundation operates a region-wide network of tech-transfer
offices to help SMEs solve technical problems.
* A well-organized Chamber of Commerce (IHK), in which membership is manda­
tory. Among other responsibilities, the Chamber plays a leading role in co-ordinat­
ing the design of training programs tailored to local industry’s needs.
* ffigbly developed and specialized, regionally organized producer associations. These
organizations conduct research on market trends, economic forecasting and
ernerging, market-ready technologies on behalf of producer firms.
302 B J 0 R N T . A S H E I M A N D M E R I C S. G E R T L E R

In addition to these regionally based elements, the entire regional innovation system
is embedded within a national regulatory fram ework that reinforces innovative айіѵііу
at the regional scale* The most important features o f this'- system are: :;
* Labor market structures that fester stable employment relations, ѣ а Ш М щ к т Щ
by doing and strengthening employers’ incentives to train. я
* An industrial relations system that formalizes worker participation in day-to-day
and longer-term strategic decision making, enabling employers to harness workers'
tacit knowledge acquired through learning by doing and using*
* Centralized collective bargaining systems that minimize interfirm variation in .щ р ,
and benefit levels, inducing firms to compete on the basis o f quality and innovative-
ness.
* Capital market structures that encourage longer-term time horizons in firm-level
decision making, thus further reinforcing stability in the workplace*

Sources: Morgan (1999); Gertler (2004)*

The region ally netw orked in n o vation system is a result o f p o lic y intervention to
increase in n o vation cap acity and collab oration . S M E s, fo r exam p le, m ay have to
supplem ent their in fo rm al know ledge (characterized b y a h igh tacit component)
w ith com petence arising from m ore system atic research an d d evelo p m en t in order
to carry out m ore radical innovations. In the lo n g ru n , m ost firm s cannot rely
exclusively o n in fo rm al localized learning, but m ust also gain access to w id er pools of
both analytical and synthetic know ledge on a n ation al an d glo bal basis. T h e creation
o f region ally netw orked in n o vatio n system s th ro u gh increased cooperation with
local universities and R & D institutes, o r th ro u gh the estab lish m en t o f technology
transfer agencies and service centers, m ay p ro vid e access to in fo rm a tio n and com­
petence that supplem ents firm s' locally derived com petence. T h is n o t o n ly increases
their collective innovative capacity, but m ay also serve to co u n teract technological
ulo ck -in " (the in ab ility to deviate fro m an established b u t o u tm o d ed techno­
logical trajectory) w ithin regional clusters o f firm s.
The third m ain type o f R IS, the regionalized national innovation system, differs
fro m the tw o preceding types in several ways. First, parts o f in d u stry and the
institutional infrastructure are m ore fu n ctio n ally integrated in to n ation al or inter­
national in n ovation system s— Le. in n o vation activity takes place p rim arily in co­
operation w ith actors outside the region. T h u s, this represents a developm en t model
in w h ich exogenous actors and relationships play a larger role. C o oke (1998)
describes this type as dirigiste R IS, reflecting a n arro w er d efin itio n o f an innov­
ation system in corp oratin g m ain ly the R & D fu n ctio n s o f un iversities, research
institutes, and corp orations. Second, the co llab o ratio n betw een organizations
w ithin this type o f R IS conform s m ore closely to the lin ear m o d el, as the cooperation
p rim arily involves specific projects to develop m ore rad ical in n o vatio n s based on
form al an alytical-scien tific know ledge. W ithin such system s, co o p eratio n is most
 T H E G E O G R A P H Y OF I N N O V A T I O N 303

likely to arise between people w ith the sam e occupational or educational back­
ground (e.g. am o n g engineers). This functional sim ilarity facilitates the circulation
and sharing o f know ledge through “ com m unities o f practice,” whose m em bership
m ay cross inter-regional and even international boun daries (A m in and Cohendet
2 0 0 4 )*
One special exam ple o f a regionalized national innovation system is the clustering
o f R & D laboratories o f large firm s and/or governm ental research institutes in
planned “ science p a r k s ” These m ay be located in close pro xim ity to universities
and technical colleges, but the evidence suggests that science park tenants typically
have lim ited linkages to local in d u stry (Asheim 1995). Science parks are, thus, a
typical exam ple o f a planned innovative m ilieu com prised o f firm s w ith a high level
o f internal resources and com petence, situated w ithin weak local cooperative envir­
onments. These parks have generally failed to develop innovative networks based on
interfirm cooperation and interactive learning w ithin the science parks them selves
(Asheim and C o o k e 1998; H en ry et al. 1995)- Technopoles, as developed in countries
such as France, Jap an , and Taiw an, are also characterized by a lim ited degree o f
innovative interaction betw een firm s w ithin the pole, and by vertical subcontracting
relationships w ith n o n -lo cal external firm s. In those rare cases where local in n o va­
tive networks arise, they have n orm ally been orchestrated by deliberate pu blic sector
intervention at the national le v e l These characteristics im ply a lack o f local and
regional em beddedness, and lead us to question the capability o f science parks and
technopoles to p ro m ote innovativeness and com petitiveness m ore w idely w ithin
local industries (especially SM E s) as a prerequisite for endogenous regional devel­
opm ent9 (A sheim and C o o k e 1998; and Longhi and Q uere 1993).

11.4 T he R ela tio n sh ip Betw een

Reg io n a l and N a t io n a l In n o v a tio n
System s

Recent w o rk ap p lyin g the R IS concept has begun to explore the linkage between the
larger institutional fram ew o rks o f the national innovation system and national
business system , and the character o f regional in n ovation systems. This question
has recently been addressed b y C o oke (2001) in studies o f biotech nology in the U K ,
the U SA , and G erm any. C o o k e has Introduced a distinction between the traditional
regional in n o vation system (w hich he refers to as the institutional regional in n o v­
ation system — IRIS) and the new econ om y in n ovation system (N E IS), which he
304 B J 0 R N Т. AS HEI M AND MERI C S. G E R T L E R

refers to as an entrepreneurial regional in n o vatio n system (E R IS ) (C o o k e 2003). The

trad ition al IR IS is m o re typ ical o f G erm an region s su ch as B ad en -W u rttem berg or
regions in the N o rd ic countries, w h ose lead in g in d u stries d ra w p rim a rily from
synthetic know ledge bases. Its effectiveness flo w s fro m the p o sitive effects o f sys­
tem ic relationships betw een the p ro d u ctio n stru ctu re and the k n o w led ge infrastruc­
ture em bedded n etw orking governance structu res and su p p o rtin g regu latory and
institu tion al fram ew orks. A cco rd in g to C o o k e, the IR IS fo rm w o rk s w ell where
tech n ology and in n o vation tends to be path dep en d en t rath er than disruptive (the
latter being m ore typical o f the E R IS set-u p ), w here in stitu tio n s have grow n incre­
m en tally to m eet needs in an evolvin g but w e ll-u n d e rsto o d sectoral innovation
system ” (C o o ke 2003: 57).
In contrast, the N E IS o r E R IS (fo u n d in the U S, the U K , an d other Anglo-
A m erican econom ies) lacks the strong system ic elem ents o f the IR IS fo rm discussed
above, and instead gets its d ynam ism from lo cal ven tu re capital, entrepreneurs,
scientists, m arket dem and, and incubators to su p p o rt in n o v a tio n that draws pri­
m arily from an analytical know ledge base. T h u s, C o o k e calls this a "Venture capital
d riven ” system . Such a system is m ore flexible and ad ju stab le and, thus does not run
the sam e risk o f ending up in “ lo ck -in ” situations. O n the o th er h an d, new economy
in n o vation system s do not have the sam e lo n g -term stab ility and provide less
system ic su p p o rt for historical tech nological trajectories, raisin g im p o rtan t ques­
tions about their long-term econom ic sustainability.

Box 11.2 US biotechnology clusters: entrepreneurial regional

innovation systems

A recent study (Cortright and Mayer 2002) o f biotechnology in the United States
concludes that innovation in this industry is dominated by just a handful o f metropol­
itan centers. Boston and San Francisco are the two largest and best-established centers,
followed by recent entrants San Diego, Seattle, and Raleigh-Durham. Philadelphia,
New York, Washington-Baltimore, and Los Angeles also have significant concentra­
tions of biotech activity. The authors of this study conclude that the two most
important overarching factors supporting the emergence of strong biotech clusters
are: (i) the presence of first-class pre-commercial medical research in a local university
or government laboratory, and (ii) local systems to support and encourage entrepre­
neurial activity leading to successful translation of research into commercially viable
outputs.
In emphasizing these two factors, this study emphasizes the key roles played by both
public and private sectors actors in such innovation systems. Government support,
through key granting councils such as the National Institutes of Health and state-level
programs to invest in university systems and research, is at least a necessary condition
for the emergence of a local biotechnology cluster. It also plays a role in the active-
recruitment of star scientists. But this process also requires the local presence of
 T H E G E O G R A P H Y OF I N N O V A T I O N 305

venture capital and managerial expertise in the development of technology-based

companies, For the reasons outlined earlier in this chapter, a high local quality of life
is also a crucial determinant o f a region's ability to attract and retain highly educated
scientific workers.
Despite these important local processes, non-local forces and relationships also play
a key role. Three o f the most important include: organizations such as the National
Institutes o f Health mentioned above; research alliances with large, global pharma­
ceutical firms with the financial resources to bankroll expensive research and clinical
trials; and non-local venture capitalists with money and expertise to identify and
support promising local firms and commercially viable research.
Once such local centers become established and attain critical mass, they begin to
attract inward investment from multinational firms, who set-up their own research
facilities in these locations in order to tap into distinctive local research competencies.
This further reinforces the technological dynamism, entrepreneurial capabilities and
commercialization potential of the region, setting in motion a virtuous circle of
increasing returns.
A dditional sources: Cooke (2001; 2003); Feldman (2001); Zucker et al. (1998).

In m aking these argum ents about a general correspondence between the m acro-
institutional characteristics o f the econ om y and the dom inant form and character o f
its regional in n o vation system s, C ooke provides a link to another useful literature on
"Varieties o f capitalism ” and national business system s (Lam 1998, 2000; W hitley
1999; H all and Soskice 2001). Soskice (1999) argues that different national insti­
tutional fram ew orks evolve to su p port particular form s o f econom ic activity— i.e.
that coordinated m arket econom ies such as G erm an y and the N ordic countries base
their com petitive advantage in “ diversified quality prod u ction ” (Streeck 1992), while
liberal m arket econom ies such as the U S and U K are m ost com petitive in industries
characterized b y science-based innovative activities. W ithin the coordinated m arket
economies, the d rivin g force is the non-m arket coordination and cooperation that
exists inside the business sphere and between private and public actors, as well as the
degree to w hich lab or is m ean in gfu lly “ in corp orated ” w ithin the production process
and the financial system is able to su p p ly long-term finance (Soskice 1999)* In a
com parison between coordinated m arket econom ies such as Sweden, Germ any, and
Switzerland on the one hand, and liberal m arket econom ies such as the U S and U K
on the other, he fo u n d that the coordinated econom ies perform ed best in the
production o f “ relatively co m p lex products, involvin g com plex production p ro ­
cesses and after-sales service in w ell-established industries (e.g. synthetically based
sectors such as the m ach in e too l indu stry). B y contrast, the U S perform ed best in
industries p ro d u cin g co m p lex system ic products such as IT and defense technology,
where the im p o rtan ce o f analytical, scientific-based know ledge often w ith the
m ajor su p po rt o f the state— is significant (Soskice i 999 : U 3- 1 4 )*
ЗОб B J 0 R N T. A S H E I M A N D M E R I C S. G E R T L E R

T h u s, Soskice argues that com petitive strength in m arkets fo r d iversified quality

p ro d u ctio n is based on p ro b lem -so lvin g know ledge d evelo p ed th ro u g h interactive
learn in g and accum ulated collectively in the w o rk fo rce (Soskice 1999 )* T h is type o f
p ro d u ctio n system is incom patible w ith an em p lo ym en t relatio n in w hich work
processes are controlled exclusively b y m an agem en t— a preference generated by
certain finance an d governance system s fo u n d in lib eral m ark et econom ies. C om ­
petitive strength in other m arkets— e.g. m arkets ch aracterized b y a h igh rate of
change th ro u gh radical in n ovation s— is based o n the in stitu tio n al freed om as well as
finan cial incentives to co n tin u ou sly restructure p ro d u c tio n system s in light o f new
m arket o p p ortu n ities (G ilp in 1996). W hile co o rd in ated m ark et econom ies on the
m acro level su p p o rt cooperative, lon g-term , and co n sen su s-b ased relations between
private as well as pu blic actors, liberal m arket econ o m ies in h ib it the developm ent o f
these relations but instead offer the o p p o rtu n ity to q u ick ly ad ju st fo rm al structures
to new requirem ents. Such institu tion al specificities b oth co n trib u te to the form a­
tion o f divergent n ation al business system s, and constitu te the co n text w ithin which
different organizational form s w ith different m ech an ism s fo r learn in g, knowledge
accu m u lation and know ledge ap p ro p riatio n h ave evolved (A sh eim and Herstad
2003).
C h risto ph erson (2002) has argued that the k in d s o f o rgan izatio n al features and
lab or m arket characteristics o f interest to Lam (1998, 2000) are shaped by the
structure o f capital m arkets and “ investm ent regim es” d eterm in ed at the national
level. M oreover, these different investm ent regim es p ro d u ce the societal conditions
fo r divergent form s o f com petitive advantage in global m arkets. A n Am erican-style
“ m arket governance m o d el” dom in ated b y the d rive to m axim ize sh ort-term invest­
m ent returns has pro m oted the em ergence o f U S strengths in analytically based
sectors such as b io tech n o logy and IC T , as w ell as in a set o f “ project-oriented”
industries such as electronic m edia and entertainm ent, ad vertisin g, management
consulting, public relations, engineering and in d u strial design, an d com puter ser­
vices.
T h e essence o f C h ristoph erson s argu m en t is that, u n d er di vergen t sets o f national
institution s govern in g capital and lab o r m arkets and co rp o ra te governance, the
kinds o f social relationships that are lik ely to d evelop betw een econ o m ic actors
locally and hence the social organization o f local in n o va tio n and production
system s w ill v a ry dram atically. Clearly, there is co n sid erab ly m o re em phasis in
the U S system (than in, say, the G erm an system ) on the role o f in d iv id u a l workers as
m obile agents o f know ledge circu lation and lo cal social learn in g, since they are the
p rin cipal actors responsible for the sharing o f know ledge betw een firm s. Grabber $
(2001, 2002) recent w ork on the project-based nature o f p ro d u c tio n organization in
the L o n d on advertising in d u stry d ocu m ents m a n y stru ctu ral sim ilarities with
C h risto ph erson s description o f U S-style, project-b ased eco n o m ic activity, strongly
suggesting the con tin u in g viab ility of a d istinctive “A n g lo - A m e ric a n ” m odel of
regionally based pro d u ction and in n ovation system s.
 T H E G E O G R A P H Y OF I N N O V A T I O N 30/

Thus, liberal m arket econom ies as represented by the U S and the U K seem to have
advantages in industries characterized by an analytical knowledge base, as well as in
those sectors that depend on a high degree o f m o b ility in labor markets. Concerning
the form er, the elite universities and education institutions, often privately o rgan ­
ized, provide strengths in R8cD, the generation o f form alized knowledge, inventions,
and radical innovations. O ther institutional features such as close u n iversity-in d u s­
try links, academ ic spin -offs and an active scientific labor m arket all operate to
prom ote the transfer and application o f scientific knowledge.
Placed in this context the classic “ trad ition al” institutional regional innovation
system typified by a region such as G erm any's Baden-W iirttem berg is m ost com pat­
ible with the institutional fram ew ork o f a coordinated m arket econom y, while the
new econom y in n o vation system (Lon d on advertising, Silicon Valley, or N ew York
City's new m edia “ Silicon A lley” ) reflects the institutional fram ew ork o f a liberal
market econom y.
This raises an im p o rtan t issue that has been the subject o f som e debate in the
literature, concerning the extent to w hich m arkedly different regional innovation
systems can em erge w ith in the sam e national institutional space. Saxenian s (1994)
landm ark study co m p arin g the electronics and com puting industries in two d o m ­
inant regions o f the U nited States— C aliforn ia's Silicon Valley and Route 128 in
M assachusetts— has reinforced the view that w idely divergent regional innovation
systems can and do em erge w ithin a single national institutional fram ew ork. She
argues that Silicon Valley o u tp erform ed R oute 128 in term s o f em ploym ent grow th
and new firm fo rm atio n because o f its m ore open, flexible, h igh -m obility system
com pared to R ou te 128's m ore closed, rigid, hierarchical, loyalty-based system.
While both regions w ere h om e to w orld -renow ned institutions o f higher learning
and research, the Silicon Valley system proved m ore effective in generating successful
innovations in response to p ro fo u n d com petitive challenges from abroad.
It is im portan t to note, how ever, that this analysis was based on the evolution o f
these two regional system s d u rin g the 1980s and early 1990s. In fact, the M assachu­
setts innovation system experienced a p ro fo u n d transform ation over the 1990s. Best
(2001) docum ents this tran sition , describing the em ergence o f new industries in
biotechnology, m edical devices, nanotechnology, and related fields. He argues that
this tran sform ation w as und erp in ned b y a m ore fundam ental shift in the social
organization o f the lead in g sectors o f the regional econom y tow ards open systems
architecture— in oth er w ord s, th rou gh its evolution tow ards a structure that m uch
more closely resem bles the new econ om y in n ovation system o f Silicon Valley. Best s
analysis suggests that Saxenian's earlier case studies captured two regional in n o v­
ation system s at a tim e w hen one o f them (R oute 128) was exhausting the innovative
capabilities o f an older, already ou tm od ed system , but before a coherent, fully-
form ed alternative had em erged (K enney and vo n B u rg 1999; Saxenian 1999)* N ow
that the organizational contours o f this new system are clear, they suggest that the
character o f d ifferent regional innovation systems within the same national
308 B F 0 R N T . A S H E I M A N D M E R I C S. G E R T L E R

institutional space may vary within a considerably narrower range than was previ­
ously thought. This variation is likely to depend primarily upon regionally specific
technological trajectories and knowledge bases.

11.5 A l t e r n a t iv e O r g a n iz a t io n a l F o rm s
a n d E m e r g in g R e l a t io n s h ip s B e t w e e n
L o c a l a n d G l o b a l K n o w l e d g e

Q uestions have lately been raised over w hether the sp atial em b ed dedn ess o f learning
and know ledge creation m ight be challenged b y alternative o rg an izatio n al form s—
in particular, tem p o rary organizations— w hich som e see as b eco m in g m ore preva­
lent in the global econ o m y (A sheim 2002; G rab h er 2002). F o r exam ple, G ann and
Salter (2000) suggest that firm s in the con stru ction and en gin eerin g sector now rely
on projects to organize the p ro d u ctio n o f kn o w led ge-in ten sive an d com plex prod­
ucts and system s. W hat im pact m ight the ad o p tio n o f te m p o ra ry fo rm s o f organiza­
tio n have on the spatial em beddedness o f learn in g an d in n o vatio n ? G rabh er's (2002)
w o rk on projects in L o n d o n advertisin g (discussed above) show s h ow co-location
facilitates the co n tin u ou s and rapid reco n figu ratio n o f p ro ject team s as well as the
circulation o f know ledge concerning the com petencies an d exp erien ce o f potential
project partners.
In contrast, A ld erm an (2004) argues that “ there are . . . im p o rta n t a priori or
theoretical reasons w hy a project-based m odel does not fit c o m fo rta b ly with ideas
ab ou t clustering, localized learning and local in n o vatio n netw orks.” H is argument
relies on a recent literature that sees “ co m m u n ities o f p ractice” as key entities driving
the firm 's know ledge-processing activities. T h is literatu re argu es that routines and
established practices shaped b y organizations (o r subset co m m u n ities w ith in organ­
izations) p ro m ote the pro d u ction and sh aring o f tacit and cod ifiable knowledge
(B row n and D u gu id 19 9 6 ,20 0 0 ; W enger 1998). C o m m u n ities o f practice are defined
as gro ups o f w orkers in fo rm ally b ou n d together b y shared experien ce, expertise and
com m itm en t to a jo in t enterprise. These co m m u n ities n o rm a lly self-organize for
the pu rp ose o f solvin g practical problem s facing the larger o rgan izatio n , and in the
process they produce innovations (both pro d u ct and process). T h e com m onalities
shared by m em bers o f the co m m u n ity facilitate the id en tification , jo in t production
and sharing o f tacit know ledge throu gh collab orative p ro b le m -so lv in g assisted by
story-telling and other narrative devices for circu latin g tacit know ledge.
A ccord in g to this view, organizational o r relational p ro x im ity an d occupational
sim ilarity are m ore im portan t than geograph ical p ro x im ity in su p p o rtin g the
 T H E G E O G R A P H Y OF I N N O V A T I O N 309

production, identification, ap p ro priation and flow o f tacit know ledge (Allen 2000;
Am in 2000; A m in and C ohendet 2004). The resulting geography o f innovation
differs from that envisioned by adherents to the learning region approach. In this
view, the jo in t pro d u ction and diffusion/transm ission o f tacit and codifiable kn ow ­
ledge across intra-organizationai boundaries is possible, so long as it is m ediated
within these com m unities. M oreover, because com m unities o f practice m ay extend
outside the single firm to include custom ers or suppliers, know ledge can also flow
across the boundaries o f in divid u al organizations.
Furtherm ore, the com m unities o f practice literature asserts that tacit know ledge
m ay also flo w across regional and national boundaries i f organizational or “ virtual
com m unity” p ro xim ity is strong enough. In other w ords, learning (and the sharing
o f tacit know ledge) need not be spatially constrained if relational pro xim ity is
present. For large, m u ltin ation al firm s with “ distributed” know ledge bases and
multiple sites o f in n ovation , the use o f com m unities o f practice, aided and su p ­
ported b y ever-cheaper and m ore pow erful ICTs and air travel, is seen as an effective
strategy for overcom ing geographical separation.
These argum ents are useful rem inders o f the im portance o f relationships and the
strength o f und erlying sim ilarities rather than geographical p ro xim ity per $e in
determ ining the effectiveness o f know ledge-sharing between econom ic actors.
However, they fail to answ er a key question: w hat forces shape or defines this
“ relational proxim ity,” enabling it to transcend physical, cultural, and institutional
divides? H o w are shared understandings produced? M uch o f the com m unities o f
practice literature is largely silent on this question. A notable exception is the w ork o f
Brow n and D u gu id (2 0 0 0 :14 3) w ho stake out a very different position on the spatial
reach o f co m m u n ities o f practice:

They are usually face-to-face communities that continually negotiate with, communicate
with, and coordinate with each other directly in the course o f work. And this negotiation,
communication, and coordination is highly implicit, part o f work practice . . . In these
groups, the demands o f direct coordination inevitably limit reach. You can only work closely
with so many people.

11.6 C o n c l u sio n s a n d Issu e s

f o r F u t u r e R e se a r c h

In this chapter we have argued that the geographical configuration o f econom ic

actors— firms, workers, associations, organizations, and governm ent agencies is
fundamentally important in shaping the innovative capabilities o f firm s and
310 B J 0 R N Т. A S H E l М A N D M E R I C S. G E R T L E R

industries. W e have distinguished between two types o f knowledge base— synthetic

and analytical and have demonstrated that, while the nature o f innovation pro-
cesses m ay differ in each case, in n ovative activ ity tends to b e sp atially clustered in
both cases, thou gh for som ew hat d ifferent reasons.
We then introdu ced the concept o f regional in n o v a tio n system s, describing the
elem ents, relationships, and system ic character that co m p rise a k ey p art o f a region's
distinct institution al endow m ent. We also exp lored the v a rie ty o f different types o f
regional in n o vation system s that can be id en tified , n o tin g h o w p articu lar regional
system s m a y b e m ore stron gly associated w ith p a rticu la r regim es o f business systems
and institutional fram ew orks at the n ation al level. A lth o u g h there is significant
variatio n in econom ic perform an ce across d ifferent region s w ith in the same na­
tion al system , the characteristics o f successful region al in n o v a tio n system s under the
sam e national regim e w ill exhibit certain consistencies fro m case to case.
D espite the em ergence o f a strong consensus a ro u n d the above issues, there
rem ain a n um ber o f contentious o r u nresolved q u estion s that are likely to provide
the focus fo r future research. First, in the o n g o in g d isco u rse o n the nature and
im pact o f globalization, som e authors have argu ed that tacit k n o w led ge has become
increasingly codified and hence u b iqu itou s, u ltim ately e ro d in g the competitive
advantage o f h igh -cost regions and nations (M askell et al. 1998; M askell 1999),
O thers m aintain that m uch strategic know ledge rem ain s “ sticky,” an d that im port­
ant parts o f the learning process continue to be localized as a result o f the enabling
role o f geographical p ro xim ity and local in stitu tion s in stim u latin g interactive
learning (A sheim 1999; M arku sen 1996).
N evertheless, global know ledge netw orks an d flow s are im p o rta n t sources of
innovative ideas for a grow in g n u m ber o f econ o m ic activities (M ackin n on et al.
2002). I f so, then h o w shou ld we u n d erstan d their im p a ct o n the geographical
d istrib ution o f innovative activity, and on the fu tu re im p o rta n ce o f regionally
based in n ovation systems? In a recent conceptual paper, B ath elt et al. (2004) seek
to reconcile these divergent view s. T h ey argue that firm s clustered in particular
locales require access to non -local sources o f k n ow ledge as an essential complement
to the know ledge they generate and share locally. T h e m e tap h o r th ey have adopted to
capture the dual nature o f em erging geographies o f in n o va tio n is “ local buzz and
global pipelines. T h ey view these global know ledge p ip elin es as extend in g between
different n odal geographical concen trations o f firm s and o th er know ledge-produ­
cing organizations arou n d the w orld. O n the one h an d, no firm — especially in
analytical, science-based sectors such as b io tech n o lo g y— can affo rd to cut them­
selves o ff from n o n -local know ledge sources. To do so w o u ld be to co u rt potential
disaster, as regional innovation system s w o u ld be p ro n e to en cou ragin g techno­
logical stagnation and lock-in tendencies. O n the oth er h an d , the abilities o f firms
to m ake the m ost effective use o f this k n ow ledge— that is, to convert it most
effectively into econom ic value— still depends on their access to im p o rtan t place-
based assets, both tangible and intangible, and the close in teractio n with other
 T H E G E O G R A P H Y OF I N N O V A T I O N 3 11

organizations arou n d them that such locations foster (see also Asheim and Hers tad
2003; C ooke et aL 2000; Freem an 2002).
This w ork is a w elcom e conceptual contribution, but the em pirical basis for its
fram ework rem ains underdeveloped. Clearly, there is a need for future research—
both case studies and aggregate statistical analyses— to investigate the prevalence o f
this “ dual geograp h y" o f innovation m ore systematically. At the sam e tim e, this
approach raises a fu rth er question that rem ains unanswered: is this m etaphor o f
local buzz and global pipelines appropriate only for those science-intensive in d u s­
tries whose in n ovation rests on an analytical know ledge base? To what extent are
non-local know ledge flow s and learning relationships extending between localized
centers o f in n ovation b ecom in g im portan t for those industries that rely m ore
heavily on a synthetic know ledge base?
A problem atic aspect o f the learning econom y and learning regions literature has
been its focus on learning by d oing and using based largely on local synthetic
knowledge w ith a high tacit content and increm ental innovations. We continue to
agree with F reem an s useful insight concerning “ the trem endous im portance o f
incremental in n o vation , learning b y doing, by using and by interacting in the
process o f technical change and diffusion o f in n o vation s" (Freem an 1993: 9 -10 ).
Yet in a highly com petitive, globalizing econom y, it m ay be increasingly difficult for
the reproduction and grow th o f a learning econom y to rely prim arily on increm ental
im provem ents to prod u cts and processes, and not on new products (i.e. radical
innovations). C revo isier (1994:259) argues that the exclusive reliance on increm ental
innovations “ w ou ld m ean that these areas w ill very quickly exhaust the technical
paradigm on w hich they are fou nd ed ."
In future studies it w ill be im p o rtan t to follow these tendencies, which u n d ou b t­
edly will be reinforced b y globalization processes (see C h. 12 by N arula and Zanfei in
this volum e). The basic rationale o f regional innovation system s is that the system ic
prom otion o f localized learn in g processes can im prove the innovativeness and
com petitive advantage o f regional econom ies. W hat rem ains to be seen is how the
capacity o f regional in n o vatio n system s to upgrade the know ledge bases o f firm s in
regional clusters w ill develop over tim e.

N otes

1. Pavitt (1984:353-65) offers a three-way taxonomy o f industries based on the predominant

nature and sources o f technical change. Supplier-dominated industries include agricul­
ture and traditional manufacturing sectors such as textiles. Production-intensive indus­
tries can be further subdivided into scale-intensive sectors such as steel, consumer
durables, and automobiles, and specialized supplier sectors such as machinery and
instruments. Science-based industries include electronics and chemicals (including
pharmaceuticals). In the discussion that follows, our observations about industries with
312 B J 0 R N T, AS H E I M A N D M E R I C S . G E R T L E R

synthetic knowledge bases correspond closely to those sectors encompassed by the first
two o f Pavitts categories (supplier-dominated and production-intensive). Similarly, the
analytical category corresponds directly to Pavitt s science-based industries.
2. We should acknowledge that many industries draw significantly upon both synthetic and
analytical forms o f knowledge. A clear example would be medical devices and technolo­
gies, whose development rests upon knowledge drawn from fields as diverse as bioscience,
ICT, software, advanced materials, nanotechnology, and mechanical engineering. For this
reason, it makes sense to conceive o f individual industrial sectors being arrayed along a
continuum between purely analytical and synthetic industries, with many— such as the
automotive industry— occupying an intermediate position along this spectrum.
3. See Feldman (2000) for a useful overview o f this literature.
4. Cortright and Mayer produce evidence to show that the degree o f geographical concen­
tration in the US life sciences industries is considerably higher than the population as a
whole. They also demonstrate— using indicators such as venture capital, funded research
conducted through inter-firm alliances, and new firm formation rates— that this concen­
tration has increased dramatically during the past two decades.
5. This conceptualization o f regional innovation systems corresponds with the one found in
Cooke et al. (2000). In their words any functioning regional innovation system consists of
two subsystems: (i) the knowledge application and exploitation subsystem, principally
occupied by firms within vertical supply-chain networks; and (ii) the knowledge gener­
ation and diffusion subsystem, consisting mainly o f public organizations.
6. There is a strong historical correspondence between these concepts and approaches and
agglomeration theories within regional science and economic geography, such as Per-
roux s (1970) growth pole theory.
7. In a recent study Carlsson (2003) shows that the majority o f theoretical as well as empirical
analyses o f innovation systems have a regional focus (see also Bathelt 2003 for a critical
discussion o f RIS).
8. A recent comparative study o f European clusters shows that firms increasingly find
relevant research activities and other supporting services inside the cluster boundaries
(Isaksen 2004).
9. See Ch. 8 in this volume by Mowery and Sampat for a similarly critical assessment of
science parks.

R efer en c es

A ld erm a n , N. (2004), Mobility versus Embeddedness: The Role o f Proximity in Major

Capital Projects, in A. Lagendijk and P. Oinas (eds.), P r o x im it y , D is ta n c e a rid D iversity:
Issues on E c o n o m ic In tera ctio n a n d L o ca l D e v e lo p m e n t , Aldershot: Ashgate.
A l le n , J. (2000), “ Power/Economic Knowledge: Symbolic and Spatial formations,” in J. R.
Bryson, P. W. Daniels, N. Henry, and J. Pollard (eds.), K n o w le d g e , Space, E c o n o m y ; London:
Routledge, 15-33.
A m in , A. (2000), “ Organisational Learning through Communities o f Practice,” Paper pre­
sented at the workshop on The Firm in Economic Geography, University o f Portsmouth,
UK, 9-11 March,
and C o h l n d e t , I. (2004), Architectures o f K n o w led g e , Oxford: Oxford University Press.
 T H E G E O G R A P H Y OF I N N O V A T I O N 313

A s h e im , В. T. (1995)? “ Regionale innovasjonssystem— en sosialt og territorielt forankret

teknologipolitikk,” N o rd is k Sa m h dllsgeo grafisk T id sk rift 20:17-34.
------(1996), Industrial Districts as 'Learning Regions’: A Condition for Prosperity?” E u ro ­
p e a n P la n n in g S tu d ies 4(4): 379-400.
------(1998), “ Territoriality and Economics: On the Substantial Contribution o f Economic
Geography” in O. Jonsson and L.-O. Olander, (eds.), E co n o m ic G eo g ra p h y in Transition,
T he S w ed ish G eo g ra p h ic a l Yearbook , vol. 74, Lund, 98-109.
------(i999)> “ Interactive Learning and Localised Knowledge in Globalising Learning Econ­
om ies” G e o jo u rn a l 49(4): 345-52.
------(2000), “ Industrial Districts: The Contributions o f Marshall and Beyond,” in G. Clark,
M. Feldman, and M. Gertler (eds.), T h e O x fo rd H a n d b o o k o f E co n o m ic G eography, Oxford:
Oxford University Press, 413-31.
------(2001), “ Learning Regions as Development Coalitions: Partnership as Governance in
European Workfare States?” C oncepts a n d T ra n sfo rm a tio n . In te rn a tio n a l Jo u r n a l o f A ctio n
R esearch a n d O rg a n iz a tio n a l R e n e w a l 6(1): 73-101.
------(2002), “ Temporary Organisations and Spatial Embeddedness of Learning and Know­
ledge Creation,” G eo gra fiska A n n a le r, Series B, H u m a n G eo g ra p h y ; 848(2): 111-24.
------and C ooke , P. (1998), “ Localised Innovation Networks in a Global Economy:
A Comparative Analysis o f Endogenous and Exogenous Regional Development
Approaches,” C o m p a ra tiv e S o c ia l R esearch 17, Stamford, Conn: JAI Press, 199-240.
-------------(1999), “ Local Learning and Interactive Innovation Networks in a Global Econ­
omy,” in E. Malecki and P. Oinas (eds.), M a k in g C on n ection s: T echnological L e a rn in g a n d
R eg io n a l E co n o m ic C h a n g e , Aldershot: Ashgate, 145-78.
------and H er st a d , S. J. (2003), “ Regional Innovation Systems, Varieties o f Capitalism and
о
Non-Local Relations: Challenges from the Globalising Economy,” in В. T. Asheim and A.
Mariussen (eds.), In n o v a tio n s , R eg io n s a n d P rojects , Stockholm: Nordregio, R 2003:3,
241-74.
— — and I s a k s e n , A. (1997), “ Location, Agglomeration and Innovation: Towards Regional
Innovation Systems in Norway?” E u ro p ea n P la n n in g Stu d ies 5(3): 299-330.
-------------(2002), “ Regional Innovation Systems: The Integration o f Local 'Sticky’ and
Global 'Ubiquitous’ Knowledge,” Jo u r n a l o f Technology T ransfer 27: 77-86.
------et al. (eds.) (2003), R e g io n a l In n o v a tio n P o licy f o r S m a ll-M e d iu m E n terp rises , Chelten­
ham: Edward Elgar.
B a t h e l t , H. (2003), “ Geographies o f Production: Growth Regimes in Spatial Perspective 1:
Innovation, Institutions and Social Systems, Progress in H u m a n G eo g ra p h y 27(6): 789-804.
* ---- M a l m b e r g , A ., and M a s k e l l , P. (2004)5 “ Clusters and Knowledge: Local Buzz,
Global Pipelines and the Process o f Knowledge Creation,” Progress in H u m a n G eography,
28(1): 31-56.
B est , M. (2001), T h e N e w C o m p e titiv e A d v a n ta g e , Oxford: Oxford University Press.
B rown , J. S„ and D u g u id , R (1996), “ Organisational Learning and Communities-of-
Practice: Towards a Unified Theory o f Working, Learning and Innovation,” in M, Cohen
and L. Sproul (eds.), O rg a n isa tio n a l L e a rn in g , New York: Sage, 58-82.
* ----------- (2000), T h e S o c ia l L ife o f In fo rm a tio n , Boston: Harvard Business School Press.*

* Asterisked items are suggestions for further reading.

314 B J 0 R N T , A S H E I M A N D M E R I C S. G E R T L E R

C arlsso n , B. (2003), “ Innovation Systems: A Survey o f the Literature from a Schumpeterian

Perspective/' in A. Pylca (eel.)* T h e C o m p a n io n to N e o -S c h u m p €te ria n E co и0mics, Chelten-
ham: Edward Elgar,
C h r is t o p h e r s o n , S. (2002), “ Why do National Labor Market Practices Continue to Diverge
in the Global Economy? The “ Missing Link" o f Investment R ules" E c o n o m ic G eography
78(1): 1-20.
C ooke , P. (1992), “ Regional Innovation Systems: Competitive Regulation in the New
Europe," G eo fo ru m 23: 365-82.
------ (1998), “ Introduction: Origins o f the Concept” in H. Braczyk, P. Cooke, and M,
Heidenreich (eds.), R e g io n a l In n o v a tio n S ystem s , London: U CL Press, 2—25.
*------ (2001), “ Regional Innovation Systems, Clusters, and the Knowledge Economy,” Indus­
tria l a n d C o rp o ra te C h a n g e 10(4): 945^74*
------ (2003), “ Integrating Global Knowledge Flows for Generative Growth in Scotland: Life
Sciences as a Knowledge Economy Exemplar,” in I n w a r d In v e s tm e n t , E n trep re n e u rsh ip and
K n o w le d g e F lo w s in S c o tla n d — In te rn a tio n a l C o m p a ris o n s , Paris: OECD.
------ and M o rgan , K. (1998), T h e A s so c ia tio n a l E c o n o m y : F ir m s , R e g io n s a n d In n ovation ,
Oxford: Oxford University Press.
-------B o ekho lt , P , and T q dtlin g , E (2000), T h e G o v e r n a n c e o f In n o v a tio n in Europe,
R e g io n a l P ersp ectives on G lo b a l C o m p etitive n ess , London: Pinter.
C o r tr ig h t , J., and M a y e r , H. (2002), Sign s o f L ife : T h e G ro w th o f B io te c h n o lo g y Centers in
the U S., Washington, DC: Center on Urban and Metropolitan Policy, The Brookings
Institution.
C ow an , R., D a v id , P A „ and F o ray , D. (2000), “ The Explicit Economics o f Knowledge
Codification and Tacitness,” In d u s tr ia l a n d C o rp o ra te C h a n g e 9: 211-53.
C r e v o isie r , O. (1994), Book review o f G. Benko and A. Lipietz (eds.), Les R egions qui
g a g n en t , Paris, 1992, E u ro p e a n P la n n in g S tu d ie s 2: 258-60.
D a v id , P A . (1994), “ Why are Institutions the 'Carriers o f History'? Path Dependence and the
Evolution o f Conventions, Organizations and Institutions,” S tr u c tu ra l C h a n g e a n d Eco­
n o m ic D y n a m ic s 5: 205-20.
D osi, G, (1988), “ The Nature o f the Innovative Process,” in G. Dosi et al. (eds.), Technical
C h a n g e a n d E co n o m ic T heory, London: Pinter, 221—38.
E n r ig h t , M. (2003), “ Competitiveness, Innovative Clusters and Positive Externalities”,
paper presented at the Sixth Annual Conference o f the Competitiveness Institute, Goth­
enburg, Sweden, 17-19 September.
E tzko w itz , H., and L eyd esd o rfe , L. (2000), “ The Dynamics o f Innovation: From National
Systems and Mode 2 to a Triple Helix o f University—Industry—Government Relations,”
R esearch P o lic y 29:109-23.
F eld m a n , M, P. (2000), Location and Innovation: The New Economic Geography of
Innovation, Spillovers, and Agglomeration,” in G. L. Clark, M. P Feldman, and M, S.
Gertler (eds.), T h e O x fo rd H a n d b o o k o f E c o n o m ic G e o g ra p h y , O xfo rd : O xford University
Press, 373-94.
(2001), Where Science Comes to Life: University Bioscience, Commercial Spin-offs,
and Regional Econom ic D evelopm ent,” Jo u r n a l o f C o m p a r a t iv e P o lic y A n a ly sis: Research
a n d P ractice 2; 345-61.
F lo rid a , R. (1995), Toward the Learning Region,” F u tu re s 27: 527-36.
— (2002a), The Econom ic G eography o f Talent,” A n n a ls o f th e A sso c ia tio n o f Am erican
G eo g ra p h ers 92: 743-55.
(2002b), f h e R ise of the C re a tiv e C la ss , New York: Basic Books.
 T H E G E O G R A P H Y OF I N N O V A T I O N 315

Fr e e m a n , C. (1987), T echnology P o licy a n d E co n o m ic P erfo rm a n ce: Lessons fr o m Ja p a n ,

London: Pinter.
----- “ (і 993)» The Political Economy o f the Long Wave” Paper presented at EAPE 1993
conference on The economy o f the future: Ecology, technology, institutions”. Barcelona,
October 1993.
-----"(1995)5 The National System o f Innovation’ in Historical Perspective,” C a m b rid g e
Jo u r n a l o f E co n om ics 19: 5-24.
*------(2002), Continental, National and Sub-National Innovation Systems— Complemen­
tarity and Economic Growth” R esearch P o licy 31:19 1-2 11.
G ann , D. M. and S a l t e r , A. J. (2000), “ Innovation in Project-Based, Service Enhanced
Firms: The Construction o f Complex Products and Systems,” R esearch P o licy 29: 955—72.
G ertler , M. S. (i995)>u 'Being there’: Proximity, Organization, and Culture in the Develop­
ment and Adoption o f Advanced Manufacturing Technologies,” E co n o m ic G eo g ra p h y 71:
1-26.
------(2003), “ Tacit Knowledge and the Economic Geography o f Context, or the Undefinable
Tacitness o f Being (there),” Jo u r n a l o f E co n o m ic G eo g ra p h y 3: 75-99.
------(2004), M a n u fa c tu r in g C u ltu re: T h e In stitu tio n a l G eo g ra p h y o f In d u s tria l P ractice ,
Oxford: Oxford University Press.
------and W o lfe , D. A. (2004), “ Ontario’s Regional Innovation System: The Evolution
of Knowledge-Based Institutional Assets,” in P. Cooke, and M. Heidenreich,
H. Braczyk (eds.), R e g io n a l In n o v a tio n System s , 2nd edn,, London: Taylor and Francis.
G ilpin , R, (1996), “ Economic Evolution o f National Systems,” In te rn a tio n a l S tu d ies Q u a r­
te r ly 40: 411-43.
G rabh er , G. (2001), “ Ecologies o f Creativity. The Group, the Village and the Heterarchic
Organisation o f the British Advertising Industry,” E n v iro n m en t a n d P la n n in g A 33:351-74.
------(2002), “ Cool Projects, Boring Institutions: Temporary Collaboration in Social Con­
text,” R e g io n a l S tu d ies 36: 205-14.
G ran o vetter , M. (1985), “ Economic Action and Social Structure: The Problem o f Embedd­
edness,” A m e ric a n Jo u r n a l o f So cio lo g y 91: 481-510.
Hall , P., and S o sk ic e , D. (2001), “An Introduction to Varieties o f Capitalism,” in Hall and
Soskice (eds.), V arieties o f C a p ita lis m : T h e In stitu tio n a l F o u n d a tio n s o f C o m p a ra tiv e A d ­
vantage, Oxford: Oxford University Press, 1-68.
H en ry , N. et al. (1995), “Along the Road: R&D, Society and Space,” R esearch P o licy 24:
707-26.
I saksen , A. (2004), “ Regional Clusters between Local and Non-Local Relations: A Compara­
tive European Study,” in A. Lagendijk and P. Oinas (eds.), P ro x im ity , D ista n ce a n d D iv e r ­
sity: Issues on E c o n o m ic In te ra c tio n a n d L o ca l D e v e lo p m e n t , Aldershot: Ashgate.
J affe , A., T r a jt e n b e r g , M., and H enderso n , R. (1993), “ Geographical Localization o f
Knowledge Spillovers as Evidenced by Patent Citations,” Q u a rterly Jo u r n a l of E conom ics
108:577-98.
J ohnson , B., L o ren z , E., and L u n d vall , B.-A. (2002), “ Why All this Fuss about Codified
and Tacit Knowledge?” In d u s tr ia l a n d C o rp o ra te C h a n g e 11: 245-62.
K en n ey , M., and von B u rg , U. (1999)5 “ Technology, Entrepreneurship and Path Depend­
ence: Industrial Clustering in Silicon Valley and Route 128, In d u s tria l a n d C o rp o ra te
C h a n g e 8(1): 67-103.
L a e s t a d i u s , S. (1998), “ Technology Level, Knowledge Formation and Industrial Compe­
tence in Paper Manufacturing,” in G. Elias son et al. (eds.), Microfoundations of Economic
Growth. A S c h u m p e te ria n Perspective, Ann Arbor: University of Michigan Pi ess, 212 —2 6 .
31 6 B J 0 R N T . A S H E 1 M A N D M E R I C S. G E R T L E R

L a m , A. (1998), “ The Social Embeddedness o f Knowledge: Problems o f Knowledge Sharing

and Organisational Learning in International High-Technology Ventures, D R U ID
W orkin g P a p e r 98-7, Aalborg,
— — (2000), “ Tacit Knowledge, Organizational Learning and Societal Institutions: An
Integrated Framework/5 O rg a n iz a tio n S tu d ie s 21(3): 487—513*
L eysh o n , A., and T h r ift , N. J. (1997). M o n e y /S p a c e : G e o g ra p h ie s o f M o n e t a r y Transform ­
a tio n , London: Routledge.
L o n gh i , C , and Q ubre , M. (1993), “ Innovative Networks and the Technopolis Phenomenon:
The Case o f Sophie-Antipolis” E n v ir o n m e n t a n d P la n n in g C : G o v e r n m e n t & P olicy 11:
317~ 3tK
L u n d v a ll , B-A. (1988), “ Innovation as an Interactive Process: From User-Producer Inter­
action to the National System o f Innovation,” in G. Dosi, C. Freeman, G, Silverberg, and
L. Soete (eds.), T ech n ica l C h a n g e a n d E c o n o m ic T h e o ry ; London: Pinter.
------ (ed.) (1992), N a tio n a l In n o v a tio n System s: T o w a rd s a T h e o r y o f In n o v a tio n a n d Inter­
a c tive L e a rn in g , London: Pinter.
------ (1996), “ The Social Dimension o f the Learning Economy,” D R U I D W o rk in g Papers 96-1,
Aalborg.
------ and J ohnson , B. (1994), “ The Learning Economy,” J o u r n a l o f In d u s t r y Studies 1:23-42.
------ and B orras , S. (1999), T h e G lo b a lis in g L e a r n in g E c o n o m y : Im p lic a tio n s f o r Innovation
P o licy , DGXII-TSER, The European Commission.
------ and M a sk e l l , P. (2000), “ Nation States and Economic Development— from National
Systems o f Production to National Systems o f Knowledge Creation and Learning,” in G. L.
Clark, M. P. Feldman, and M. S. Gertler (eds.), T h e O x fo r d H a n d b o o k o f Econom ic
G eo g ra p h y , Oxford: Oxford University Press, 353-72.
M a c k in n o n , D., C u m be r s , A., and C h a pm a n , K. (2002), “ Learning, Innovation and
Regional Development: A Critical Appraisal o f Recent Debates/5 P ro gress in H um an
G e o g ra p h y 26(3): 293-311.
M a r k u se n , A, (1996), “ Sticky Places in Slippery Space: A Typology o f Industrial Districts/5
E c o n o m ic G e o g ra p h y 72(3): 293-313.
M a s k e l l , P, (1999), “ Globalisation and Industrial Competitiveness: The Process and Conse­
quences o f Ubiquitification,” in E, Malecki and P, Oinas (eds.), M a k in g Connections:
T ech n o log ical L e a rn in g a n d R e g io n a l E c o n o m ic C h a n g e , Aldershot: Ashgate, 35-59.
et al. (1998), C o m p etitiven ess, L o c a lis e d L e a r n in g a n d R e g io n a l D e v e lo p m e n t, London:
Routledge.
* anc^ Ma l m b e r g , A. (1999), “ Localised Learning and Industrial Competitiveness”
C a m b rid g e Jo u r n a l o f E co n o m ics 23:167-86.
M o rga n , K. (1997), The Learning Region: Institutions, Innovation and Regional Re­
newal,” R e g io n a l S tu d ie s 31: 491-504,
------ ( 1999), “ Reversing Attrition? The Auto Cluster in Baden-Wurttemberg,” in T. J. Barnes
and M. S. Gertler (eds.), T h e N e w In d u s tr ia l G e o g ra p h y : R eg io n s, R e g u la tio n a n d Insti­
tutions, London: Routledge, 74-97.
N elso n , R. (ed.) (1993b N a tio n a l In n o v a tio n System s: A C o m p a r a t iv e A n a ly sis, Oxford:
Oxford University Press.
Nonaka , I., and T akeuchi, H. (1995), T h e K n o w le d g e C r e a tin g C o m p a n y , Oxford and New
York: Oxford University Press.
-*~-et al. (2000), SECI, Ba and Leadership; A Unified Model o f Dynamic Knowledge
Creation, L o n g R a n g e P la n n in g 33; 5—34,
 T H E G E O G R A P H Y OF I N N O V A T I O N 317

P a v it t , K. (1984),
Sectoral Patterns of Technical Change: Towards a Taxonomy and a
Theory,” Research Policy 13: 343-73.
------(2002), “ Knowledge about Knowledge since Nelson and Winter: A Mbced Record,”
Electronic Working Paper Series Paper No. 83, SPRU, Uni versity of Sussex, June.
P e r r o u x , R (1970), "Note on the Concept of Growth Poles” in D. McKee et al. (eds.),
Regional Economics: Theory and Practice, New York: The Free Press, 93-103.
P io r e , M., and S a b e l , C. (1984), The Second Industrial Divide: Possibilities for Prosperity, New
York: Basic Books.
P o l a n y i , M. (1958), Personal Knowledge: Towards a Post-Critical Philosophy, London: Rou-
tledge & Kegan Paul.
------(1966), The Tacit Dimension, New York: Doubleday.
P o r t e r , M, (1990), The Competitive Advantage of Nations, New York: Basic Books.
P o r t e r , M. E. (1998), “ Clusters and the New Economics o f Competition ” Harvard Business
Review (Nov.-Dee.): 77-90.
*S a x e n ia n , A. (1994), Regional Advantage: Culture and Competition in Silicon Valley and
Route 128, Cambridge, Mass.: Harvard University Press.
------(i 999)» “ Comment on Kenny and von Burg, 'Technology, Entrepreneurship and Path
Dependence: Industrial Clustering in Silicon Valley and Route 128’,” Industrial and
Corporate Change 8(1): 105-10.
S c h o e n b e r g e r , E. (1997), The Cultural Crisis of the Firm, Oxford: Blackwell.
S o s k ic e , D. (1999), “ Divergent Production Regimes: Coordinated and Uncoordinated
Market Economies in the 1980s and 1990s,” in H. Kitschelt et al. (eds.), Continuity and
Change in Contemporary Capitalism, Cambridge: Cambridge University Press, 101-34.
S t o r p e r , M. (1997), The Regional World: Territorial Development in a Global Economy>New
York: Guilford Press.
------and S c o t t , A. (1995), “ The Wealth of Regions,” Futures, 27(5): 505-26.
*------and V e n a b l e s , A. J. (2003), “ Buzz: The Economic Force o f the City,” Paper presented
at the DRUID Summer Conference 2003 on “ Creating, sharing and transferring know-
ledge: the role o f geography, institutions and organizations,” Elsinore, Denmark.
S t r e e c k , W. (1992), Social Institutions and Economic Performance—Studies of Industrial
Relations in Advanced Capitalist Economies, New York: Sage.
von H i p p e l , E. (1988), The Sources of Innovation, Oxford: Oxford University Press.
W e n g e r , E. (1998), Communities of Practice: Learning, Meaning and Identity, Cambridge:
Cambridge University Press.
W h it l e y , R. (1999), Divergent Capitalism: The Social Structuring and Change of Business
Systems, Oxford: Oxford University Press.
Z u c k e r , L. G., and D a r b y , M. R. (1996), “ Star Scientists and Institutional Transformation:
Patterns o f Invention and Innovation in the Formation of the Biotechnology Industry,”
Proceedings of the National Academy of Science 93:12709-16.
-------------- and A r m s t r o n g , J. (199В), “ Geographically Localized Knowledge: Spillovers or
Markets?” Economic Inquiry 36: 65-86.
-------------- an(i B r e w e r , M. B. (1998), “ Intellectual Human Capital and the Birth o f U.S.
Biotechnology Enterprises,” American Economic Review 88: 290-306.
Z ysm an , J. (1994), “ How Institutions Create Historically Rooted Trajectories o f Growth,”
Industrial and Corporate Changes: 243-83.
 CHAPTER 12

GLOBALIZATION
OF INNOVATION
TJl H
jl.JL F
.н У -ROT
Ж.м
ч M
l. .F
п»'кі/ O F
Ш Jrl

M ULTINATIONAL
ENTERPRISES

RAJNEESH NARULA
ANTONELLO ZANFEI

12.1 In t r o d u c t i o n

E c o n o m ic globalization implies a growing interdependence o f locations and eco­

nomic units across countries and regions. Technological change and multinational
enterprises (M N Es) are among the prim ary driving forces o f this process. In this
apter we attempt to evaluate the changing extent and importance of MNEs as
conduits tor cross-border knowledge flows.
MNEs affect the development and diffusion of innovations across national
кГмп rS 1 r0Ugh 3 number o f mechanisms, am ong which FD I (through which
;s acquire existing assets abroad or set-up new wholly or majority owned
^ m 'f 1 m niarkets) is only one. International knowledge flows also move
rough trade, licensing, cross-patenting activities, and international technological
 G L O B A L I Z A T I O N OF I N N O V A T I O N 319

and scientific collaborations. These other m odalities involve a w ide variety o f

econom ic actors, but the M N E occupies a central role am ong these actors. This
chapter em phasizes the M N E s m ultifaceted role in the m ore general process o f
globalization o f innovation.

1 2 . 2 T r e n d s i n THE
I n t e r n a t i o n a l i z a t i o n of
In n o v a tiv e A c tivities

A useful taxon o m y pro p o sed by A rch ibugi and M ichie (1995) identifies three m ain
categories o f the globalization o f innovation (Table 12.1). A lthough a variety o f
econom ic actors undertake innovation and are engaged in its internationalization,
the M N E is the o n ly institu tion w hich b y definition can carry out and control the
global generation o f in n o vation w ithin its boundaries. We briefly discuss each o f the
three categories below,

12.2*1 The Cross-Border Commercialization of

National Technology
The first category involves national and m ultinational firm s as well as individuals
engaged in the in tern ation al com m ercialization o f technology developed at “ home.”
Key indicators o f these activities are international trade flow s and cross border
patenting, both o f w h ich are responsible for grow ing levels o f global transfer o f
technology.
The share o f h igh -tech produ cts (including electrical and electronic equipm ent,
aerospace produ cts, p recision instrum ents, fine chem icals and pharm aceuticals) in
world exports rose fro m 8 per cent in 1976 to 23 per cent in 2000. Exports o f
inform ation and co m m u n icatio n s technology products showed the highest annual
growth rate am o n g all p ro d u cts in 19 85-2000 (U N C T A D 2 0 0 2 :14 6 -7 ), The rise in the
share o f w orld trade represented b y R&D-intensive sectors suggests that the global ­
ization o f tech n o logy flow s is in creasin g.1
Table 12.2 reveals a grow th in the “ internationalization o f patenting: the share o f
non-resident patentin g in v irtu ally all O E C D econom ies has grow n d u rin g the 1980s
320 RAJNEESH NARULA AND ANTONELLO ZANFEI

Table 12.1 A taxonomy of the globalization o f innovation

Categories Actors Forms

In te rn a tio n a l E x p lo ita tio n o f Profit-seeking Exports of innovative goodfe

Nationally Produced Innovations (national and Cession of licenses and
multinational) firms patents.
and individuals Foreign production of
innovative goods internally
designed and developed.
Global Generation of Innovations MNEs REtD and innovative activities
both in the home and the host
countries.
Acquisitions of existing RED
laboratories or green-field
RED investment in host
countries.
Global Techno-Scientific Universities and Joint scientific projects.
Collaborations Public Scientific exchanges,
Research Centers sabbatical years.
International flows of
students.
National and Joint ventures for specific
Multinational Firms innovative projects.
Productive agreements with
exchange of technical
information and/or equipment

Source: elaboration on Archibugi and M ichie 1995.

and early 1990s, and external patenting (i.e. patent ap p licatio n s o f n ation al inventors
abroad) has also rapidly increased.

12*2*2 Technological and Scientific Collaborations

D om estic and international technical and scientific co llab o ratio n s involve both
private and pu blic institutions, in clu d in g n ation al and m u ltin a tio n a l firm s, univer­
sities and research centers. Since the 1970s, the use b y in d u strial firm s o f “ non­
internal o p tio n s that include co op eration w ith co m p etito rs, su p p liers, customers,
and other external institutions (e.g. universities), w hich w e d en ote as strategic
tech nology partnering (ST P ), has grow n. A vailable in d icators o f international STP
Table 12.2 Rates o f growth o f industrial REfcD and patenting in the OECD countries

Countries Average annual rates of change (per cent)

Industrial RftD (1) Resident patents (2) Non-resident patents (3) External patents (4)

1970-80 1985-95 1970-80 1984-94 1970-80 1984-94 1970-80 1985-95

United States 2.0 1.3 -2 .0 5.7 5.0 6.6 -0 .6 15.6

Japan 6.1 5.4е 5.1 2.2 -0 .8 5.1 5.5 8.3
Germany 4.9a 1.1 -0.7 1.4 0.8 4.6 1.7 8.0
France 3.7 3.2 -2 .4 1.0 0.2 5.3 3.0 8.4
United Kingdom 3.0b 0.3е -2 .4 -0 .4 0.8 4.8 -1.7 16.2
Italy 3.6 —0.5 n.a. 2.51 n.a. 3.8 1.8 10.3
Netherlands 1.4 3.3е -2.1 -1 .5 1.5 6.8 0.1 14.1
Belgium 6.7C 1.7f -3 .0 -1 .6 -0.1 7.7 0.5 13.4
Denmark 3.8 7.49 1.7 3.0 -0.3 19.9 1.0 22.5

G L O B A L I Z A T I O N OF I N N O V A T I O N
Spain 12.7 1.8е -4 .5 2,0 0.2 19.2 1.3 16.0
Ireland 5.2C 15.4 6.8 2.3 4.9 31.1 6.7 24.3
Portugal 4.6d 2.2h —6.4 0.9 -0.5 37.2 -24.2 52.4
Greece n.a. —1.4* -0.8 —13.4m 2.4 37.0 n.a. 21.5
Sweden 5.9C 0.29 -0 .5 0.0 2.5 7.1 3.0 14.2
Austria 9.8a 5.19 0.3 -1 .6 3.4 9.0 1.4 10.1
Finland 6,8C 5.1 4.7 2.7 0.7 13.4 5.7 23.1
Switzerland 0.8a -0.5' -3,1 -1.5 2.2 7.8 -1.3 5.5
Norway 7.3 1.39 -2.7 0.9 -0.1 11.1 0.8 21.1
Australia n.a. 8.9е 5.2 1.5 -2.0 7.5 6.7 21.7
Canada 5.5 4.9 -1.1 2.2 -2.1 4.5 -0.5 21.5
OECD weighted average n.a. n.a. 1.3 2.7 0.9 9.3 0.9 13.3

Notes : n.a. — not available

a1970~81 b 1972-81 c 1971-81 d 1971-80 e 1985-94 f 1985-91 9 1985-93 h 1986-92 11986-93
1992-94 m 1984-93 (1) Million
$USat 1995
PPP
(2) Resident patents: inventors in their home country
(3) Non-resident patents: foreign inventors in the country
(4) External patents: national inventors patenting abroad
Source: Archibugi and lam m arm o (2002) based on OECD, MSTI, various years.
322 RAJNEESH NARULA AND ANTONELLO ZANFEI

have a n u m b er o f w ell-k n ow n draw backs due to the q u a lity o f d ata available.2 In

spite o f these draw backs, there is a general agreem ent in the literatu re that global
in ter-firm alliances have becom e increasingly p o p u la r o ver the past tw o decades
(Hagedoorn 2002; see also Ch. 3 b y Pow ell and G rod al, this v o lu m e ).3 International
S T P has grow n con siderab ly in absolute term s, alth o u gh its sh are o f all ST P has
rem ained steady in the 1970s and 1980s, oscillatin g a ro u n d 60 p er cent o f all
agreem ents, w hile the share has declined in the 1990s to a b o u t 50 p er cent (H age­
d o o rn 2002). T h ere has also been a grad u al shift in the types o f agreem ents favored
b y firm s o ver tim e, according to the M E R IT -C A T I database. T h e percentage o f
eq u ity agreem ents in the total has declined fro m ab o u t 70 p er cent to less than 10
per cent betw een the m id -19 70s and the end o f the 1990s. T h e in creasin g share o f
n o n -eq u ity alliances m ay indicate grow in g use b y M N E s o f S T P s as relatively rapid,
sh ort-term vehicles to gain access to n o n -d o m estic kn o w led ge sources (see Section
12.5 fo r fu rth er d iscussion).
S T P agreem ents appear to be m ost co m m o n in the d o m a in o f n ew materials,
b io tech n o lo gy and in fo rm atio n technology, an d largely in volve T riad econom ies
rather than d evelopin g econom ies. D eve lo p ed -co u n try firm s p articip ate in 99 per
cent o f the ST P agreem ents in the M E R IT -C A T I dataset (H ag ed o o rn 2002).
A lth o u gh R & D and m an u factu rin g o u tso u rcin g agreem ents w ith d eveloping-coun-
try firm s have expanded in n u m ber d u rin g the last tw o decades, the share o f these
firm s in ST P has rem ained arou n d 5 -6 per cent since the 1990s (N aru la and Sadowski
2002). Seventy per cent o f all S T P since the 1960s have h ad at least on e U S partner,
w ith collab oration s between E uropean and N orth A m e ric a n firm s increasing from
18.5 to 25.2 per cent o f overall technological alliances betw een the 1970s and the 1990s
(H aged o orn 2002).

12.2.3 The Role of M NEs in the Cross-Border Generation

of Innovation
The globalization o f in n o vation has been associated w ith grow th in M N E activity
and FD I since W orld W ar II. F D I stocks as a percentage o f G D P 4 stood at 21.46 per
cent in 2001, up from ju st 6.79 per cent in 1982 (Table 12.3). F u rth erm o re, M N Es
^ considerable intra- and in ter-firm trade (Table 12.3). T h e p rim a ry source
o f o u tb o u n d FD I alm ost 90 per cent o f the total in 2 0 0 1— con tin u es to be the
industrialized countries. T h e EU accounted for the largest share o f ou tw ard FDI,
with N etherlands, U K , France, and G erm an y accou n tin g fo r fu lly 41.3 p er cent of all
outw ard 1 DI stock from the developed w orld. A ro u n d 68 per cent o f inw ard FDI is
also directed tow ards Iriad countries. The d evelo pin g e co n o m ies’ increase in the
*hare o f inw ard FD I d u rin g the period 19 8 2 -2 0 0 1 is alm ost entirely due to a small
jro u p o f d eveloping countries, p rim arily the A sian N IC s and C h in a.
 G L O B A L I Z A T I O N OF I N N O V A T I O N 323

Table 12.3 Selected indicators o f FDI and international production,

1 9 8 2 -2 0 0 1 ($US Billion at current prices and percentage values)

1982 2001

FDI inflows 59 735

FDI outflows 28 621
FDI inward stock 734 6846
FDI outward stock r sJX
V
ro, 6582
Sales of foreign affiliates 2541 18517
Gross product of foreign affiliates 594 3495
Total assets of foreign affiliates 1959 24952
Exports of foreign affiliates 670 2600
Employment of foreign affiliates (thousands) 17987 53581

Inward FDI stocks to GDP ratio 6,790/0 21t460/o

Foreign affiliates' export to total exports 32,200/q 34,99°/o

Source: UNCTAD, based on its FDI/TNC database and UNCTAD estimates.

The figures for R & D activity reflect sim ilar patterns, since m any o f the largest
firm s engaged in F D I are key actors in the generation and diffu sion o f innovation.
M ore than o n e-th ird o f the top too M N E s are active in the m ost R & D -in ten sive
industries, such as electronics and electrical equipm ent, pharm aceuticals, chem icals
(U N C TA D 2002). Furthermore, large M N Es play a dominant role in the innovative
activities o f their h om e countries. For instance, Siem ens, Bayer and H oechst per­
formed 18 per cent o f the total manufacturing R & D expenditures in Germany in 1994
(Kumar, 1998). In 1997 three M N E s accounted for more than the 30 per cent o f the
overall U K R & D investment in manufacturing. These same M N Es also undertake a
growing share o f their total R& D activities outside their home countries.
Significant cross-national differences are also apparent in indicators o f inter­
national R&D. The share o f national R& D expenditures accounted for by non­
domestic sources varies substantially within the industrialized and developing areas
(see Table 12.4 for some details). The origins o f international R&D investment flows
also differ considerably among industrialized economies (Table 12.5). Cantwell
(1995) suggests that countries such as Switzerland, UK, and the Netherlands,
which have historically been home to large M N Es and that were long-time inter­
national investors in R&D, have greatly expanded their offshore R&D investments
since World War II. Another group o f countries (such as France and Germany ) has
relatively few large M N Es, and their outward R& D investments have grown more
324 R A JN E E SH NARULA AND A N TO N ELLO ZANFEI

Table 12.4 R&D expenditure of foreign affiliates as a percentage of total R&D

expenditures by ail firms in selected host economies. I f 98 or latest
year

Country Percentage of R a B

Canada 34.2
Finland (1999) 14.9
France 16.4
Japan 1.7
Netherlands 21.8
Spain (1999) 32.8
UK (1999) 31.2
US 14.9
Czech Republic (1999) 8.4
Hungary 78.5
India (1994) 1.6
Turkey 10.1

Source: UNCTAD (2002), table 1.10.

Table 12.5 Shares of US patenting of largest nationally owned industrial

firms due to research located abroad, 1 9 2 0 - 1 9 9 0 (°/o)

1920™1939 1940-1968 1969“ 1990

US 6.81 3.57 6.82

Europe 12.03 26.65 27.13
UK 27.71 41.95 43,17
Germany 4.03 8.68 13.72
Italy 29.03 24.76 14.24
France 3.35 8.19 9.55
Sweden 31.04 13,18 25.51
Netherlands 15.57 29.51 52.97

Source: Cantwell (1995).

rad ually d u rin g the last eigh ty years. A third gro u p inclu des co u n tries that were
major investors in offshore R & D d u rin g the first fourteen years o f the twentieth
„entury. O ffsh ore investm ent by these econom ies actu ally declined after 19 14 and
eturned to pre-W orld War I levels o n ly recently. T h is gro u p inclu des the United
 G L O B A L I Z A T I O N OF I N N O V A T I O N 325

States, hom e o f a n u m ber o f M N E s w hich have a relatively low p ro po rtion o f their

R& D and patenting activity ab ro ad .5
On average, firm s fro m E U countries obtain a larger share o f patents from their
foreign subsidiaries than is true o f U S or Japanese com panies (Table 12.6). D uring
the 1969-95 p eriod, the share o f total patents o f EU firm s attributable to foreign
affiliates grew from 26.3 to 32.5 per cent. European firm s tend to concentrate a
considerable share o f their international R & D activities in the U S (over 50 per cent
o f their foreign R & D investm ent on average, w ith G erm an, British, and Swiss firm s
showing the highest concen tration o f their foreign activities in the United States).
The foreign patenting activity o f U S firm s also increased during this period, but
rem ained below 10 per cent.6 A lth ough U S foreign R & D activities are relatively low
com pared to E U firm s, they are m uch larger than Japanese com panies, whose
offshore patenting declined from 2.1 per cent in 19 6 9 -77 to approxim ately 1 per
cent o f their total patentin g activity in 1987-95.
O verall, M N E s have increasingly internationalized their innovative activities,
with a few relevant exceptions (m ost notably, Japanese M N E s). The im portance o f
R & D activities o f foreign affiliates has grow n in m ost host econom ies over the 1990s.
R & D b y foreign firm s is especially high in the U K , Ireland, Spain, H ungary, and
Canada, and low est in Jap an , w ith other countries (including the U S, France, and
Sweden) in interm ediate positions. N evertheless, m ost R & D and patenting activities
are still largely concentrated in the M N E s' hom e countries, and in a few host
countries. W ell over 90 per cent o f the R & D expenditures o f m ost M N E s is located

Table 12.6 Share of US patents of the world's largest firms attributable to

research in foreign locations by main area of origin of parent firms,
1 9 6 9 -1 9 9 5 (%)

Nationality of parent firm 1969-77 1978-86 1987-95

us 5.4 6,9 8,3

Japan 2.1 1,2 1.0
European countries3 26.3 25.6 32.5
Total alt coun|riesb 10.3 10.7 11,3
Total all countries excluding Japan 11,1 13.0 16.2

^ flerthany, UK, Italy, France, Netherlands, Belgium, Luxembourg, Switzerland, Sweden,

Denmark, Ireland, Spain, Portugal, Greece, Austria, Norway, Finland.
M gta! includes f it the 784 world's largest firms recorded by the University of Reading
database, base year 1984
Source: Cantwell and Jarme (2000).
326 R А JN EESH N A R U L A A N D A N T O N E L L O Z A N FE I

w ith in the T riad .7 W hile there are significant differences in the in tern atio n al disper­
sion o f in n ovative activity across indu stries, firm s have gen erally n o t international­
ized their in n ovative activity at the sam e rate as th eir p ro d u c tio n activities.
Excep tion s to this rule are M N E s o rigin atin g fro m sm all econ o m ies, such as Bel­
giu m , the N etherlands, and Sw itzerland. A large p ro p o rtio n o f even the most
internationalized M N E s concentrate at h o m e their m o re strategic activities,
such as R & D and h eadquarters fu n ction s (B en ito et al. 2003).
T h is relatively low — but increasing— degree o f in tern atio n alizatio n is associated
inter alia w ith the co m p lex nature o f system s o f in n o vatio n , an d the em beddedness
o f the M N E ’s activities in the h om e en viron m en t (see e.g. N a ru la 20 0 2a), the need
fo r internal cohesion w ithin the M N E (Blanc and Sierra 1999, Z a n fe i 2000), and the
high q u ality o f local infrastructures and a p p ro p ria b ility regim es that R & D activities
tend to require. These factors, together w ith the d ifficu lties o f m an ag in g complex
technological po rtfolio s, im p ly that the in tern atio n alization o f in n o va tio n occurs at
a slow er pace than the in tern ation alization o f p ro d u ctio n .

12.3 O v e r se a s In n o v a t iv e A c t iv it ie s
o f MNEs: T h e o r e t ic a l a n d

E m p ir ic a l Iss u e s

T h e extensive literature on international R & D investm ent h igh ligh ts tw o broad

firm -level m otives. First, firm s internationalize their R & D to im p ro ve the w ay in
w hich existing assets are utilized. T h at is, firm s seek to p ro m o te the use o f their
technological assets in con ju n ction w ith, or in respon se to, sp ecific foreign loca­
tional conditions. T h is has been dubbed as asset-exp lo itin g R & D (D u n n in g and
N arula 1995) o r h om e-base exp loitin g (H B E ) a ctiv ity 8 (K u em m erle 1996). For
exam ple, som e m o d ificatio n in these firm s’ p ro d u cts o r processes m a y be necessary
to m ake them com petitive in the relevant foreign m arket. T h is typ e o f offshore R&D
investm ent typ ically is based on the tech nological advan tages o f the source firm,
w hich in turn reflect those o f its h om e country.
A sset-exp loitin g strategies correspond to trad itio n al view s o f the organization of
in n ovative activities and foreign direct investm ent, m a n y o f w h ich w ere rooted in the
prod u ct life cycle th eo ry o f such investm ent. R eferrin g m a in ly to US-based
m ultinationals, Vernon (1966), K in d leb erger (1969), and Sto p fo rd and W ells (1972)
suggested that an M N E s foreign su b sid iary replicated the p a ren t’s non-strategic
 G L O B A L I Z A T I O N OF I N N O V A T I O N

activities abroad, with strategic decisions— including R&D and innovation— being
rigidly centralized in the home country. Vernon emphasized that coordinating
international innovative activities was too costly, due to the difficulties of collecting
and controlling relevant information across national borders. The R&D activities of
foreign subsidiaries were limited largely to the adoption and diffusion of centrally
created technology.
The second b road m otive for offshore R & D investment is strategic asset-
augmenting activity (D u n n in g and N arula 1995), also know n as hom e-base aug­
menting (H BA ) activity (K uem m erle 1996). Firm s use these types o f R & D invest­
ments to im prove existing assets o r to acquire (and internalize) or create com pletely
new technological assets through foreign-located R & D . The assum ption in such
cases is that the foreign lo cation provides access to complementary location-specific
advantages that are less available in its p rim a ry o r “ h om e” base (letto-G illies 2001).
In many cases, the strategic assets sough t b y the investing firm are associated w ith the
presence o f other firm s. A location w hich is hom e to a m ajor com petitor m ay attract
asset-augm enting investm ents b y other firm s in the sam e or in other related in d u s­
tries (see C antw ell in this v o lu m e on the im plications o f these patterns o f FD I for the
competitiveness o f host countries). A sset-augm enting m otives and technology
sourcing have been p artially incorporated in form al m odels o f the FD I decision.9
The asset-augm enting perspective, w hich considers local contexts m ore as
sources o f com petencies and o f technological opportu nities, and less as constraints
to the action o f M N E s, m arks a fundam ental departure from the conventional
wisdom. In a sem inal co n trib u tio n , H edlund (1986: 2 0 -1) caught the essence o f
this new w ay o f con cep tu alizin g the role o f local contexts: “ The m ain idea is that the
foundations o f com petitive advantage no longer reside in any one country, but in
many. N ew ideas an d pro d u cts m ay com e up in m an y different countries and later be
exploited on a global scale.” (See K ogut 1989 for a sim ilar view .)
There are several reasons w h y such asset-augm enting R & D activities are hard to
achieve through m eans oth er than FD I. Som e o f these reasons are associated w ith the
nature o f technology. W h en the know ledge relevant for innovative activities is
clustered in a certain geograph ical area and is “ sticky,” foreign affiliates engage in
asset-augm enting activities in these areas in order to benefit from the external
economies and kn ow ledge spillovers generated b y the concentration o f production
and innovation activities in the relevant clusters. The tacit nature o f technology
implies that even w here know ledge is available through m arkets, it m ay still require
m odification to be efficiently integrated w ithin the acquiring firm s portfolio o f
technologies. T h e tacit nature o f know ledge associated w ith pro d u ction and in n o v­
ation activity in these sectors also m eans that “ physical or geographical p ro xim ity
m ay be im p o rtan t fo r accessing and absorbing it (Blanc and Sierra 1999)- The
m arginal cost o f tran sm ittin g codified know ledge across geographic space does
not depend on distance, b u t the m argin al cost o f transm itting, accessing, and
328 r a jn e e s h n a r u l a AND ANTONELLO ZANFEi

ab sorb in g tacit know ledge increases w ith distance. T h is leads to the clustering of
in n o vatio n activities, especially in the early stages o f an in d u stry life cycle where tacit
know ledge plays an im po rtan t role (A udretsch and F eld m an 199b).
In general, asset-exploiting activities are p rim a rily associated w ith demand-
d riven in n ovative activities (e.g, lo calization o f the p a ren t-firm products for
a specific offshore m arket). A sset-au gm en tin g activities, o n the other hand,
are p rim a rily undertaken w ith the intention to acq u ire an d internalize technological
spillovers that are host lo cation -sp ecific. A sset-ex p lo itin g activity, broadly
speaking, represents an extension o f R & D w o rk u n d ertak en at h om e, while asset-
augm en tin g activity represents a d iversificatio n into n ew scien tific problem s, issues
o r areas.
A n extensive literature has suggested that asset-au gm en tin g internationalization
o f R & D has becom e m o re significant d u rin g the past tw o decades as a result of
several factors that include: ( a) the increasing costs and c o m p le x ity o f technological
d evelopm ent, leading to a grow in g need to exp an d tech n o lo g y so u rcin g and inter­
action w ith different and geograph ically dispersed actors en d o w ed w ith comple­
m en tary bits o f know ledge; ( b) the faster pace o f in n o vative activities in a num ber of
industries, sp u rrin g firm s to search fo r ap p licatio n o p p o rtu n ities w h ich are mainly
location -specific; (c) grow in g pressures fro m h o st g o vern m en ts, w hich have led
M N E s to increase the interaction w ith local partn ers as k ey co n d itio n s to gain access
to foreign m arkets.
A lth ough the conceptual differences betw een these tw o m otives fo r offshore R&D
investm ent are clear, indicators o f the im p o rtan ce o f these tw o m o tives are scarce.
U ntil recently, m ost em pirical studies o f in tern atio n al R & D investm ent (Mansfield
et al. 1979; Tail 1979; W arrant 1991) reflected the v iew that the role played b y foreign
R & D units was determ ined b y m arket or d em an d -sid e facto rs, i.e., asset-exploiting
m otives were assum ed. M ore recent em pirical w ork, how ever, has focused on asset-
augm en tin g m otives fo r R & D investm ents. D etailed analyses carried o u t by Miller
(1994), O dagiri and Yasuda (1996), and F lo rid a (1997) argu e that technology
sourcing strategies play an im p o rtan t role in a n u m b er o f m a n u fa ctu rin g industries
in N orth A m erica, E urope and A sia .10 Som e studies fin d that “ m arket-oriented”
R & D units established for asset-exploiting m otives have evolved into asset-aug­
m enting ones (R ondstad t (1978 ). B u t oth er foreign R & D units experien ce no major
shift in their characters (K uem m erle 1999).
Several studies have used m ultivariate techniques to id en tify the relative import-
ance o f asset-augm entin g vs. asset-exploiting m otives fo r o ffsh o re R & D investment.
U sing patent citations A lm eida (1996) fou nd that foreign firm s in the sem iconductor
in d u stry not only learnt m ore from local sources, but they did so to a greater extent
than their dom estic counterparts. T h is stu d y also fo u n d that, w ith the significant
exception o f subsidiaries o f Japanese M N E s, foreign firm s locate their technological
activities overseas in areas where these firm s exhibited a h om e c o u n try disadvantage
 G L O B A L I Z A T I O N OF I N N O V A T I O N 329

(measured in term s o f'R e v e a le d Technological A dvantages' (R T A )). Using a sim ilar

methodology, C antw ell and N o o n an (2002) show ed that M N E subsidiaries located
in Germ any between 1975 and 1995 sourced a relatively high p roportion o f know ­
ledge (especially new, cuttin g-edge technology) from this host country.
Data such as this lend su p p o rt to the idea that foreign owned technological
activities undertaken in G erm an y are often asset-augm enting. H owever, Patel and
Vega (1 999 ) obtained different results from their study o f US patenting activities in
high technology fields. B y co m p arin g the RTA o f the M N E at hom e and the host
location, they show ed that a m a jo rity o f firm s undertook foreign innovative activ­
ities in the technological fields in w hich they were strong at hom e. T h ey interpreted
this as evidence that asset exp loitin g m otives, i.e. adapting products and processes
for foreign m arkets and p ro vid in g technical su p p o rt to offshore m anufacturing
plants, rem ained d o m in an t in M N E s' foreign innovative activities. Their findings
were supported b y an extensive interview -based survey carried out by Pearce (1999).
Employing a m eth o d o lo g y sim ilar to that o f Patel and Vega, Le Bas and Sierra (2002)
confirmed that M N E s rarely internationalize R & D to com pensate for technological
weaknesses at hom e. H ow ever, their research also show ed that the lion's share o f
these investments w ent to tech nologically advanced locations, indicating that asset
augmenting is v ery im p o rtan t and can coexist w ith asset exploiting in m any cases.
This may be interpreted as signalling the fo rm ation o f global "centers o f excellence"
in specific technological fields (see B o x 12 .1 for details on the m ethodology used to
measure alternative in tern atio n al R & D strategies).

Box 12.1 Asset exploiting, asset augmenting or both?

In one of the most extensive empirical exercises to date, Le Bas and Sierra (2002)
studied the R&D investment strategies o f the 345 MNEs with the greatest patenting
aflhdty in Europe between 1988 and 1996. These companies, which accounted for
about one half o f total patenting through the European Patent Office (EPO) over
this period, were predominantly of US, European, or Japanese origin.
To measure the technological strength o f companies and locations, the authors used
a patent-based indicator ("Relative Technolological Advantage" RTA). For a company,
HomeRTA is defined as the firm's share o f total European patents in a particular
technological field relative to its overall share o f all European patents. Patents from
"'".'foreign affiliates o f the firm (filed from outside the country in question) were excluded
v fro ip the calculation. For a location (country) in which a given firm has invested,
/.^HostRTAIs defined as the host country's share o f all European patenting in that field,
divided by its share of all European patents in all fields. In all cases an RTA> 1 signals a
>;y $efeive advantage o f the country (firm). Based on these definitions, four different R&D
identified:
330 RAJNEESH NARULA AND A N T O N E LIO ZANFEI

Corporate Technological activities in the host country

technological
activities in the Weak Strong ; ' :
home country

Weak Type 1: market-seeking Type 2: technology-seeking

HomeRTA < 1 HomeRTA < 1
HostRTA < 1 HostRTA > t
(Technology is not a driver (13%)
of FDD
(10%)

Strong Type 3: asset-exploiting Type 4: asset-augmentin§>

HomeRTA > 1 HomeRTA > 1
HostRTA < 1 HostRTA > 1
(Efficiency-oriented FDl in (Learning-oriented FDl in
Rao) R&D)
(30%) (47%)

Source: adapted from Patel and Vega (1999, p. 152) and from Le Bas and Sierra (2002 p. 606).

The numbers in brackets indicate the frequency of the strategy in question for the
sample of firms considered. As is evident from the table, Le Bas and Sierra found that
the great majority of MNEs located their activities abroad in technological areas or
fields for which they were strong at home (strategies 3 and 4). However, the most
frequent strategy is clearly number 4, in which case not only the firm but also the host
country has a relative technological advantage (HostRTA > 1 ) . This may indicate the
formation of “centers of excellence” in which strong domestic research environments
function as global attractors.

12.4 F o r c e s S u p p o r t in g C o n c e n t r a t io n

a n d D is p e r s io n o f R&D

The literature on the location o f R& D activities views the location of MNEs
innovative activities as affected by centrifugal and centripetal forces that determine
whether the MNE centralizes (in the home location) or internationalizes to create
additional centers abroad. But all too often, this dichotom y— while substantially
correct presumes that the M N E has a single center in the first place. In order to
allow for the possibility that the M N E may have multiple home bases or several
locations of R&D concentration rather than a single “ hub", this section uses the
terms “ concentration" and “ dispersion,"
 G L O B A L I Z A T I O N OF I N N O V A T I O N 331

We can single out at least fo u r b ro ad sets o f factors affecting the concentration and
dispersion o f innovative activities. These forces are active at both the m acro-level o f
countries, regions and system s o f firm s involved in the globalization o f innovation;
and the m icro-level o f in d ivid u al firm s and o f their internal networks o f innovati ve
activities across n ation al borders.

12.4.1 The Costs of Integrating Activities in Local Contexts

When firm s engage in R & D in a foreign location to avail them selves o f com ple­
mentary assets that are lo catio n -sp ecific (in cluding those that are specific to local
firms or institutions), they seek to internalize several aspects o f the system s o f
innovation o f the host lo cation . D evelop in g and m aintaining strong linkages with
external netw orks o f local counterparts is expensive and tim e consum ing. N etw orks
o f governm ent-funding institu tion s, suppliers, university professors, private re­
search teams, in fo rm al netw orks o f like-m in ded researchers take considerable
effort to create, b u t once developed, links w ith these entities or netw orks are less
costly to m aintain. Even w here the host location is potentially superior to the hom e
location, the high costs o f b eco m in g fam iliar w ith and integrating into a new
location m ay be p ro h ib itive. Firm s are constrained b y resource lim itations, and
some m inim u m threshold size o f R & D activities exists in every distinct location. As
such, m aintaining m o re than one facility w ith a ‘"critical m ass” o f researchers
requires that the new (host) lo cation offer significantly superior spillover o p p o rtu n ­
ities, or provide access to co m p lem en tary resources that are unavailable elsewhere
and cannot be acqu ired b y less costly m eans m o re efficiently.11

12.4л Local Technological Opportunities and Constraints

The high costs o f integration into a host location's system s o f innovation— in
contrast to the low m argin al cost o f m ain tain in g its em beddedness in its hom e
locations in n o vation system — m ay increase the fixed costs firm s have to overcom e
in order to exp an d in tern atio n ally (N arula 2003). H owever, these costs m ust be
tempered by other su p p ly-sid e considerations. Fo r exam ple, developm ent o f the
technologies in qu estion m ay benefit fro m diversity and heterogeneity in the
knowledge base, w hich m igh t com e fro m com petitors, from interaction w ith cu s­
tomers and fro m oth er co m p lem en tary technologies in the offshore site. A single
national in n o vation system , especially in a sm all country, m ay be unable to offer the
full range o f interrelated tech nological assets required for this diversification strat-
egy (see B o x 12.2 o n the in teraction s between in n o vation system s and R & D inter­
nationalization strategies).
332 RAJN EE5H N ARULA AND A N T O N E L L O Z A N FE I

Box 12.2 How innovation systems affect the internationalization of І Ш Э Д

.... •........ . — .. ..i: ...................... ....... .1 .... ...... '' ."lni" ........................ . ■■

Innovation systems are built upon a relationship of trust* iteration* and імешгійэд
between firms and the knowledge infrastructure* within the framework of institution^
based on experience of and familiarity with each other over relatively long periods of
time. In engaging in foreign operations in new locations* firms which already face
opportunities and constraints created by their home innovations systems gradually
become embedded in the host environment The self-reinforcing interaction between
firms and infrastructure perpetuates the use o f a specific technology or technologies* or
production of specific products* and/or through specific processes. Increased ffed#*
ization often results in a systemic lock-in. Institutions develop that support шй
reinforce the interwoven relationship between firms and the knowledge infrastrucfore
through positive feedback* resulting in positive lock-in. When innovation systems
cannot respond to a technological discontinuity, or a radical innovation that has
occurred elsewhere, there is a mismatch between what home locations can provide
and what firms require, this is known as sub-optimal lock-in (Narnia 2002$).
In general, national innovation systems and industrial and technological specializa­
tion of countries change only very gradually, and— especially in newer, rapidly evolv­
ing sectors— much more slowly than the technological needs of firms. In other words,
there may be systemic inertia. Firms have three options open to them (Narnia 2002a).
Firms may seek either to import and acquire the technology they need from abroad, or
venture abroad and seek to internalize aspects of other countries" innovation systems,
thereby utilizing an “exit” strategy. Of course, firms rarely exit completely, preferring
often to maintain both domestic and foreign presence simultaneously. There are costs
associated with an exit strategy. On the one hand, they would weaken their contact
with their home market and by so doing they might reduce their ability to absorb
external knowledge. On the other hand, it must suffer the costs of entry in another
location (in terms of effort, capital, and time), and firms may minimize this through a
cooperative strategy with a local firm. Developing alternative linkages and becoming
embedded in a non-domestic innovation system takes considerable time and effort.
They can also use a “voice” strategy which is to seek to modify the home-country
innovation system. For instance, establishing a collective R&D facility, or by political
lobbying. Firms are inclined towards voice strategies, because it may have lower costs,
especially where demand forces are not powerful, or where the weakness of the
innovation system is only a small part of their overall portfolio. But voice strategies
have costs, and may not be realistic for SMEs, which have limited resources and
political clout. Such firms usually cannot afford an “ exit” strategy either, and end up
utilizing a loyalty strategy, relying instead on institutions to evolve, or seeking to
free-ride on the voice strategy of industry collectives, or larger firms.

Where local technological opportunities are sufficiently high, asset-augmenting

activities are likely Capturing foreign opportunities may require that a firm develop
proximity to local technology leaders (see Ch. 20 by Cantwell, this volume) whose
competences are rooted in the offshore system o f innovation.12 Whenever products
 G L O B A L I Z A T I O N OF I N N O V A T I O N

are m ulti-technology-based, one firm m ay be m arginally ahead in one technology,

and its com petitor in another; but on a m acro-level, both m ay be associated with
“powerful” in n o vation system s (Criscuolo et a l 2005). Thus, technology leadership
can change rapidly. T h is is another reason w hy firm s often engage in both asset-
augmenting and asset-exploiting activities sim ultaneously.13

12*4-3 Firm Size and Market Structure

An im portant factor affecting internationalization is the size o f the firm . The expan­
sion o f R & D activities— both at hom e and in overseas locations— requires consider­
able resources o f capital and m anagem ent expertise that sm aller firm s often lack.
Ceteris paribusy large firm s have m ore m oney and resources to use in overseas
activity. As they have higher R & D budgets at hom e, they are also m ore likely to
have the absorptive capacity to set up linkages w ith both foreign and dom estic
science bases. R & D is a costly and slow affair, and overseas R & D facilities are an
expensive and risky o p tio n that is hard to ju stify fo r SM Es. Indeed, Belderbos (2001)
finds that there is a n o n -lin ear relationship between firm size and overseas R & D ,
with m edium -sized Japanese firm s show ing a h igher propensity (in relative term s)
to internationalize R & D than sm all- or large-sized firm s. M an y sm all firm s operate
as part o f a d om estic su p plier netw ork for larger firm s, and are thus also boun d to
their hom e location (o r the location o f their m ain custom ers) (N arula 20 0 2 b).
Internationalization o f su p p lier firm s often occurs in tandem w ith the international­
ization o f their p rim a ry custom er, especially where the custom er dom inates their
market. T h is m o tive w as apparent in the investm ent b y Japanese autom obile firm s'
supplier firm s in U S and E u rop ean p ro d u ction facilities d u rin g the 1980s and early
1990s (Florida 1997).
In dustry-specific factors also encourage or discourage the locational concentra­
tion o f innovative activities. The industrial structure o f countries is path dependent,
and technological specialization changes o n ly gradually over tim e (C antw ell 1989;
Zander 1995). At one extrem e, m ature technologies evolve slow ly and dem onstrate
m inor but consistent in n o vation s over tim e. The tech nology is to a great extent
codifiable, w id ely dissem inated, and the p ro p erty rights well defined. U nder these
circum stances, constant and close interaction w ith custom ers is not an im portant
determ inant o f R & D : p rofits o f firm s depend on the costs o f inputs, and pro xim ity to
the source o f these in pu ts is often m ore significant than that o f custom ers. At the
other extrem e, rap id tech nological change in “ n ew er" technologies or engineering
industries m ay requ ire closer interaction between p ro d u ction and R & D (Lall 1979)?
o r between users and produ cers o f technology. In som e circum stances both new
technology and ap p lication s environ m en ts have a high tacit, uncodified elem ent,
requiring extensive interaction d u rin g new produ ct developm ent, design, and
334 RAJNEESH N A R U LA AND A N T O M E LLO Z A N FEI

testing T h is factor m ay account for the frecjuent estab lish m en t o f b oth manufactur­
ing and R & D plants close to app lications abilities in fo reig n telecommunications
m arkets (Ernst 1997)- lit other industries, how ever, a large v a rie ty o f international
linkages are required for R & D and in n o vatio n , as ap p ears to be particu larly the case
o f biotechnology (Arora and GambardeUa 1990).

Organizational Issues
Another micro-level determinant is associated with the difficulties of managing
cross-border R&D activities. It is not sufficient for the foreign affiliate to internalize
spillovers if it cannot make these available to the rest o f the MNE (Blanc and Sierra
1999). A dispersion o f R&D activities across the globe requires extensive coordin­
ation between them, and particularly with headquarters, if they are to function
efficiently. This acts as a centripetal force on R&D, and accounts for a tendency of
firm s to locate R & D (or at least the m ost strategically sig n ifican t elem ents) closer to
headquarters.
Com plex linkages, both within the firm, and between external networks and
internal netw orks, require co m p lex co o rd in a tio n i f th ey are to provide optimal
benefits (Zan fei 2000). Such co ord in atio n m a y req u ire exp ertise, m anagerial and
financial resources that are m ost likely available to larger firm s w ith m ore experience
in transnational activity (Castellani and Z a n fe i 2004), Large firm s tend to engage in
both asset-augm enting and asset-exploitin g activities. In d eed , large M N E s may have
several sem i-au ton om o u s sister affiliates in the sam e lo catio n that operate in similar
technological areas. Lastly, M N E s tend to engage in p ro d u c tio n activities (whether
in the sam e or another physical facility) in the h ost lo catio n , an d this prompts a
certain level o f asset-exploiting activity. T h u s, an M N E in a given location may seek
to internalize spillovers from n on-related firm s and to e xp lo it in trafirm knowledge
transfers w ith in the sam e m u ltin atio n al gro u p (C risc u o lo et al. 2005).

12.5 In n o v a t io n t h r o u g h In t e r n a t io n a l

S t r a t e g ic T e c h n o l o g y P a r t n e r in g

The previous sections have discussed the g ro w in g in tern atio n al d im en sio n o f R&D,
concentrating on the in tra -M N E aspect o f this d evelo p m en t. H ow ever, not all
in n o vato ry activity is undertaken w ithin hierarchies; d u rin g the last two decades,
 non-internal R & D activities that rely on interfirm cooperative agreem ents have
grown rapidly in num ber.
Fully exam in in g the role o f (international) netw orks m the generation and
diffusion o f in n o vatio n is beyon d the scope o f this chapter (see Powell and G rodal,
Ch. 3 in this volu m e, fo r a m o re com prehensive discussion). A key issue for this
discussion is w h eth er and to w hat extent there is substitution or com plem entarity
between internal in n ovative activities and technological collaborations on a global
scale,
In som e circum stan ces, international S T P m ay substitute for internal innovative
activities. O ne such circu m stan ce is that o f R & D alliances aim ing to enter foreign
markets protected b y n o n -ta riff barriers, as is the case for environm ental regulations
in the chem ical industry. N onetheless, there are lim its to how m uch a firm can
substitute S T P fo r in -h o u se R & D , and by extension, international STP for overseas
R& D facilities.14 S T P tends to develop in areas in w hich partner com panies share
com plem entary capabilities, and these alliances create a greater degree o f interaction
between the p artn ers' respective paths o f learning and innovation (M ow ery et al.
1998; C antw ell and C o lo m b o 2000; Santangelo 2000).
One w ay to lo o k at this issue is to tackle the problem o f firm size, technological
capabilities, and co llab o ration s. P articipation in STPs tends to be correlated w ith
firm size in tech n o logy-in ten sive sectors. In these sectors, cooperation is a w ay to
keep up with the tech n o logical frontier: by associating com plem entary resources and
com petencies, it m akes it possible to explore and exploit new technological o p p o r­
tunities. But sm aller tech n o logy-b ased M N E s also are involved in such agreem ents,
and their grow in g significance raises n u m erou s conundrum s (N arula 2002b; see also
Ch. 5 b y Lam , an d C h . 3 b y Pow ell and G rod al, this volu m e). Firm s— regardless o f
size— m ust m ain tain a g ro w in g breadth o f technological com petences, and this m ay
require p articip atio n in in tern ation al internal and external networks. SM Es need to
rely on n o n -in tern al sources, as they often experience w ider gaps in term s o f
com petencies and d evelo pm en t abilities than their larger counterparts (Zanfei
1994) but m ust be m o re skilfu l at m anaging their po rtfolio o f technological assets,
because they have lim ited resources (N aru la 2002b). Indeed, the costs o f m anaging a
web o f cross-b ord er agreem ents h igh ligh t the im portance o f transaction-type o w n ­
ership advantages fo r the M N E . T h is com plem en tarity between firm size, tech no­
logical capabilities and the d evelopm ent o f innovation netw orks is consistent w ith
some o f the trends h igh ligh ted in Section 12,2.2. In particular, the geographical
concentration o f S T P activ ity w ith in the Triad reflects inter alia the fact that firm s
from these areas ten d to be larger and account for a m ajor share o f R & D activity.
The issue o f complementarity or substitution between the internal and non­
internal innovative activities of MNEs can also be examined by looking at the
interdependencies between multinational expansion and international STP. Draw­
ing on the transaction-cost literature, several works on international market entry
strategies argue that multinational experience may lower the risks faced by an MNE
336 RAJNEESH NARULA AND ANTONELLO 2 A N FE I

in entering a new foreign mftrket. In the ftbsence o f rnultinfttionftl experience,

cooperative ventures m ay be m ore effective m arket e n try too ls th an hierarchical
con trol strategies. As M N E s accum ulate greater exp erien ce in fo reign markets,
the in fo rm atio n -gath erin g and risk -sh arin g advantages o f co llab o ra tio n w ill decline.
A s a result, the organizational costs o f co o p eratio n , in term s o f shirking and
conflicts o f interest betw een partners, w ill exceed the b en efits o f this strategy
fo r experienced M N E s (e.g. G o m ez-C asseres 1989; H en n art an d L arim o 1998).
In summary, m u ltin atio n al experience is su p posed to im p act n egatively on collab­
o rative ventures and p o sitively on equity-based, co m m itm en t-in ten sive linkages.
This view is largely— but not exclusively— consistent with the argument that multi-
n ation al experience helps facilitate the exp lo itatio n o f M N E s assets in foreign
m arkets. T h at is, M N E s resp on d to u n certain ty in h ost eco n o m ies by utilizing
their o w n assets as a m eans to penetrate these m arkets. Su ch a v ie w regards STP as
a second-best option.
A second b o d y o f literature focuses m ain ly on the e vo lu tio n o f h igh technology
industries, and h ighlights an im p o rtan t m o tive fo r in terfirm linkages, i.e. the need to
explore and rap id ly exp loit new o p p ortu n ities, either n ew b u sinesses o r new tech­
n ological developm ents. From this perspective, strategic allian ces provide “ an
attractive o rgan isation al fo rm fo r an en viron m en t ch aracterised b y rap id innovation
and geographical d ispersion in the sources o f k n o w h o w ” (Teece 1992: 20). As the
relevant know ledge sources are dispersed globally in a n u m b er o f industries, this
perspective explain s the fo rm atio n o f som e types o f in tern atio n al S T P agreements.
From this perspective, m u ltin ation al experience— w h ich is associated w ith the
establishm ent and activity o f foreign subsidiaries over tim e— m a y increase a firm s
cap acity to search fo r and absorb external know ledge (C an tw ell 1995; Castellani and
Z an fei 2004). T h is view is consistent w ith a n u m b er o f stu d ies o n h igh technology
industries w h ich h igh light the m u tu ally rein forcin g nature o f in tra- an d interfirm
netw orks. M u ltin atio n al experience thus m ay exp an d a firm ’s exp lo ra tio n potential
and hence exp an d its use o f international ST P .15
Som e o f the trends in the developm ent o f S T P h igh ligh ted in section 12.2.2 seem to
be consistent w ith the vie w that firm s w ith m u ltin atio n al exp erien ce are m ore likely
to use alliances as an e xp lo rato ry strategy. A s we have sh ow n , the fractio n o f non
equity S FPs is grow ing, p articu larly in high tech n o lo gy in du stries. T h is trend may
constitute evidence o f the fact that lo w co m m itm en t inten sive agreem ents are more
effective as a m echanism to gain tim ely and extensive access to rap id ly evolving
tech n o logy across borders. From this perspective, S T P m a y represent a “ first-best”
o p tio n to M N E s (N arula 2003), especially w here in n o vative activities are concerned.
In other w ords, firm s do not n ecessarily resort to these strategies b ecau se they cannot
have access to m ore effective and m ore profitable channels o fte c h n o lo g y transfer (as
uncertainty is too high or institutional b arriers con strain “ in tern a l” strategies).
Instead, STPs, especially non equ ity agreem ents, are m ore flexible an d m ore apt
for know ledge developm ent and learning.
 12.6 C o n c l u s i o n s a n d P o l i c y I s s u e s

This chapter has discussed the internationalization of innovative activities, and

highlighted that it has been driven by a myriad of factors. One of the most recurrent
among these factors is the need to respond to different demand and market condi­
tions across locations, and the need for the firms to respond effectively to these by
adapting their existing product and process technologies through foreign-located
RMX
Nevertheless, supply factors and the need to gain access to local competencies
have become an increasingly important motivation to engage in asset-augmenting
R&D abroad. This is due, in te r alia> to the growing tendency for multi-technology
products, and to the fact that patterns of technological specialization are distinct
across countries, despite the economic and technological convergence associated
with economic globalization.
As a result, there is a growing mismatch between what home locations can provide
and what firms require. In general, innovation systems and the industrial and
technological specialization of countries change only very gradually, and— espe­
cially in newer, rapidly evolving sectors— much more slowly than the technological
needs of firms. Firms must seek either to import and acquire the technology they
need from abroad, or venture abroad and seek to internalize aspects of other
countries' innovation systems. A third option, lobbying for modification of the
home-country innovation system, is expensive and difficult (Narula 2002a). Thus,
in addition to proximity to markets and production units, firms venture abroad to
seek new sources of knowledge, which are associated with the innovation system of
the host region. The interdependence of markets and the cross-fertilization of
technologies— whether through arms-length means, cooperative agreements, or
equity based affiliates— means that that few countries have truly “ national” systems.
Of course, some innovation systems are more “ national” than others, and the term is
indicative rather than definitive (see also Ch. 7 by Edquist and Ch. 14 by Malerba in
this volume for a discussion). Furthermore, firms need a broader portfolio of
technological competences than they did in the past.
The internationalization o f R & D raises crucial welfare issues, since it provides
opp ortu n ities fo r spillovers betw een the M N E and its host econom y, and in certain
circum stances betw een the M N E affiliate and its h om e country. There has been som e
concern in the U S w ith the potential loss o f com petitiveness o f dom estic firm s and
w ith the Im poverish m ent o f the “ national know ledge base” w h ich w ou ld be associ­
ated w ith the increasing local R & D presence o f foreign-ow n ed M N E s (e.g., D alton
et al. 1999). In oth er countries and areas o f the w orld, the perception is very different,
as a local presence o f foreign R & D and value-added activities is expected to co n trib ­
ute to the up grad in g o f n ation al technology system s. A few em pirical studies seem to
provide so u n d evidence on the existence o f positive spillovers o f m u ltinational
33$ RAJNEESH NARULA AND ANTONELLO ZANFEI

presence in som e em erging econom ies such as K orea, T aiw an, And Singapore
(H o b d a y 2000; L im 1999)? and som e EU m em b er states (B a rry and Strobl 2002;
C astellani and Z an fei 2003). H ow ever the evidence in the case o f m o st developing
countries does not p o in t to significant spillovers (see H arriso n 1999)* Indeed,
accord in g to a recent su rvey on econ om etric studies o f p ro d u c tiv ity spillovers
fro m F D I, the n u m ber o f studies in w h ich negative o r n o n -sig n ifican t results are
obtained is ap p ro xim ately as h igh as cases w here p o sitive sp illo vers w ere observed
(G o rg and Strobl 2001). T h is suggests a cau tiou s ap p ro ach to this issue, and calls for a
refinem ent o f analytical tools (see B o x 12.3). T h ere is a need to develop more
app ro p riate m easures o f technological spillovers, w h ich are n o t p ro p e rly captured
b y p erfo rm an ce in dicators like pro d u ctivity. T h e channels th ro u g h w h ich spillovers
occu r also need to be exam ined m o re carefully, i f F D I-related sp illovers are to be
exp licitly used as m eans for tech nological u p grad in g.
A n o th er p o sitio n in this p o licy debate argues that the in tern atio n alizatio n o f R& D
m ay lead to a “ h o llo w in g o u t” o f the h om e co u n try's in n o v a to ry cap acity when the
dom estic in n o vation system does not m eet the needs o f firm s in certain industries.

Box 12.3 Host country effects: technology gaps, technological upgrading,

and absorptive capacity

One of the strongest and most popular arguments in favor of inward investment as a
vehicle for local technological upgrading is that foreign firms usually outperform
domestic ones (see Beliak 2 0 0 2 for a review on empirical evidence on this aspect).
The underlying policy issue is whether or not foreign presence can generate techno­
logical opportunities for the local economy. There is a dear connection here to the
literature on technology gaps and catching up (Godinho and Fagerberg, Ch. 19 in this
volume). On the one hand, some works suggest that the larger the productivity gap
between host country firms and foreign-owned firms, the larger the potential for
technology transfer and for productivity spillovers to the former. This assumption,
can be derived from the original idea put forward by Findlay (1978), who formalized
technological progress in relatively “backward” regions as an increasing function of the
distance between their own level of technology and that of the “advanced regions,” and
of the degree to which they are open to direct foreign investment.
On the other hand, scholars have argued that the lower the technological gap
between domestic and foreign firms, the higher the absorptive capacity of the former,
and thus the higher the expected benefits in terms of technology transfer to domestic
firms. It is worth noting that the role of absorptive capacity is also implicitly recognized
in the catching up tradition, when it is acknowledged that a sort of lower bound of local
technological capabilities exists, below which foreign investment cannot be expected to
have any positive effects on host economies.16 The “ technological accumulation
hypothesis goes beyond this simplistic view of absorptive capacity and places a new
emphasis on the ability to absorb and utilize foreign technology as a necessary condi­
tion for spillovers to take place.
 G L O B A L I Z A T I O N OF I N N O V A T I O N 339

Although there is currently little evidence to support or refute the hollow ing out
hypothesis, this has been raised by po licy m akers in several countries, and represents
an im portant area for future research. The consequences o f a potential hollow ing out
m ay be especially significant in sm all open econom ies that are specialized around a
few products, and/or concentrated around a few large firm s. A noth er related and
potentially im p o rtan t area for future research is the need to distinguish between
hollow ing-out as a sym p tom o f su b -op tim al lock-in and the internationalization o f
innovation to supplem ent dom estic supply lim itations (N arula 2003). A fter all, no
country can p ro vid e w orld-class com petences in all technological fields. Even the
largest, m ost tech nologically advanced countries cannot provide strong innovation
systems to all their industries, and w orld-class com petences in all technological
fields. Som e coun tries regard im ported technologies as a sign o f national weakness,
and have sought to m ain tain and develop in -co u n try com petences, often regardless
o f the cost (N arula 2002a). R elyin g largely on in -co u n try com petences m ay how ever
lead to a sub -op tim al strategy, especially in this age o f m u lti-tech n ology products. In
fact, the cross-b ord er flo w o f ideas is fundam ental to firm s, and this im perative has
increased w ith grow in g cross-b ord er com petition, and international production.

N otes

1. Both changes in the composition o f world trade, and sectoral correlations between
R8cD intensity and internationalization should be considered with caution since
definitions o f industries change over time (see Von Tunzelman and Acha, Ch. 15 in this
volume).
2. For instance, press releases are often used to construct data-sets, and these are not always
factual, sometimes reflecting the public relations objectives o f the firms; the coverage of
large firms is higher than for smaller firms; STP failures are not reported as accurately (or
as often) as STP formation; large databases are hard to update and are frequently subject
to changes in the methodology o f data collection over time.
3. STP refers to interfirm cooperative agreements where R8cD is at least part o f the collabora­
tive effort, and which are intended to affect the long-term product-market positioning of
at least one partner.
4. Strictly speaking, the two numbers are not comparable, because GDP is a flow figure.
Nonetheless, it is generally accepted that FDI stock is a monotonic function o f value
added, so the change in this ratio gives a general idea of how the significance o f FDI
activities has changed.
5. Paradoxically, perhaps, this group also includes Swedish MNEs, whose much higher
shares o f offshore R&D and patenting throughout the twentieth century, nevertheless,
display a sharp drop after 1940 and a recovery by 1969—90 to a share that is lower than that
o f 1920-39.
340 RAJ N E E S H N A R U L A AND A N T O N E L L O Z A N F E I

6. Although the degree o f R&D internationalization o f US firms is below average, it more

than doubled between the mid-1960s and the end o f the 1980s (Creamer 1976; Pearce
1990)*
7. Even where MNEs do engage in R&D in developing countries (e.g. industries where
demand considerations and regional variations are especially significant, such as food
products and consumer goods), these tend to agglomerate in just a few locations such as
China, India, Malaysia, Brazil, South Africa, and the Asian NICs.
8. Although “ home-base exploiting” (HBE), and “ home-base augmenting” (HBA) (which
we define later) have become dominant in the literature, this terminology is less accurate
than “ asset exploiting” and “ asset augmenting”. H BA and H BE hold to a very traditional
view o f the M NE as centered in a dominant home base. In fact, by emphasizing the role of
home bases, the H BA -H BE jargon cannot be easily made consistent with the possibility
that firms are evolving towards network structures, hence reducing the importance of a
single home and, by the same token, expanding the number o f countries wherein the
firm ends up being based. This chapter takes the view that being accurate is more
important than being fashionable, and avoids using the H B E -H B A terminology except
where necessary for historic accuracy.
9. Fosfuri and M otta (1999) and Siotis (1999) show that a tech n ological laggard m ay choose
to enter a foreign m arket by FD I because there are po sitive spillover effects associated
with locational p ro xim ity to a technological leader in the foreign country. Where the
beneficial know ledge spillover effect is sufficiently strong, Fo sfu ri an d M otta show that it
m ay even pay the laggard firm to run its foreign su b sid iary at a loss to incorporate the
benefits o f advanced technology in all the m arkets in w h ich it operates.
10. Miller (1994: 37) studied the factors affecting the location o f R& D facilities o f twenty
automobile firms in North America, Europe, and Asia, and found that an important
motivation is to establish “ surveillance outposts” to follow competitors' engineering and
styling activities. In their study o f 254 Japanese manufacturing firms, Odagiri and Yasuda
(1996:1074) note that R&D units are often set up in Europe and in the US to be kept
informed o f the latest technological developments. Similar results are obtained by
Florida (1997: 90) analyzing 186 foreign affiliated laboratories in the US.
n. With few exceptions (e.g. Narula 2002a), the costs and inertia o f offshore R&D networks
is a topic which has not as yet been properly explored and represents an important area
for further research.
12. Technology leaders are not always synonymous with industry leaders: firms— particu­
larly in technology intensive sectors— increasingly need to have multiple technological
competences (see e.g. Granstrand 1998; Granstrand et al. 1997).
13. This is another area w hich has not as yet been fully studied (for an exception, see Zander
1999) and represents an im portant area for further research.
14. The attem pt to understand the reasons behind a firm 's choice between non-internal and
internal technological developm ent is not new. The w ork o f Teece (1986) presents a
pioneering analysis o f this issue, w hich builds on A bernathy and U tterback (1978), Dosi
(1982) am ong others. See also further developm ents by P isan o (1990), H enderson and
C lark (1990), N agarajan and M itchell (1998), Veugelers and C assim an (1999), Gambar-
delia and Torrisi (1998), N ooteboom (1999), N arula (2001) and B ru so n i et al. (2001).
15* Castellani and Z anfei (2004) have tried to provide som e em p irical basis to this view with
reference to the electronics industry. T hey m easure w hat they call “ specific experience”
in term s o f the num ber o f subsidiaries a M N E has established in a given country, which in
 G L O B A L I Z A T I O N OF I N N O V A T I O N 341

their view w ould reduce uncertainty about the foreign market. Controlling for a num ber
o f sources o f heterogeneity, they show that this factor is positively correlated with the
creation o f new subsidiaries and o f equity agreements. By contrast, what they call
“variety experience," reflecting the heterogeneity and geographical dispersion o f markets
where a M N E is active, should increase the firm's exploratory capacity. They find that, in
the exam ined industry, variety experience has a positive and significant im pact on non
equity technical alliances.
16. As Findlay (1978:2-3) notes: “ Stone age communities suddenly confronted with modern
industrial civilisation can only disintegrate or produce irrational responses. . . Where the
difference is less than some critical minimum, admittedly difficult to define operation-
ally, the hypothesis does seem attractive and worth consideration." Findlay also observes
that the educational level o f the domestic labour force, which is a good proxy for what is
currently named country s “ absorptive capacity," might also affect, inter alia, the rate at
which the backward region improves its technological efficiency (Findlay 1978: 5-6).

R eferen ces*

A ber n a th y , W., and U t t e r b a c k , J. (1978), “ Patterns o f Industrial Innovation," Technology

Review 80: 97-107.
A lm eid a , P. (1996), “ Knowledge Sourcing by Foreign Multinationals: Patent Citation An­
alysis in the Semiconductor Industry," Strategic Management Journal 17:155-65.
A rc h ibu g i , D., and Ia m m a rin o , S. (2002), “ The Globalisation o f Technological Innovation:
Definition and Evidence," Review of International Political Economy 9(1): 98-122.
----------— (2000), “ Innovation and Globalisation: Evidence and Implications," in F. Ches-
nais, G. letto-Gillies, and R. Simonetti (eds.), European Integration and Global Corporate
Strategies, London: Routledge, 95-120.
*------ and M ic h ie , J. (1995), “ The Globalisation o f Technology: A New Taxonomy," Cam­
bridge Journal of Economics 19 :121-4 0.
A rora , A., and G a m b a r d ella , A. (1990), “ Complementarity and External Linkages: the
Strategies o f the Large Firms in Biotechnology," The Journal of Industrial Economics 38(4):
361-79.
A u d retsch , D., and F eld m a n , M. (1996), “ R&D Spillovers and the Geography o f Innov­
ation and Production," American Economic Review 86: 253-73.
Ba r r y , E, and S tro bl , E. (2002), “ FDI and the Changing International Structure o f
Employment in the EU Periphery," CEPR-LdA workshop on Labour Market Effects of
European Foreign Investments, Turin, 10 -11 May.
B eld erbo s , R. (2001), “ Overseas Innovations by Japanese Firms: An Analysis o f Patent and
Subsidiary Data," Research Policy>30(2): 313-32.
B ella k , C. (2002), “ How Performance Gaps between Domestic and Foreign Firms Matter
for Policy,” ELBA Annual Conference, Athens, December.
B en ito , G., G ro gaard , B., and N a r u l a , R. (2003), “ Environmental Influences on MNE
Subsidiary Roles: Economic Integration and the Nordic Countries," Journal of Inter­
national Business Studies 34: 443-56.

* Asterisked items are suggestions for further reading.

342 R A J N E E S H N A R U L A AND A N T O N E L L O Z A N F E I

Blan c , H „ and S ierra , C. (1999). “ The Internationalisation o f R&D by Multinationals:

A Trade-Off between External and Internal Proximity,” Cambridge Journal o f Economics гу
l 87 - 2 0 6 .
B rusoni, S., P r e n c ip e , A., and P a v i t t , K. (2001), “ Knowledge Specialization and the
Boundaries o f the Firm: Why do Firms Know More Than They Do?” Administrative
S cien ce Q u a rte rly 46: 597-621.
B u r e t h , A., W o l f f , S., and Z a n f e i , A. (i999)> Cooperative Learning and the Evolution of
European Electronics Industry,” in A. Gambardella and F. Malerba (eds.), The O rganisa­
tion of Inventive Activities in Europe, Cambridge: Cambridge University Press, 202-38.
* C a n t w e l l , J. (1989), Technological In n o v a tio n a n d M u lt in a t io n a l C orporations^ Oxford:
Basil Blackwell.
------ (1995), "The Gobalisation o f Technology: What Remains o f the Product Cycle Model,”
Cambridge Journal of Economics, 19,155-174.
------ and C olombo , M. G. (2000), “ Technological and Output Complementarities, and
Inter-firm Cooperation in Information Technology Ventures,” Journal of Management
a n d G o v e rn a n c e 4 :117-4 7.
------ and J a n n e , O. (2000), “ The Role o f Multinational Corporations and National States in
the Globalisation o f Innovatory Capacity: The European Perspective,” Technology Analysis
a n d S trateg ic M a n a g e m e n t 12(2): 243-62.
------ and N oonan , C. A. (2002), “ Technology Sourcing by Foreign-Owned MNEs in Ger­
many: An Analysis Using Patent Citations” El BA Annual Conference, Athens, December.
C a s t e l l a n i , D., and Z a n f e i , A. (2003), “ Technology Gaps, Absorptive Capacity and the
Impact o f Inward Investments on Productivity o f European firms,” Economics of Innov­
ation and New Technology 12: 555-76.
------------- (2004), “ Choosing International Linkage Strategies in Electronics Industry: The
Role o f Multinational Experience,” Jo u r n a l o f E c o n o m ic B e h a v io u r a n d O rgan isa tion 53:
447-75.
C r e a m e r , D. (1976), O versea s R esearch a n d D e v e lo p m e n t b y U n ite d S ta tes M u ltin atio n a ls
1966-1975, New York: The Conference Board Inc.
C riscuolo , R, N a r u la , R., and V e r sp a g e n , B, (2005), “ The Relative Importance of Home
and Host Innovation Systems in the Internationalisation o f M N E R8cD: A Patent Citation
Analysis,” Economics of Innovation and New Technologies (forthcoming).
D alto n , D., S e r a pio , M., and Y o sh ida , P. (1999), G lo b a liz in g I n d u s t r ia l R&DyU.S. Depart­
ment o f Commerce, Technology Administration, Office o f Technology Policy.
D osi, G. (1982), “ Technological Paradigms and Technological Trajectories: A Suggested
Interpretation of the Determinants and Directions of Technical Change,” R esearch Policy
1 1 : 147-62.
D u n n in g , J. H. (1994), “ Multinational Enterprises and the Globalization o f Innovatory
Capacity,” R esearch P o lic y i y 67-88.
* an<^ N a r u l a , R. (1995), The R8cD Activities o f Foreign Firms in the United States,”
In te rn a tio n a l S tu d ies o f M a n a g e m e n t & O rg a n iz a tio n 25(1—2): 39—73.
E r n s t , D. (1997)> From Partial to Systemic Globalisation: International Production Net­
works in the Electronics Industry, B R I E W orkin g P a p e r 98, Berkeley Roundtable on the
International Economy, University of California at Berkeley.
E- IAN (1998), Technology P o lic y in the C o n te x t of In te rn a tio n a lis a tio n o f R & D a n d Innovation.
How to Stren g th en E u ro p e s C o m p e titiv e A d v a n ta g e in Technology. Brussels: European
Commission, Directorate-General Science, Research and Development.
 G L O B A L I Z A T I O N OF I N N O V A T I O N 343

F in d l a y , R. (1978), “ Relative Backwardness, Direct Foreign Investment and the Transfer o f

Technology: A Simple Dynamic M odel;' Quarterly Journal of Economics 92:1-16.
F l o r id a , R. (1997), "The Globalisation o f R&D: Results of a Survey o f Foreign-Affiliated
R M ) Laboratories in the USA ” Research Policy 26: 85-103.
Fo s f u r i , A., and М о т т а , M . (1999), “ M ultinationals w ithout Advantages,” Scandinavian
Journal of Economics 101(4): 6 17-30.
G a m b a r d e l l a , A., and T o r r is i , S. (1998), "Does Technological Convergence Imply Con­
vergence in Markets? Evidence from the Electronics Industry,” Research Policy 27: 445-63.
G o m e z -C a s s e r e s , B. (1989), “ Ownership Structures of Foreign Subsidiaries,” Journal of
Economic Behavior and Organization 11:1-2 5 .
G org H., and S t r o b l , E. (2001), “ Multinational Companies and Productivity Spillovers:
A Meta-analysis,” The Economic Journal 11 (Nov): F723-F739.
G r a n s t r a n d , O. (1998), “ Towards a T h eo ry o f the Technology Based Firm ” Research Policy
27: 465-90.
------ P a t e l , R, and P a v i t t , K. (1997), “ Multi-Technology Corporations: Why they have
"Distributed' rather than "Distinctive Core' Competencies,” California Management
Review 39(4): 8-25.
H a g e d o o r n , J, (2002), “ Inter-Firm R&D Partnerships: An Overview of Patterns and Trends
since i960” Research Policy31: 477-92.
H a n n a n , M., and F r e e m a n , J. (1984), “ Structural Inertia and Organisational Change,”
American Sociological Review 49:149-64.
* H e d l u n d , G. (1986), ""The Hypermodern M NC— a Heterarchy,” Human Resource Man­
agement 25: 9-35.
H e n d e r s o n , R., and C l a r k , K. (1990), ""Architectural Innovation: The Reconfiguration of
Existing Product Technologies and the Failure o f Established Firms,” Administrative
Sciences Quarterly 35: 9-30,
H e n n a r t , J. F., and L a r im o , J. (1998), ""The Impact o f Culture on Strategy of Multinational
Enterprises: Does National Origin Affect Ownership Decisions?” Journal of International
Business Studies 29(3): 515-38.
H o b d a y , M. M. (2000), ‘"East vs. South East Asian Innovation Systems: Comparing OEM-
and MNE-led Growth in Electronics ” in L. Kim and R. Nelson (eds.), Technology Learning
and Innovation., Cambridge: Cambridge University Press, 129-69.
I e t t o -G i l l i e s , G. (2001), Transnational Corporations: Fragmentation amidst Integration,
London: Routledge.
K in d l e b e r g e r , C. P. (1969), American Business Abroad. Six Lectures on Direct Investment,
New Haven: Yale University Press.
K o g u t , B. (1989), “A Note on Global Strategies,” Strategic Management Journal 10: 383-9*
K o k k o , A. (1994), “ Technology, Market Characteristics and Spillovers,” Journal of Develop­
ment Economics 43(2): 279-93.
K u e m m e r l e , W. (1996), “ Home Base and Foreign Direct Investment in R & D ” Unpublished
Ph.D. dissertation, Boston: Harvard Business School.
------ (1999), “ Foreign Direct Investment in Industrial Research in the Pharmaceutical and
Electronic Industries: Results from a Survey o f Multinational Firms ” Research Policy 28:
U 9- 93*
K u m a r , N. (1998), Globalization, Foreign Direct Investment, and Technology Transfen
London: Routledge.
 R A JN EESH NARULA AND A N T O N E L L O Z AN FEI
344

L a l l , S. (1979), “ The International Allocation o f Research Activity by U.S. Multinationals”

Oxford Bulletin of Economics and Statistics 41: З^З-З1*
L e B a $, C.> and S i e r r a , C. (2002), Location versus C ountry Advantages in R&D Activities;
Some Further Results on Multinationals' Locational Strategies,” Research Policy 31;
589-609.
L im , Y. (1999), Technology and Productivity: The Korean Way of Learning and Catching Up,
Cambridge, Mass.: M IT Press.
M a n s f f e l d , E „ T e e c e , D., and R o m e o , A. (1979), “ Overseas Research and Development by
US-Based Firms,” Economica 46 (May): 1 87-96.
M i l l e r , R, (1994), “ Global R&D Networks and Large Scale Innovations: The Case of
Automobile Industry,” Research Policy 23(1): 27-46.
M it c h e l l , W., and S in g h , K. (1992), “ ‘Incumbents’ Use o f Pre-Entry Alliances before
Expansion into New Technical Subfields of an Industry,” Jo u rn a l o f Econom ic Behaviour
and Organisation 18:347-72.
M o w e r y , D. C., O x l e y , J. £., and S il v e r m a n , B. S. (1998), “ Technological Overlap and
Interfirm Cooperation: Implications for the Resource-Based View o f the Firm ” Research
Policy 27(5): 507-24.
N a g a r a ja n , A., and M it c h e l l , W. (1998), “ Evolutionary Diffusion: Internal and External
Methods used to acquire Encompassing, Complementary, and Incremental Technological
Changes in the Lithotripsy Industry,” Strategic Management Journal 19:1063-77.
N a r u l a , R. (2001), “ Choosing between Internal and Non-internal R&D Activities: Some
Technological and Economic Factors,” Technology Analysis & Strategic Management 13:
365-88,
------ (2002a), “ Innovation Systems and 'Inertia' in R&D Location: Norwegian Firms and the
Role o f Systemic Lock-in,” Research Policy 31: 795-816.
------ (2002b), “ R&D Collaboration by SMEs: Some Analytical Issues and Evidence,” in
F, Contractor and E Lorange (eds.), Cooperative Strategies and Alliances, Kidlington:
Pergamon Press, 543-68.
*------ (2003), Globalisation and Technology Cambridge: Polity Press,
------ and H a g e d o o r n , J. (1999), “ Innovating through Strategic Alliances: Moving towards
International Partnerships and Contractual Agreements,” Technovation 19: 283-94.
------ and Sad o w ski , B. (2002), “ Technological Catch-up and Strategic Technology Partner­
ing in Developing Countries,” International Journal of Technology Management 23:
599-617.
N o o t e b o o m , B. (1999), “ Inter-firm Alliances: Analysis and Design,” London: Routledge.
O d a g ir i , H., and Y a s u d a , H. (1996), “ The Determinants o f Overseas R&D by Japanese
Firms: An Empirical Study at the Industry and Com pany Levels,” Research Policy 25(7):
1059—79,
P a t e l , P. (1996), Are Large Firms Internationalising the Generation o f Technology? Some
New Evidence, IEEE Transactions on Engineering Management 43: 41—7.
and Р а ѵ іт г , K. (2000), National Systems o f Innovation Under Strain: The Inter­
nationalisation o f Corporate R&D,” In R, Barrel!, G. Mason, and M. O'Mahoney (eds.),
Productivity, Innovation and Economic Performance, Cambridge: Cambridge University
Press, 135-60.
and V e g a , M. (1999), “ Patterns of Internationalisation and Corporate Technology.
Location versus Home Country Advantages,” Research Policy 28:145-55.
 G L O B A L I Z A T I O N OF I N N O V A T I O N 345

* P e a r c e , R. (1990)? The In terna ti0 na1isa tio n of Resditch and Development, London: Mac-
millan.
------ (1999), “ D ecentralised R & D and Strategic Com petitiveness: Globalized Approaches to
Generation and use o f Technology in M ultinational Enterprises (M N Es),” Research Policy
28(2-3): 157-78,
P isa n o , G, (1990)>The R & D Boundaries o f the Firm : An Em pirical Analysis,” Adminis­
trative Science Quarterly 3 5 :15 3 -7 6 .
* R o n s t a d t , R. C. (1978), “ International R & D : The Establishm ent and Evolution o f Re­
search and D evelopm ent A broad by Seven U S M ultinationals,” Journal of International
Business Studies 9(1): 7-24.
S a n t a n g e l o , G . (2000), “ C orporate Strategic Technological Partnerships in the European
Inform ation and C om m un ication s Technology Industry,” Research Policy 2 9 :10 15 -3 1.
S e r a p io , M ., and D a l t o n , D. (1999), “ Globalisation and Industrial R & D : A n Exam ination
o f Foreign Direct Investm ent in R & D in the United States,” Research Policy 28: 303-16.
S io tis , G. (i999)> “ Foreign D irect Investm ent Strategies and Firm s' Capabilities,” Journal of
Economics and Management Strategy 8(2): 251-70,
S joholm , F, (199b), “ International Transfer o f Knowledge: The Role o f International Trade
and G eographic Proxim ity,” Weltwirthschaftliches Archiv 132: 97—115.
St o p f o r d , JL M ., and W e l l s jr., L. T (1972), Managing the Multinational Enterprise: Organ­
isation of the Firm and Ownership of SubsidiariesyN ew York: Basic Books.
* T e e c e , D. J. (1986), “ Profiting from Technological Innovation: Im plications for Integra­
tion, C ollaboration, Licensing and Public Policy,” Research Policy 15: 285-305.
------ (1992), “ C om petition , C ooperation and Innovation. O rganizational Arrangem ents for
Regimes o f R apid Technological Progress,” Journal of Economic Behavior and Organization
18 :1-2 6 .
------ (1996), “ Firm O rganisation, Industrial Structure and Technological Innovation,” Jour­
nal of Economic Behavior and Organization 3 1:19 3 -2 2 4 .
------ R u m e l t , R ,, D o s i , G ,, and W in t e r , S. G. (1994), “ U nderstanding Corporate C oher­
ence: T h eo ry and Evidence,” Journal of Economic Behavior and Organization 2 3 :1- 3 0 .
U N C T A D (2001), World Investment Report: Promoting linkages, N ew York: United Nations.
— — (2002), World Investment Report: TNCs and Export Competitiveness, N ew York: United
Nations.
V e r n o n , R. (1966), “ International Investm ent and International Trade in Product C y c le ”
Quarterly Journal of Economics 8 0 :19 0 -2 0 7 .
V e u g e l e r s , R ., and C a s s im a n , B. (1999). “ M ake and Buy in Innovation Strategies: Evidence
from Belgian M an ufacturin g Firm s,” Research Policy 28: 63-80.
W a r r a n t , E (1991), he Deploiement mondial de la R&D industrielle. Brussels, C om m ission
des C om m unautes Europeennes— Fast, December.
Z a n d e r , I. (1995), The Tortoise Evolution of the Multinational Corporation—Foreign Techno­
logical Activity in Swedish Multinational Firms 1890-1990, Stockholm : IIB.
* ---- (1999), “ H o w do you M ean ‘Global*? An Em pirical Investigation o f Innovation
N etw orks in the M ultinational C orporation,” Research Policy 2 8 :19 5 -2 13 .
Z a n f e i , A. (1994), “ Technological Alliances Between W eak and Strong Firm s: Cooperative
Ventures w ith A sym m etric Com petences,” Revue dyEconomic Industrielle 7: 255-79.
* ---- (2000), “ Transnational Firm s and the Changing O rganisation o f Innovative A ctiv­
ities,” Cambridge Journal of Economics 2 4 :5 15 -4 2 .
 In t r o d u c t i o n to P a r t III

T h e existence and significance o f differences acro ss in d u stry and

over tim e in the structure and o rg an izatio n o f the in n o vation
process has been a central top ic in the literatu re o n in n o vatio n .
A w ell-kn ow n distinction in m an y d iscu ssion s o f in n o v a tio n p r o ­
cesses is betw een so-called “ Sch u m p eter M a rk I ” in d u stries, ch ar­
acterized by n u m erou s sm all, en trep ren eu r-led firm s, and
“ Schum peter M ark II” indu stries, w hich are d o m in a ted b y large,
oligo p o listic firm s w ith extensive organ ized R & D . A s B ru la n d and
M o w e ry p o in t out in C h ap ter 13, how ever, such d ifferen ces are the
results o f lengthy processes o f h istorical ch an ge an d d ep en d on
both technological and in stitu tion al factors. M ale rb a , in C h ap ter
14, extends this analysis to a sam ple o f c o n te m p o ra ry “ h igh -tech ”
industries, and show s h o w differences in in n o va tio n an d in d u strial
d yn am ics m ay be analyzed as the in terp lay w ith in d ifferen t “ sec­
toral in n ovation system s” am o n g tech nology, actors an d in stitu ­
tions. C h apter 15 b y von Tunzelm an and A ch a, in contrast, lo o k at
so-called “ low -tech ” and “ m ed iu m -tech ” in d u stries, w h ich are
often assum ed to be less in n o vative than their “ h ig h -te ch ” c o u n ­
terparts. T h e authors, how ever, em phasize that in n o va tio n (often
in volvin g different processes) is p ervasive w ith in these sectors as
well. S im ilar findings are reported b y M iles in C h a p te r 16, w h ich
surveys the rap id ly exp an d in g literature on in n o va tio n in services.
H all, in C h apter 17, discusses the d iffu sio n o f tech n o logy, w h ich is
closely related to in n o vation , and u n d erp in s the p ervasiven ess o f
in n ovation th rou gh ou t the econom y.
 PART III

HOW
INNOVATION
DIFFERS
 In tr o d u ctio n to P a r t III
# # ♦ # * ♦ ♦ * * ♦ * * * # # » #**********»**•******•*******

T h e existence and significance of differences across industry and

over time in the structure and organization of the innovation
process has been a central topic in the literature on innovation.
A well-known distinction in many discussions of innovation pro­
cesses is between so-called “ Schumpeter Mark I” industries, char­
acterized by numerous small, entrepreneur-led firms, and
“ Schumpeter Mark И” industries, which are dominated by large,
oligopolistic firms with extensive organized R&D. As Bruland and
Mowery point out in Chapter 13, however, such differences are the
results of lengthy processes of historical change and depend on
both technological and institutional factors. Malerba, in Chapter
14, extends this analysis to a sample of contemporary “ high-tech”
industries, and shows how differences in innovation and industrial
dynamics may be analyzed as the interplay within different “ sec­
toral innovation systems” among technology, actors and institu­
tions. Chapter 15 by von Tunzelman and Acha, in contrast, look at
so-called “ low-tech” and “ medium-tech” industries, which are
often assumed to be less innovative than their “ high-tech” coun­
terparts. The authors, however, emphasize that innovation (often
involving different processes) is pervasive within these sectors as
well. Similar findings are reported by Miles in Chapter 16, which
surveys the rapidly expanding literature on innovation in services.
Hall, in Chapter 17, discusses the diffusion of technology, which is
closely related to innovation, and underpins the pervasiveness of
innovation throughout the economy.
 C H A P T E R 13

INNOVATION
T H R O U G H TI ME

K R IST IN E B R U LA N D

D AVID C. M O W E R Y

13.1 In t r o d u c t i o n

M o s t analysts o f in n o vatio n em phasize the im portance o f a historical approach,

with good reason. First, in n o vation is tim e consum ing, based on conjectures about
the future, and its ou tcom es typically are uncertain fo r long periods. A nalysis o f any
innovation therefore requires an understanding o f its history. Second, innovative
capabilities are developed th rou gh com plex, cum ulative processes o f learning.
Finally, in n o vation processes are shaped b y social contexts, as Lazonick has pointed
out: “ The social co n d itio n s affecting in n ovation change over tim e and v a ry across
productive activities; hence theoretical analysis o f the innovative enterprise m ust be
integrated w ith h istorical stu d y” (Lazon ick 2 0 0 2 :3 ).
H istorical patterns o f in n o vation are characterized b y com plexity, reflecting the
heterogeneous nature o f econ om ic activity, and the diversity o f processes o f tech­
nology creation across sectors and countries. These characteristics m ake it p ro b lem ­
atic to construct o verarchin g schem as o f historical developm ent. N evertheless, som e
historians and analysts o f in n o vation have developed taxonom ies o f epochs, often
based on “ critical tech n ologies” that define w hole periods o f developm ent. O ne
350 K R I S T I N E BRULAND AND DAVID С» MOWERY

form o f this is the w ave th eo ry p ro p o sed b y Sch u m p eter in Business Cycles, in which
steam po w er drove the First In du strial R evo lu tio n , electricity the Secon d Industrial
R evo lu tio n , an d so on. O ther w o rk that does n o t rely on w ave th eo ries also stresses
the role o f a sm all n u m ber o f technologies in d riv in g b ro ad er processes o f econom ic
grow th. A lth o u gh valuable, m an y o f these fram ew o rk s o verem p h asize the im port­
ance o f the allegedly critical technologies w h ile sligh tin g o th er areas o f innovation
and econ om ic activity that are no less im p o rtan t. In w h at fo llo w s w e challenge some
o f the h istorical discussions that stress the tran sfo rm ative effects o f “ critical innov­
a tio n s ” Instead, w e em phasize the co m p lex m u ltisecto ral ch aracter o f innovation,
and hence the need to take seriou sly the coexistence o f a ran ge o f in n o vatio n modes,
in stitu tion al processes, and organizational form s.
O u r discussion o f in n o vatio n th rou gh tim e h igh ligh ts chan ges in the structure
o f the in n o vatio n process in successive perio d s, and is In fo rm ed b y the innovation
system concept (discussed in Ch. 7 b y E dqu ist, C h . 11 b y A sh e im an d Gertler, and
Ch. 14 b y M alerb a). In ad optin g this fram ew o rk, we fo cu s o n the ch an gin g structure
o f econ om ic activity, changes in relevant in stitu tion s, an d ch an ges in patterns o f
know ledge generation and flow s w ith in em ergent in d u strial e c o n o m ie s.1
We begin the discussion below b y review ing recent historical interpretations o f
the im pulse tow ard industrialization in the w orld econom y. W e then discuss the
changing structure o f the innovation process in different phases o f industrialization,
focusing on the First Industrial Revolution in B ritain from rou gh ly 1760 to 1850, the
so-called Second Industrial Revolution during the late nineteenth and early twentieth
centuries, and w hat m ight be called the T hird Industrial R evolu tion after W orld War
II. O ur discussion thus includes the w idespread appearance of shop-
floor-driven technological innovation in eighteenth-century G reat Britain and
m oves forw ard to consider the invention o f the art o f invention, to use the philosopher
A. N . W hitehead s phrase (W hitehead 1925), in the late nineteenth an d early twentieth
centuries, w ith the em ergence o f organized industrial R8cD w ith in the firm . The Third
Industrial Revolution, m ost clearly illustrated by the p o st-w ar U nited States, is one in
w hich private and public institutions com pete and collaborate In new fields o f
innovation, a m ode o f innovation that is not yet exhausted.

13.2 T he F i r s t In d u s t r i a l R e v o l u t i o n

13.2.1 Institutions, Innovation and the Impulses to Growth

Sustained in n ovation -b ased developm ent is a recent and u n even ly distributed
historical ph enom en on . A substantial literature on uw o rld h isto ry ” has sought to
explain the rise o f the West, and p articu larly the E u ro p ean breakthrough to
 INNOVATION THROUGH TIME 351

sustained p ro d u ctivity grow th, in the late eighteenth century* W hy were som e
human societies able to break out o f a M althusian trap, shifting from 'ex ten sive ”
economic grow th that relied on increased labor input and a w ider division o f labor
to innovation-based intensive grow th w ith sustained rises in real output per head?
An im portant con trib u tion to this historical debate is Pom eranz (2000), who argues
that prior to the m id-eighteenth century, Europe, Japan, China, and India were at a
broadly sim ilar level o f econom ic developm ent— this was "a w orld o f surprising
resemblances*” W hy did o n ly N orthw estern Europe m ake the transition to in n o v­
ation-based growth? Pom eranz suggests that tw o factors were crucial: the acquisi­
tion by the m ajor E u rop ean pow ers o f colonies as m arkets for m anufactures and
sources o f fo o d and raw m aterial, and the developm ent w ithin Europe o f coal as a
new energy sou rce.2
An alternative exp lan ation for the industrialization o f N orthw est Europe stresses
institutional changes (see Braudel 1984; W allerstein 1974; Landes 1998), focusing on
the em ergence o f p ro p erty rights as im pulses to innovation. A varian t o f this insti­
tutional analysis is pro vid ed b y Jones (2003) who argues that technology-based
growth has occu rred at several points in w orld h istory; the challenge is less to
understand grow th than to understand the forces that prevent grow th. He stresses
the inh ibitory role o f political institutions that are based on surplus extraction by
political and m ilita ry elites. O nly w hen such rulers are weakened b y crisis do
opportunities arise for gain from innovation. Since the political pow er o f established
political elites in N orthw est E urope eroded during the fourteenth to seventeenth
centuries, the em ergence o f sustained, innovation-based econom ic grow th first
occurred in this region o f the w o rld econom y.
There is disagreem ent w ith in this literature over the tim in g o f the divergence, as
well as the relative im po rtan ce o f different factors in su p portin g the grow th o f such
institutions as private p ro p erty rights and the weakening o f rent-seeking political
and m ilitary elites. B u t all o f the scholars adopting this approach em phasize insti­
tutional change as an indispensable precondition for sustained innovation-led
growth.

13.2.2 Innovation in the First Industrial Revolution

M ost econom ic h istorian s regard the developm ents in B ritain and N orthw estern
Europe from aro u n d 1760 as an econom ic and technological watershed. In novation
during this p erio d is best conceptualized as an econom y-w ide process that involved
technological, organ ization al, and institutional change, spanning m any sectors and
product groups. T h is v ie w o f British industrialization contrasts w ith the classic
historical accounts that em phasize epochal technological breakthroughs in steam
pow er and textile technologies (see e.g. M an to u x 1961). The debate is a significant
one for the b ro ad er study o f in n o vation , since im portan t scholarly pieces in the field
352 K R I S T I N E B R ULA ND A N D DAVID C. M O W E R Y

o f in n o vation studies have follow ed the *key in n o va tio n s in terp retatio n o f the First
In dustrial R evo lu tio n (e.g. Freem an an d Lou^a 2002; fo r an e co n o m ic h istory o f
in d u strializatio n in this fram ew ork, see L lo y d -Jo n e s an d L ew is 1998).

13.2.3 Sectoral Patterns of Technological Advance:

The Patenting Evidence
O ne im p o rtan t source o f evidence on the pace an d sectoral d istrib u tio n o f innova­
tive activities d u rin g the In d u strial R evo lu tio n is patent statistics fro m the period.
A lth o u gh the high cost o f paten tin g (ap p ro xim ately £ 12 0 fo r E n g lan d , at a tim e when
the an n ual in com e o f a skilled w o rk er w as ab o u t £50) an d lim ited access to patent
attorneys b y m an y inventors argu ab ly m ake patent d ata a biased so u rce o f evidence,
no oth er co m p arab ly com prehensive sources exist on in n o vative a ctiv ity d urin g the
In d u strial R evolu tion . M acL eo d (1988) finds that p aten tin g grew rap id ly after 1750,
especially in capital goods. T h e tw o technologies fo r w h ich p aten tin g grew most
rap id ly d u rin g this perio d are po w er sources and textile m ach in ery. B u t patenting
also expand ed sign ifican tly in other cap ital-go o d s categories; in clu d in g agricultural
equipm en t, brew ing, sh ipb u ildin g, canal b u ild in g, and m etallurgy. A lth o u gh the
share o f all patentin g accounted fo r b y capital go od s grew d u rin g the 1750-1800
p eriod, this category nevertheless accounted fo r no m ore than 40 p er cent o f British
patents in the h a lf cen tu ry betw een 1750 and 1800.
A great deal o f inventive activity d u rin g this p erio d fo cu sed o n co n su m er goods.
A cco rd in g to B erg (B erg 1998; see also Su llivan 1990), m u ch o f this consum er-good s
paten tin g affected a vast n u m b er o f sm all, novel p ro d u cts such as buckles and
fasteners, cabinets and fu rn itu re, and spectacle fram es. In deed m u ch o f the patent
activity w ith in the textiles sector— ro u g h ly o n e-th ird — in volved n ew products
(G riffith s et al. 1992). M u ch o f the inventive activity in this k ey sector w ithin the
In dustrial R evo lu tio n involved new thread types and fab rics, an d focused on a
co n su m er m arket.
Patent evidence thus suggests that the p erio d o f the In d u strial R evo lu tio n was a
period o fb ro a d technological change. N evertheless, in reco gn itio n o fth e lim itations
o f patent data fo r tracking in n ovative activity, we tu rn n o w to m o re qualitative
evidence on the sectoral structure o f in n ovation .

13.2.4 Sectoral Patterns of Change: Technological Histories

13.2.4.1 Steam Pow er a n d Textiles

Fo u r in n o vation s the sp in n in g jenny, the w ater fram e, the sp in n in g m ule, and the
autom atic mule- were associated w ith d ram atic grow th in the B ritish textiles
 INNOVATION THROUGH TIME 353

industry o f the First Industrial R evolution. Between the late eighteenth and the
middle o f the nineteenth century, the cotton textile indu stry grew spectacularly in
the size o f its output, in lab or productivity, in the scale o f enterprises, in capital
employed, and in its share o f national incom e. Value added in cotton rose from less
than £500,000 in 1760 to about £25,000,000 by the m id-i82o$. In spinning, the
num ber o f direct lab or h ou rs required to process 100 pounds o f cotton declined
from 300 in 1790 to 135 in 1820 (M o kyr 2002: 5 0 -1), and the average annual input o f
raw cotton per facto ry rose by over 1,000 per cent during 1797-1850 (C hapm an, 1972:
70). D ram atic as these changes were, they should be kept in proportion: textiles
made up about 25 per cent o f m anu factu ring output at their peak. In novation and
productivity w ere grow in g elsewhere as well.
Another critical in n ovation o f this period was the steam engine o f Jam es Watt,
first introduced in 1775. W att s innovation is com m on ly described as the em blem atic
technology o f the In dustrial R evolution (see Toynbee 1908; Deane 1965). Yet von
Tunzelm anffs study o f steam pow er (1978) show ed that the m achine diffused
relatively slowly, that it had only m odest econom ic advantages over existing pow er
technologies (and hence could not significantly affect econom ic grow th), and had
limited backw ard and fo rw ard linkages w ith the rest o f the British econom y, further
reducing its “ catalytic” effects (see B o x 13.1). As we noted earlier, the innovations that

Box 13.1 Technological diffusion in the First Industrial Revolution

Since the economic effects of innovations depend on their widespread adoption

(see Ch. 17 Hall by in this volume), it is important to recognize that many of the
important innovations of the First Industrial Revolution in fact diffused relatively
slowly. For example, the Watt steam engine, described above as an emblematic innov­
ation of the First Industrial Revolution, diffused gradually through the British econ­
omy. By 1800, twenty-five years after the introduction of the Watt steam engine,
Manchester (a central locus of industrialization in textiles) had about 32 engines,
and Leeds (another emergent textiles center) about 20 . By 1817 Glasgow had 45
engines, by 1820 Birmingham had about 60 engines, and by 1825 Bolton had 83.
Growth rates of steam-generated horsepower averaged between 6 and 10 per cent per
year in the late 1830s, more than 50 years after the Watt engine's introduction. Von
T te e lm a rm ( 1978) argued that this gradual pace o f diffusion reflected the high costs
o f steam engines and their fuels through the 1850s, long after the introduction of the
engine. Similar points apply to other important innovations of this period. The
: .'Roberts automatic spinning machine, said by no less an observer than Karl Marx to
^o pp iu p a completely new epoch in the capitalist system,” was a major innovation
. wortdft first truly automatic power machine. But it diffused slowly; fifty years
before ibis machine accounted for a majority o f the output of the UK cotton
354 K R I S T I N E B R U L A N D A N D D A V I D C. M OW ER Y

co n trib u ted to B ritish eco n o m ic grow th an d in d u strializatio n sp an n ed a broader

gro u p o f technologies and sectors,

13.2.4,2 Innovation in O ther Sectors

A lth o u gh in du strialization necessarily w as associated w ith a fall in the share o f
national o u tp u t flo w in g fro m agriculture, B ritish ag ricu ltu re g rew in absolute
term s d u rin g 17 50 -18 50 and w as h ig h ly in n ovative. D u rin g this p erio d , key innov­
ations w ere developed in farm tools, cu ltivatio n im p lem en ts (p lo w s, harrows,
m o w ers), sow in g im plem ents, harvesting eq u ip m en t (reapers, rakes, hoes, scythes,
w in n o w in g and threshing devices, etc.), and d rain age eq u ip m en t (fo r a detailed
overview , see B ru lan d 2004). A gricu ltu ral in n o vatio n w as associated w ith the emer­
gence by the 1830s o f a specialized agricu ltu ral eq u ip m en t in d u stry, w h ich in turn
su p p o rted the grow th o f n u m erou s sm all en gin eerin g w o rk s an d fou n d ries.
C lo sely linked w ith technical change in the ag ricu ltu ral sector w ere innovations in
the processing, d istrib u tion , and co n su m p tio n o f fo o d , w h ich d u rin g the Industrial
R evo lu tio n (and after) d om in ated B ritish m a n u factu rin g . T ech nological innov­
ations in fo o d preservation , refrigeration, b ak in g, b rew in g , and g rain m illin g sup­
ported exp an sion in the scale o f p ro d u ctio n establish m ents an d organizational
in n o vation o f p ro d u ctio n and firm s. B ak in g w as the first B ritish in d u stry to develop
and use the p ro d u ctio n line, based on new techniques that su p p o rte d m o re accurate
tim in g o f operations. B rew in g and m illin g w ere the first sectors to d eploy large,
p ro fession ally m anaged enterprises w ith n ation al d istrib u tio n system s.
A sim ilarly innovative sector was the glass industry, w h ich m anufactured widely
used and differentiated products— w indow s, bottles and containers, lam ps, and
spectacles. Glass w as one o f the few large-scale p ro d u ction activities in early industri­
alization, and relied o n experim entation and research to a degree n o t w idely appreci­
ated in m an y accounts o f the role o f science in technological in n o vatio n during this
period. The m ost know ledge-intensive segm ent o f glass p ro d u ctio n w as optical glass,
where developm ents o f the technology deployed optical theory, pio n eerin g the inte­
gration o f science w ith production. A lth ou gh the first Industrial R evolu tion overall
was far fro m a science-based phenom en on , developm ents in at least som e o f the key
innovative sectors prefigured subsequent changes in the organization o f innovation.
These exam ples o f in n o vation cou ld easily be exp an d ed to in clu d e sectors such as
iron and steel, chem icals (alkalis and ch lorin e), p o ttery and ceram ics, m achinery
and m achine tools, instrum ents, m in in g, and p ap er and p rin tin g . T h e pervasiveness
and extent o f in n o vation in these and oth er indu stries o n ce again suggests that
in n o vation d u rin g the First In du strial R evo lu tio n w as n ot co n fin ed to "leading
sectors o f the econom y, but was present in v irtu a lly all sectors d u rin g the period.
We cannot ignore the sectors such as textiles and steam p o w e r that have driven so
m uch o f the h isto rio grap h y o f in du strialization ; but their role needs to be kept in
econ o m ic and technological perspective.
 INNOVATION THROUGH TIME 355

13.2.5 The Organization of Innovation and Learning

in Early Industrialization
How was innovation organized during the First Industrial Revolution? Analyses o f
patent data suggest that virtually all inventions in the late eighteenth and early
nineteenth centuries resulted from the efforts o f individual inventors. These inventors
may have w orked alone in individual w orkshops or in larger enterprises, but the key
point was the individ u ality o f inventive and innovative effort, and its integration with
shop-floor production. Inventors' skills and knowledge bases were rooted in existing
trades, such as w atchm aking, carpentry, blacksm ithing, m etalworking, and w o od ­
working. Textile m achinery in particular was fabricated largely within the existing
textile-producing firm s. A specialized capital goods sector appeared only in the 1820s.
The upsurge in inventive and innovative activity du ring the Industrial R evolution
did not depend in any general w ay on “ science" as we now understand it, although
isolated instances o f the integration o f science and in d u stry du ring the late eight­
eenth and early nineteenth centuries are apparent in such areas as glassm aking. The
period saw the em ergence o f form al and inform al scientific societies in Great Britain
and w ide d iffu sion o f scientific ideas (see U glow 2002), but this early “ scientific
revolution" p ro d u ced few practical applications. A lthough the search and learning
processes em ployed by inventors during this period are best described as “ trial and
error," this characterization inaccurately m inim izes the extent and sophistication o f
the know ledge required for in n ovation in early industrialization. Indeed, M o k yr has
proposed that a central factor in the Industrial R evolution was an “ Industrial
Enlightenm ent," associated w ith im provem ent in the quantity and accessibility o f
knowledge concerning indu strial techniques. This “ E nligh tenm ent" included the
surveying and co d ificatio n o f artisanal techniques in published m anuals, h an d ­
books, textbooks, and pam phlets on industrial practices (M okyr 2002: 34--5). P at­
terns o f learning and know ledge accum ulation during the Industrial Revolution m ay
have begun as tacit and practical, but d u rin g the late eighteenth and early nineteenth
centuries, m ore and m ore o f this learning was codified, accelerating the d iffusion o f
industrially relevant know ledge across sectors.

13.2.6 Institutions and the Organization of Enterprise

during the First Industrial Revolution
Institutional change that affected the organization o f firm s and produ ction p ro ­
cesses played an im p o rtan t role in the upsurge o f innovation du ring early in du stri­
alization. T h is is a vast top ic, and we focus on tw o crucial institutional changes— the
developm ent o f new fo rm s o f co m p an y law and finance that supported the grow th o f
356 K R I S T I N E B R U L A N D A N D D A V I D C. M O W E R Y

co rp o rate firm s; and the rise o f m an agerial con trol o f p ro d u c tio n , w h ich trans­
fo rm ed w orkplace o rgan ization and scale. These in stitu tio n al in n o v a tio n s together
m ad e possible the subsequent grow th in facto ry p ro d u ctio n .
M o st in du strial enterprises o p eratin g d u rin g the eigh teenth cen tu ry were
extrem ely sm all. Large-scale factories w ere u n co m m o n b efo re the early nineteenth
century, and the sm all-scale w o rk sh o p o r p ro d u ctio n u n it w as the p rim a ry organiza­
tional fo rm for m ost o f the p erio d o f early in d u strializatio n . T h ese sm all firm s were
in d ivid u ally ow n ed o r were partnerships, lo cally fin an ced , in w h ich liab ility for
debts was the personal resp on sib ility o f ow n ers w h o u su ally acted as m anagers. Two
in stitution al fo rm s m ade possible an exp an sio n in the scale o f enterprises: joint-
stock (i.e. lim ited liability) organ izatio n an d the grow th o f fin an cial netw orks.
Jo in t-sto ck association s em erged in the m edieval p e rio d in B ritain , but were
perm itted o n ly via the explicit au th orization o f the state. A series o f piecemeal
reform s after 1825 w ere follow ed b y legislation p erm ittin g the creatio n o f com panies
w ith separate legal identity, lim ited liab ility an d tradeable shares. G en eral legislation
fo r the jo in t stock fo rm w as passed in 1844 and co n so lid ated in the statutes o f 1856
and 1862 (M ath ias 1983:325; see H arris 2000 for a co m p reh en sive acco u n t). Although
m uch in du strial fin an cin g rem ained local and sm all in scale (see H u d so n 1986 for an
account o f local n etw orks5 role in fin an cin g the w o o llen in d u stry ), these legal
reform s enabled substantial grow th in the fin an cin g an d scale o f in d u strial enter­
prises. But jo in t stock o rgan ization and access to finance w ere n ecessary rather than
sufficient co n d itio n s for enterprise grow th. Even m o re sign ifican t w as the develop­
m ent o f m anagem en t system s and m an agerial control.
M anagers o f these early indu strial enterprises co n fro n ted serio u s challenges in the
assem bly and m aintenance o f a suitable w o rk fo rce, the co n tro l o f w o rk , and the
ad o p tion o f new techniques and organ ization al stru ctu res fo r p ro d u c tio n activities
by a restructured w orkforce. Pollard h igh lights “ tw o distinct, th o u g h clearly overlap­
p in g difficulties; the aversion o f w ork ers to enterin g the n ew large enterprises with
their unaccustom ed rules and discipline and the sh ortage o f skilled and reliable
lab o u r (Pollard 19 6 5 :16 0 ). The em ergence o f rule-b ased d isc ip lin a ry m ethods, the
lab orio u s constru ction o f su p erv iso ry system s, and the h ab itu atio n o f w orkers to an
organized and controlled w o rk in g day em erged slo w ly b u t w ere central develop­
m ents ot early industrialization. N ew m anagem en t tech n iq u es that app eared during
the In dustrial R evo lu tio n perm itted the d evelopm ent o f larger, centralized pro d u c­
tion sites and o f the m echanized factory. In turn, such sites p erm itted the application
o f pow er, the ad o p tion o f new indu strial techniques, an d closer m an agerial control
over the organization and pace o f w ork.
These organizational and m anagerial in n o vatio n s w ere d efin in g characteristics o f
the First In dustrial R evolution. In p o ttery fo r exam ple, the m o st im p o rta n t m an­
agerial in n o vato r was Josiah W edgw ood, w h o d eveloped a n u m b er o f product
in n o vation s— new designs, new glazes and finishes, and n ew basic m aterials— and
pioneered new m arketing m ethods. But his m ost im p o rta n t in n o vatio n s were
 INNOVATION THROUGH TIME 357

organizational— the creation o f an integrated w orkforce, the design o f a plant

organized around a set o f pro d u ction sequences, and above all the creation o f a
workforce subject to control and discipline (M cK endrick 1961; see Box 13.2), W edg­
wood s innovations strengthened m anagerial pow er over the production workforce,
which form ed the central context for innovation (and struggle) in the later n in e­
teenth century.

Box 13.2 Josiah Wedgwood and "modern" management in pottery fabrication

In the second half o f the eighteenth century, rising incomes and increased coffee and
tea consumption accelerated growth in the market for china and other types o f glazed,
fired clay plates, cups, and related items. This was part o f a wider growth In demand for
"lu xu ry5 consumer goods (Berg and Eger 2003). The production o f pottery was
concentrated in Staffordshire in central England, and was dominated by small enter­
prises operated by craftsmen, often producing on a piecework basis. Production was
controlled by individual craftsmen, and production rhythms and volumes were hap­
hazard. Josiah Wedgwood transformed the industry by developing factory-based
production techniques that supported the creation o f an enterprise o f unprecedented
scale. Wedgwood's success rested on two achievements. First, he successfully lobbied
the British government to improve regional transportation infrastructure (a publicly
financed turnpike was built in 1763 and a canal, on which Wedgwood sited his factory,
was completed in 1771), thereby enabling his factory to serve the British market while
reducing formerly exorbitant breakage rates. Second, he introduced radical organiza­
tional innovations, developing new techniques for organizing production and man­
aging the workforce (Bruland 1989).
Wedgwood, an acquaintance of Matthew Boulton, the entrepreneur who formed the
successful steam-engine firm of Boulton and Watt, modeled his new production
organization on Boulton’s factory, emphasizing a physical layout that separated and
sequentially organized the various operations that went into production of his china
(Langton 1984). Consistent with this organization, Wedgwood assigned workers to
specific tasks, relying on specialization to enhance skill and consistency in the per­
formance of these tasks. Workmen "were not allowed to wander at will from one task to
another as the workmen did in the pre-Wedgwood potteries. They were trained to one
task and they had to stick to it” (McKendrick 1961: 32).
Having reorganized the structure of production and jobs within his organization,
Wedgwood had to develop techniques to encourage and/or force workers to adapt to
this new system. He invested heavily in the retraining o f experienced workers (with
mixed results) and in the training of new employees, many of whom were young
women (women accounted for 25 per cent of his employees as of 1790). Even more
important, however, was Wedgwood’s emphasis on codification o f technical guidelines
for the performance o f the various tasks in his factory and development of extremely
Леія\\рЛ written rules for worker behavior, Wedgwood also introduced sanctions and
rewards for punctuality and absenteeism on the part o f workers, going so far as to
develop an early prototype o f a timeclock for monitoring workers attendance.
358 K R I S T I N E R R U L A N D A N D D A V I D C. M O W E R Y

Bok1 3 J (cont) ___ ■ " -■ ^

W edgwoods new methods were significant organizational changes in the produc­
tion o f kitchenware and china, resting on a transformation o f the nature and character
of work itself* The new methods encountered considerable resistance from experienced
workers, but he successfully created a production system without equal in the industry,
employing 200 workers* By 1790, less than twenty-five years after its foundation*
Wedgwood himself was enormously wealthy, and the firm survived as an independent
entity into the twentieth century.

A n essential ingredient in the tran sfo rm atio n o f eco n o m ic a n d in n o vative activity

that characterized the First In d u strial R evo lu tio n thus w as the d evelo p m en t o f new
techniques o f econ om ic organization and m anagem en t, A n o th e r w ave o f institu­
tion al change and m anagerial in n o vation pro ved in d isp en sab le to the Second Indus­
trial R evo lu tio n and an organizational in n o vatio n that w as at its heart— the
in d u strial research laboratory*

13.3 T h e S e c o n d I n d u s t r i a l R e v o l u t i o n

13.3.1 A Second Phase of Industrialization

In the late nineteenth cen tu ry in du strial tech nologies began to chan ge, and a range
o f n ew technologies and industries em erged. T h is Seco n d In d u strial Revolution
too k place on the continent o f Eu rop e and in the U SA . In E u ro p e it w as led by the
em ergence o f new indu strial regions in France and G erm an y, such as the Ruhr* It
involved a shift aw ay fro m the basic indu stries that h ad d evelo p ed in B ritain before
d iffu sin g to E u rop e and the U n ited States (iron, steel, coal, textiles, an d m echanical
engineering), to new indu strial sectors (such as ch em icals, o p tics, an d electricity),
and signaled the passing o f tech nological leadersh ip fro m B ritain to the United
States and Germ any*
The Second In dustrial R evo lu tio n w as characterized b y o rgan izatio n al innov-
alions that laid the g ro u n d w o rk for links betw een in d u stry and fo rm a l science that
becam e stronger d u rin g the course o f the tw entieth century. T h e developm ent o f
these stronger links tran sform ed the in n o vatio n process in several w ays: (1) Form al
.raining for w o u ld -b e inventors becam e far m o re im p o rtan t and the role o f artisanal
n gen u ity d im in ish ed ;3 (2) the role o f in stitu tion s extern al to the firm that
 INNOVATION THROUGH TIME 359

conducted such form al training and research increased; and (3) bodies o f em pirically
grounded, codified scientific and technological know ledge internal to the firm
became pow erful engines for expansion and diversification.
The technological shifts o f the late nineteenth century were accom panied by
changes in firm structure. Large-scale, vertically integrated enterprises em erged in
Germ any and the U nited States that incorporated specialized research and develop­
ment departm ents or laboratories. W ithin such firm s scientific w ork was carried out
by teams o f researchers and depended on netw orks o f scientific contacts in the
education (particu larly university) system s. These professionally m anaged firm s o f
unprecedented size becam e the agents o f Schu m peter’s “ creative destruction” by the
m id-twentieth century, as industrial innovation becam e a core com ponent o f
corporate strategy.

13.3.2 Was the Second Industrial Revolution

a “ Science-Driven” Phenomenon?
Although im po rtan t scientific breakthroughs and an expanded application o f sci­
ence to indu stry did em erge in the late nineteenth century, for m ost o f the century
these two trends w ere m ore loosely coupled to one another than is com m on ly
thought. The constru ction o f a bridge between recent scientific discoveries and
technological in n o vatio n typ ically requires considerable tim e. For exam ple, no
significant technological applications follow ed Faraday s dem onstration o f electro­
magnetic in d u ction in 1831, w ith the exception o f the telegraph. Yet this scientific
discovery laid the fo u n d atio n s for one o f the defining industries o f the Second
Industrial R evolu tion , electrical equipm ent and electric pow er generation.
As the exam ple o f electricity suggests, technological exploitation o f new scientific
understanding often requires considerable tim e, since additional applied research is
needed to translate a new but abstract form u lation into econom ically useful k n o w ­
ledge. In other im p o rtan t cases, such as Perkin's accidental synthesis in 1856 o f
m auveine, the first synthetic dyestuff, exploitation o f scientific advances required
the developm ent o f co m p lex process technologies for w hich no scientific foundation
existed. A lth o u gh chem ical science was vitally im portan t to industrial developm ents
during the p eriod, m u ch o f the actual tim in g o f innovation, i.e. the translation o f
scientific breakthroughs into com m ercial products, depended on advances in
m anufacturing technologies that rem ained p o o rly understood through m uch o f
the nineteenth century.
In other indu stries, the linkage between science and technological innovation
rem ained w eak, sim p ly because technological innovation did not require scientific
knowledge. T h is w as true o f a b ro ad range o f m etal-using industries in the second
h a lf o f the nineteenth century, a period d u rin g which A m erica took a position o f
36O K R I S T I N E B R U L A N D A N D D A V I D C. M O W E R Y

tech nological leadership. T h e d evelopm ent o f this n ew m ach in e tech n o logy rested
o n m echanical skills o f a high order, as w ell as co n sid erab le in g e n u ity in conception
and design. It required little o r no recourse to the scien tific k n o w led ge o f the time,
an d U S success in this and oth er indu stries, such as ch em icals, m eatpacking, and
co n su m er go od s, relied on access to a large d o m estic m arket.
T h e creation o f a tru ly “ n atio n al" m arket in the U n ited States in turn was
facilitated b y the con stru ction o f a reliable n atio n al in fra stru ctu re fo r com m uni­
cations and tran sportation . T h e enterprises m o st h ea v ily in v o lv ed in the creation of
this in frastru ctu re w ere them selves am o n g the largest in d u strial firm s organized in
the U nited States, and the organizational and fin an cial in n o va tio n s developed by
firm s such as W estern U n io n and the P en n sylvan ia R a ilro a d w ere w id e ly emulated in
other industries as new firm s o f unprecedented scale w ere created (C h an d ler 1977).
But few i f any o f these econ o m ically crucial o rg an izatio n al in n o va tio n s relied on
science.

13*3.3 The Origins of Industrial Research

A defin in g characteristic o f the “ new in d u stries" o f the S eco n d In d u strial Revolution
w as their increased reliance on organized exp erim en tatio n . T h e pion eers in this
organizational innovation were the large German chemicals firms that grew rapidly
in the last quarter of the nineteenth century, based on innovations in dyestuffs. By
the first decade of the twentieth century, a number of large US firms had established
similar in-house industrial research laboratories. In both nations, the growth of
industrial research was linked to a broader restructuring of manufacturing firms that
tran sfo rm ed their scale, m anagem en t structures, p ro d u c t lines, an d global reach.
B u t the development o f indu strial research in the G e rm a n ch em icals and electrical
equipment industries also relied on complementary changes in the institutional
structure of the nascent “ German" innovation system that occurred before and after
German unification in 18 7 0 .4
Scientific advances in physics and chemistry during the last third of the nineteenth
cen tu ry created considerable poten tial for the p ro fitab le a p p lica tio n o f scientific and
technical knowledge in industry. The first in-house industrial R & D laboratories
were established by German firms seeking to commercialize innovations based on
the rapidly developing field o f organic chemistry. Kekules 18 6 5 model of the
molecular structure of benzene, a key co m p o n en t of organic chemistry and synthetic
dyestuffs, provided the first scientific foundation for developing new products. But
scientifically trained personnel were needed to translate Kekule’s breakthrough into
new products. 1 he rapid expansion in Germany’s network of research and technical
universities during the second half of the nineteenth century thus was critically
im p o rtan t to the growth of industrial research, particularly in the chemicals Indus­
 INNOVATION THROUGH TIME 361

try, G erm an universities produced a large pool o f scientifically trained researchers

(many o f w h om sought em ploym ent in France and G erm an y during the 1860s),
university faculties advised established firm s, and university laboratories provided a
site for industrial researchers to conduct scientific experim ents in the early stages o f
the creation o f in -h ou se research laboratories.
The G erm an universities pioneered the developm ent during the nineteenth
century o f the m odern m odel o f the “ research u n iversity” in w hich faculty research
was central to the train in g o f advanced degreeholders. In addition, the G erm an
polytechnic institutes that had been founded during the 1830s b y the various
Germ an principalities were b y the 1870s transform ed into technical universities
that played a central role in training engineers and technicians for the chem icals
and electrical equipm en t industries. B y the 1870s, according to M u rm ann (1998),
Germ any had nearly th irty university and technical university departm ents in
organic chem istry, and seven m ajo r centers o f organic chem istry research and
teaching. A n d technically trained personnel m oved into senior m anagem ent p o s­
itions w ithin G erm an industry, in contrast to the situation in Great Britain, further
strengthening the links between corporate strategy and industrial research.
The contrast between G erm an y and Great Britain in the role o f universities is
especially strikin g.5 British universities received far less pu blic funding, supported
less technical education, and w ere less closely linked w ith dom estic chem icals firm s
than was true in G erm an y by the 1880s. British university enrollm ent increased b y 20
per cent between 1900 and 1913, far less than the 60 per cent increase in G erm an
university enrollm ent d u rin g the sam e period. Enrollm ent at the “ redbrick” British
universities (largely fo u n d ed d u rin g the nineteenth century, this group excludes the
ancient English universities o f O xford and C am bridge) grew from rough ly 6,400 to
9,000 during 18 9 3 -19 11, b u t on ly 1,000 o f the students enrolled in these universities as
o f 1911 w ere engineering students, w hile 1,700 w ere pu rsuing degrees in the sciences
(H aber 19 7 1:5 1). B y contrast, the G erm an technical universities alone enrolled 11,000
students in engineering and scientific degree program s by 1911. British governm ent
funding o f high er edu cation am ou nted to rou gh ly £26,000 in 1899, w hile the
Prussian govern m en t alone allocated £476,000 to support higher education. B y
1911, these respective am ou nts stood at £123,000 and £700,000 (H aber 1971: 45
and 51).
The institutional tran sfo rm ation o f G erm an y's national innovation system was
both a cause and an effect o f the grow th o f the chem icals and electrical equipm ent
industries. W erner von Siem ens o f the Siem ens electrical equipm ent firm w as a
founder o f the G erm an A ssociation for Patent Protection in 1874* and the first
national patent law in the new G erm an state was passed in 1877. A lthough the law
did not cover d yestu ff products, stronger intellectual p ro p erty protection increased
the ease w ith w h ich firm s could app ropriate the returns to their R & D , and m any o f
the largest G erm an chem icals firm s established form al in-house R8cD laboratories
after its passage.
362 K R I S T I N E BR U L A N D A N D D A V I D C. MOWERY

The large, profitab le firm s that em erged in these scien ce-b ased in d u stries actively
lobbied the G erm an govern m en t fo r increased su p p o rt o f h ig h e r edu cation (The
“ C lu b o f G erm an C h e m ists;5 draw n largely fro m sen io r m an ag em en t o f German
chem icals firm s, lobbied fo r ad d ition al facu lty a p p o in tm en ts in ch em istry) and for
other fo rm s o f su p p o rt o f research related to their enterprises. W ern er v o n Siemens
d onated the land for the Im perial Institute o f Physics and T ech nology, located near
his firm 's h ead qu arters in B erlin and the city's tech n ical u n iversity, and the Institute
w as fo rm a lly established w ith pu blic fu nds in 1887. S im ila r lo b b y in g b y the chemicals
in d u stry led to the an n ou n cem en t in 19 10 b y the G e rm an e m p e ro r o f the foundation
o f the K aiser W ilhelm Institute for C h em istry, staffed larg ely b y acad em ic chemists
an d fu n ded b y industry. B o th the Institute o f Physics an d T ech n o lo gy and the Kaiser
W ilhelm Institute were dedicated to “ m issio n -o rie n ted ” fu n d am en tal research,
m uch o f w hich was lon ger-term in nature than the R & D p e rfo rm e d in industry
b u t nonetheless m ore applied than the w o rk o f u n iversity faculties (Beyerchen 1988).
The creation b y G erm an chem icals firm s o f in -h o u se in d u strial research labora­
tories also w as associated w ith change in the m an agem en t an d stru ctu re o f these
firm s (see B o x 13.3). F am ily m anagers w ere replaced b y p ro fe ssio n al m anagers and,
eventually, b y p rofession al chem ists. T h eir in -h o u se R & D activities produced new
products in fields other than dyestuffs, e.g. B ayer s asp irin . A n d the im portance o f

Box 13.3 The foundation of R&D laboratories by Bayer and Du Pont

Bayer's foundation o f a laboratory was triggered in part by a realization among the

firm's senior management that it was unable to compete effectively with Hoechst and
BASF (which had founded research laboratories respectively in 1877 and 1878), as well
as the growing difficulties that Bayer faced in forming strong linkages with leading
university research chemists. In 1883, Carl Duisberg, who later served as the first
director o f Bayer's in-house research facility and the firm’s CEO, was sent by Bayer
managers to work with the chemistry faculty at the University o f Strasbourg (then a
German university), before returning to Bayer to begin work in the firm's R&D
laboratory (a small room just off the main production floor in Bayer's plant). At the
same time, Bayer sought to strengthen its links with German university chemists
through other tactics, including the negotiation o f contracts with leading research
chemists and the funding o f research by new Ph.D. degreeholders in university or
technical university laboratories.
Duisberg s first laboratory was at best an appendage to Bayer's main production
facility, but his success in dyestuff synthesis led to an expansion in his staff. Neverthe-
less, Duisberg s group had important responsibilities in production engineering and
problem-solving, as well as marketing, until roughly 1890. Only in 1891 was a
dedicated laboratory established at Bayer and a clear distinction made within the
organization between R&D and workaday technical support (see Meyer-Thurow
1982).
 INNOVATION THROUGH TIME 363

The US chemicals firm Du Pont established its first industrial research facility, the
Eastern Laboratory, in 1902, and founded the Experimental Station in 1903. Creation
of the Eastern Laboratory followed the acquisition of control of the Du Pont Company,
founded in the early nineteenth century, by T. Coleman Du Pont, Pierre S. Du Pont,
and Alfred I. Du Pont from other family members. The Company’s transformation
from a loose holding company to a multifunction, diversified industrial corporation
began with this 1902 change in control.
Du Pont s Eastern Laboratory was the first laboratory to be physically and organiza­
tionally separated from the manufacturing operations of the firm. Its R&D activities
were devoted almost entirely to improvements in manufacturing processes for Du
Pont’s existing product line of dynamite and high explosives. By contrast, the Experi­
mental Station, founded one year later, focused on the development of new products
and improved applications of Du Pont’s smokeless-gunpowder products. The Experi­
mental Station also monitored and evaluated inventions from sources outside of the
Du Pont Company.
AUS government antitrust suit against Du Pont forced the divestiture of a portion of
its black powder and dynamite businesses in 1913, and the firm used its R&D
laboratories to diversify its product lines through R&D and the acquisition of tech­
nologies from external sources during and after World War I.

dose links between research personnel and the users o f these new products, as well as
the proliferation o f new products, triggered expansion in these firm s’ internal
distribution and m arketing capabilities. Just as U S firm s were d oing by the end o f
the nineteenth century, the G erm an chem icals firm s expanded their boundaries to
incorporate new fu n ction s and a m uch broader and m ore diversified product line.
A sim ilar sequence occu rred at rou gh ly the same tim e in the G erm an electrical
equipm ent industry, as Siem ens and A E G , am ong other leading firm s, established
in-house research laboratories d u rin g the 1870s and 1880s.
The developm ent o f indu strial research w ithin U S m anu factu ring firm s follow ed
these developm ents in the G erm an chem icals and electrical m ach in ery industries.
M any o f the earliest U S corporate investors in industrial R & D , such as General
Electric and A lcoa, w ere fo u n d ed on product or process innovations that drew on
recent advances in physics and chem istry. The corporate R & D lab orato ry brought
m ore o f the process o f developing and im p ro vin g industrial technology into the
boundaries o f U S m an u factu rin g firm s, reducing the im portance o f the independent
inventor as a source o f patents (Sch m ookler 1957).
But the in -h ou se research facilities o f large U S firm s were not concerned exclu­
sively w ith the creation o f new technology. Just as the G erm an dyestuff firm s
laboratories had, these U S indu strial laboratories also m on itored technological
developm ents outside o f the firm and advised corporate m anagers on the acq uisi­
tion o f extern ally developed technologies.
 K R I S T I N E BRULAND AND DAVI D C. M O W E R Y

A s Pavitt notes in his chapter in this vo lu m e, in -h o u se R & D in U S firm s developed

in parallel w ith independen t R & D lab orato ries that p e rfo rm e d research on a con­
tract basis (see also M o w e ry 1983)* But o ver the co u rse o f the tw entieth century
contract-research firm s 5 share o f in d u strial research e m p lo y m e n t declined. The
co m p lex and uncertain projects u n d ertaken w ith in m a n y in -h o u se research facilities
did not lend them selves to “ arnTs-length” o rg an izatio n .
A s had been the case in G erm an y, the d evelo p m en t o f in d u strial research, as
well as the creation o f a m arket fo r the acq u isitio n an d sale o f in d u strial technologies,
benefited fro m a series o f reform s in U S paten t p o lic y betw een 1890 and 1910
that strengthened and clarified paten th o ld er righ ts (See M o w e ry 1995). Judicial
tolerance fo r restrictive patent licensin g p o licies fu rth er increased the value of
patents in co rp orate research strategies. A lth o u g h the search fo r new patents
p ro vid ed one incentive to p u rsu e in d u strial research, the im p e n d in g expiration of
these patents created anoth er im p o rtan t im petu s. B o th A m erican Telephone
and Telegraph and G eneral Electric, fo r exam p le, establish ed o r expanded their
in -h ou se lab oratories in response to the in ten sified co m p etitive pressure that
resulted from the exp iratio n o f key patents (R eich 1985; M illa rd 19 9 0 :15 6 ). Intensive
efforts to im p ro ve and protect co rp o rate tech n o logical assets w ere com bined with
increased acqu isition o f patents in related tech n o logies fro m oth er firm s and inde­
pendent inventors.
Schum peter argu ed in Capitalism , Socialism and Democracy that in-h ouse indus­
trial research had supplanted the in ven to r-en trep ren eu r (a h yp o th esis supported by
Sch m o ok ler 1957) and w ou ld reinforce, rather than erode, the p o sitio n o f dominant
firm s. The data on research em p lo ym en t and firm tu rn o v er a m o n g the 200 largest
U S m an u factu rin g firm s suggest that d u rin g 19 2 1-4 6 at least, the effects o f industrial
research were consistent w ith his predictions. D isp lacem en t o f these firm s from the
ranks o f the very largest w as sign ifican tly less likely fo r firm s w ith in-house R&D
laboratories (M o w e ry 1983).

13.3*4 Innovation in the Interwar Chemicals Industry

A s w e noted in the previou s section, one o f the critical scien ce-based industries
associated w ith this Second In d u strial R evo lu tio n w as ch em icals. A com parison of
U S and G erm an innovative p erfo rm an ce in this in d u stry h igh ligh ts m an y o f the
p oin ts m ade above concerning the new in stitu tion al an d o rg an izatio n al underpin­
nings o f in n ovation d u rin g this period. A lth o u g h G e rm a n an d U S chem icals firms
had pioneered in the d evelopm ent o f a new stru ctu re fo r in n o va tio n that relied on
in -h ou se R & D and the froutinization o f in n o vatio n ,” these tw o grou p s o f firms
pursued som ew hat different in n ovative strategies d u rin g the in terw ar period
fo llo w in g the creation o f their R & D facilities. T h ese differences highlight the
 INNOVATION THROUGH TIME 365

influence o f cross-nation al contrasts in m arket structure and resource endow m ents,

factors that receded som ew hat in im portance after 1945.
One im portant poin t o f contrast was the quality o f scientific research in chem istry
(as opposed to technological innovation) o f leading firm s and universities in the two
nations. T h rough 1939, G erm an scientists received fifteen out o f the thirty Nobel
Prizes awarded in chem istry, U S scientists received only three, and French and
British scientists each accounted for six. Between 1940 and 1994, U S scientists
received th irty-six o f the sixty-five chem istry Prizes awarded, G erm an scientists
received eleven, British scientists received seventeen, and French scientists received
one (Encyclopaedia B ritannica 19 9 5 :7 4 0 -7 ). A lthough the situation was beginning to
change during the 1930s and w o u ld change dram atically after 1945, the United States
remained a scientific backw ater du ring this p ortion o f the twentieth century.
Technological change in the A m erican chem ical indu stry was shaped b y several
features: (1) the large size and rapid grow th o f the A m erican m arket; (2) the
opportunities afforded b y large m arket size for exploiting the benefits to be derived
from large-scale, co n tin u ou s process produ ction; and (3) a natural resource endow ­
ment— oil and gas— that provided unique opportunities for transform ing the
resource base o f the organic chem ical in d u stry and achieving significant cost
savings, if an app ro priate new tech nology could be developed.
The in troduction and rapid ad option o f the internal com bustion autom obile in
the opening years o f the tw entieth centu ry in the U nited States brought in its wake an
almost insatiable dem and for liqu id fuels. T h is dem and in turn spurred the grow th
o f a new sector o f the petroleu m refining indu stry that was specifically calibrated to
accom m odate the needs o f the autom obile in the first tw o decades o f the twentieth
century. Petroleum refining had tw o im portant, related features. First, it was h igh ly
capital-intensive; b y the 1930s it had becom e the m ost capital-intensive o f all
Am erican industries. Second, produ ctive efficiency required that sm all batch p ro ­
duction, so characteristic o f other chem ical products, such as synthetic organic
materials, be discarded in favor o f h igh -volu m e production m ethods that required
continuous-process technologies. A m erican leadership in petroleum refining p ro ­
vided the critical know ledge and the engineering and design skills to su p port the
chem icals in d u stry’s shift fro m coal to petroleum feedstocks in the interw ar years.
The large size of the American market had introduced American firms at an early
stage to the problems involved in the large-volume production of basic products,
such as chlorine, caustic soda, soda ash, sulfuric acid, superphosphates, etc. The
early American experience with large-scale production contributed to the US
chemical industry’s transition to petroleum-based feedstocks. The dominant par­
ticipants in this industrial transformation were Union Carbide, Standard Oil (New
Jersey), Shell and Dow. But the shift to petroleum benefited as well from the
adoption by US petroleum firms, notably Humble Oil (an affiliate of Standard Oil
of New Jersey), of new techniques for discovery and exploitation of petroleum
deposits. The availability of low-cost petroleum and natural gas, coupled with the
366 K R I S T I N E B R U L A N D A N D D A V I D C. M O W E R Y

new, l«irgc-sc<ilc process in n o vation s d eveloped b y U S firm s, enabled significant

reductions in m an u factu rin g costs.
B y contrast, the G erm an chem icals in d u stry fash io n ed n ew technologies that
com pensated fo r the absence o f such d o m estic feedstocks. D u rin g W orld War II,
G erm an tanks and airplanes were fueled b y syn th etic g aso lin e an d ran on tires made
fro m synthetic rubber derived fro m coal feedstocks. O n ly in the w ake o f W orld War
II an d the creation o f a set o f m u ltilateral in stitu tion s g o ve rn in g international trade
and finance d id the revival o f in tern ation al trad e an d U S guarantees o f access to
foreign petroleum sources su p p o rt a shift b y the G e rm a n ch em icals industry to
p etroleu m feedstocks (Stokes 1994). H ere as w ell as elsew here, th e post-1945 revival
o f international trade and investm ent flow s relaxed so m ew h at the constraints on
tech nological in n o vatio n im posed b y reliance o n d o m e stic n a tu ral resources. As
A b ram o vitz (1994) and other scholars (N elson an d W righ t 1994) have pointed out,
econ o m ic co n d itio n s in E urope and elsew here in the g lo b al e c o n o m y gradually came
to resem ble m ore closely those that had given rise to U S su p re m a c y in this sector.

13.4 A “ T hird In d u s t r i a l R e v o l u t i o n ”?
R&D a n d I n n o v a t i o n d u r i n g
the posT-1945 Period

13.4.1 The Post-war Transformation

The structure o f the in n o vation process in the in d u strial e co n o m ies w as transformed
after 1945. G lo b al scientific leadership shifted d ecisively fro m W estern Europe to the
U nited States. A new set o f industries, focused o n IC T an d b io m ed ical innovation,
grew rapidly. As global trade and investm ent flo w s revived after the 19 14 -4 5 period of
w ar and depression, international flow s o f tech n o lo g y also exp an d ed , and by the
1980s and 1990s enabled econom ies such as Jap an , S o u th K orea, and Taiwan to
advance to the front rank as sources o f in d u strial in n o va tio n . D evelopm ents in the
U nited States illustrate these trends m ost vivid ly, an d h ig h lig h t the developm ent of
a US national innovation system” that contrasted sharply with its 1900-194°
counterpart.
The structure ( if not the scale) o f the p re-19 40 U S R & D system resem bled those of
other leading industrial econom ies o f the era, such as the U n ited K ingdom , Ger­
m any, and France— indu stry was a significant fu n d er an d p e rfo rm e r o f R& D and
central governm ent fu n din g of R & D was m odest. In n o v a tio n an d R & D during the
 INNOVATION THROUGH TIME 367

post-1945 period in the U nited States were transform ed by the dram atic increase in
central governm ent spen ding on R & D , m uch o f which was allocated to indust ry and
academic research. A s Lu ndvall and B orras point out in C hapter 22 o f this volum e,
this expansion in pu blic R & D fu nding was m otivated prim arily by national security
and public-health concerns and secondarily by political support for basic research.
The post-w ar U S R & D system differed from those o f other industrial econom ies in at
least 3 aspects: (1) sm all, new firm s were im portant entities in the com m ercialization
o f new technologies; (2) defense-related R & D fu nding and procurem ent exercised a
pervasive influence in the h igh -tech n ology sectors o f the U S econom y; and (3) US
antitrust p olicy d u rin g the p o st-w ar period was unusually stringent.
The prom inence o f new firm s in com m ercializing new technologies in the post­
war United States contrasts w ith their m ore m odest role during the interw ar period.
In industries that effectively did not exist before 1940, such as com puters, sem icon­
ductors, and biotechnology, new firm s were im portan t actors in R & D and com m er­
cialization, in contrast to po st-w ar Jap an and m ost Western European econom ies.
Moreover, in both sem icond u ctors and com puters, new firm s grew to positions o f
considerable size and m arket share.
Several factors contribu ted to this prom inen t role o f new firm s in the post-w ar U S
innovation system . The large basic research establishm ents in universities, govern­
ment, and a n u m b er o f private firm s served as im portant “ incubators” for the
developm ent o f in n ovation s that “ w alked out the d o o r” w ith individuals w ho
established firm s to com m ercialize them . Relatively w eak form al protection for
intellectual p ro p erty d u rin g the 1945-80 period paradoxically also aided the early
growth o f new firm s. C o m m ercialization o f m icroelectronics and com puter h ard ­
ware and softw are in n o vation s b y new firm s was aided b y a perm issive intellectual
property regim e that facilitated tech nology diffusion and reduced the burden on
young firm s o f litigation over inventions that originated in part w ithin established
firms. In m icroelectronics and com puters, liberal licensing and cross-licensing
policies were b yp ro d u cts o f antitrust litigation, illustrating the tight links between
these strands o f U S governm ent policy. U S m ilitary procurem ent policies also
contributed to the grow th o f new firm s in m icroelectronics and contributed to
high levels o f tech nological spillovers am ong these firm s.

134*2 Electronics and ICT

Advances in electronics tech n o logy created three new industries— electronic co m ­
puters, com pu ter softw are, and sem iconductor com ponents -in the post-w ar
global econom y. T h e electronics revolution can be traced to two key in n o v­
ations— the tran sistor and the com puter. Both appeared in the late 1940s, and the
exploitation o f both w as spurred b y C o ld W ar concerns over national security.
368 K R I S T I N E B R U L A N D AND DAVID C. M O W E R Y

N ew firm s played a p ro m in en t role in the in tro d u c tio n o f n e w p ro d u cts, reflected

in their o ften -d o m in an t share o f m arkets in new se m ic o n d u c to r devices, signifi­
cantly o u tstrip p in g that o f larger firm s. M o reo ver, the ro le o f n ew firm s grew in
im p o rtan ce w ith the developm ent o f the in tegrated circu it (IC ). T h e U S m ilitary’s
w illin gn ess to pu rchase sem icon d u ctor co m p o n en ts fro m u n tried suppliers was
accom p an ied b y co n d itio n s that effectively m an d ated su b stan tial tech n o logy trans­
fer and exchange am o n g U S sem icon d u ctor firm s. To red u ce the risk that a system
designed aro u n d a particu lar IC w o u ld be delayed b y p ro d u c tio n p ro b lem s or by the
exit o f a supplier, the U S m ilitary requ ired its su p p liers to d evelo p a “ second source”
for the product, i.e. a dom estic pro d u cer that co u ld m a n u fa ctu re an electronically
and fu n ctio n ally identical produ ct. To c o m p ly w ith secon d so u rce requirements,
firm s exchanged designs and shared sufficient process k n o w led g e to ensure that the
co m p o n en t produ ced b y a second source w as id en tical to the o rig in a l product. These
requirem ents spurred interfirm “ sp illo vers” o f k n o w led ge and k n o w h o w within the
sem icon d u ctor industry.
The d evelopm ent o f the U S co m p u ter in d u stry also b en efited fro m C old War
m ilitary spending, but in other respects the o rig in s an d early years o f this industry
differed fro m sem iconductors. D u rin g the w ar years, the A m e ric a n m ilitary spon­
sored a n u m b er o f projects to develop h igh -sp eed calcu lato rs to solve special
m ilita ry problem s. T h e E N IA C — generally co n sid ered the first fu lly electronic US
digital co m p u ter— w as fu nded b y A rm y O rdn an ce, w h ich w as concerned with the
co m p u tatio n o f firin g tables for artillery. From the earliest d ays o f th eir support for
the developm ent o f com pu ter technology, the U S arm ed forces w ere anxious that
technical in fo rm atio n on this in n o vation reach the w id est p o ssib le audience. The
technical plans fo r the m ilitary-sp o n so red IA S co m p u te r w ere w id ely circulated
am o n g U S governm ent and academ ic research institutes, an d sp aw n ed a num ber of
“ clon es” (e.g. the IL L IA C , the M A N IA C , A V ID A C , O R A C L E , an d JO H N IA C — see
Flam m , 1988: 52).
A lth ou gh business dem and for com pu ters g rad u ally exp a n d ed d u rin g the early
1950s, govern m en t pro cu rem en t rem ained crucial. T h e p ro jected sale o f fifty ma­
chines to the federal governm ent (a substantial p o rtio n o f the total forecast sales of
250 m achines) influenced IB M ’s decision to initiate the d evelo p m en t o f its first
business com puter, the 650 (Flam m 1988). N ew firm s played a m u ch m ore modest
role in the early years o f the business co m p u ter in d u stry, h ow ever, in part because
established firm s such as IB M had developed p o w erfu l m ark etin g organizations for
electrom echanical business equipm en t sales that w ere (w ith som e difficulty)
adapted to selling com puters to their business cu sto m ers (U sselm an 1993). The
appearance o f new m arkets for com puters that resulted fro m the developm ent o f
scientific com puters and m in icom pu ters d u rin g the 1950s an d 1960s and desktop
system s d u rin g the 1970s and 1980s, how ever, p ro vid ed o p p o rtu n itie s for entry by
m an y new firm s, such as D igital E qu ip m en t, C ra y System s, C o m p u te r Research
A ssociates, and Apple.
 INNOVATION THROUGH TIME 369

The progress o f co m p u ter technology since the 1950s has been driven by the
interaction o f several trends: (1) dram atic declines in the price—perform ance ratios
o f com ponents, inclu ding central processing units and such essential peripherals as
data storage devices; (2) resulting in part from (1), the rapid extension o f com puting
technology into new applications; and (3) the increasing relative costs o f software.
Throughout the b rie f h isto ry o f the electronic com puter industry, these trends have
created bottlenecks that have influenced the path o f technological change. The I BM
360 m ainfram e com puter, for exam ple, w hich cem ented IB M 's dom inance o f the US
com puter in d u stry d u rin g the 1960s and 1970s, created a 'p ro d u c t fam ily” o f
com puters in different perform an ce and price classes that all utilized a com m on
operating system and other software.
The Intel C o rp o ra tio n s com m ercialization o f the integrated circuit m icroproces­
sor in 1971 tran sform ed the structure o f the U S com puter indu stry during the next
twenty-five years. D evelopm ent o f the m icroprocessor at Intel resulted from a search
for an integrated circuit that could be used in a w ide range o f applications. Rather
than designing a cu stom "ch ip set” for each application, the m icroprocessor m ade it
possible for Intel to p rodu ce a pow erful, general-purpose solution to m any diverse
applications, breakin g a bottleneck that lim ited technological progress and
diffusion.
The m icrop rocessor was the fou n d ation for a new generation o f com puting
technology that tran sfo rm ed the IC T industry. D esktop com puters diffused rapidly
throughout m ost indu strial econom ies, and the sheer size o f the resulting "installed
base” radically changed the econom ics o f the com puter hardw are and software
industries. C o m p u te r softw are, fo rm erly dom inated by relatively sm all independent
vendors o r subsidiaries o f established com puter system s producers, becam e a
m ass-m arket in d u stry that supported entry b y a flood o f new firm s, som e o f
which proved to be en o rm o u sly profitable. Established producers o f com puter
hardware w ere severely affected b y the rise o f desktop system s, which encroached
rapidly on m arkets fo r m ain fram e and m inicom puters. B y the end o f the twentieth
century, o n ly a few o f the d om in an t com pu ter system s firm s o f the 1970s rem ained
active.
The final dramatic transformation in computing technology, the Internet,
depended on the appearance of the desktop computer (see Box 13.4). The develop­
ment of computer networking and the Internet began during the 1960s and was
sponsored by governments in the United States, France, and the United Kingdom.
The large-scale and early commitment to deployment of computer networks in the
United States, which relied entirely on public funds, as well as the rapid diffusion of
desktop computers in the United States during the 1980s, created a huge domestic
network by the late 1980s. Although the HTTP and HTML "hypertext” software
protocols were invented at the European research center, CERN, in 199L A e A st
commercial application of these protocols as part of a "Web browser” emerged from
a U S u n iversity in 1993.
370 K R I S T I N E BRULAND AND DAVID C. MOWERY

Box 13.4 The Internet

^ ................ _ _ _ _ ..................... ....■ ........................ 1........................... — if - v—

A collection o f independent but interconnected networks built and managed by a

variety o f organizations, the Internet owes its success to institutional as well as
technological innovations (Mowery and Simcoe 2002 ). Research in computer net*-
working was supported by governments in France, the United Kingdom, and the
United States during the 1960s. A central goal o f Defense Department com pute-
networking research was to enable more effective use by researchers in universities,
government, and industry o f the small number o f large research computers then
available.
Research and early experiments in computer-networking technology in all three
nations led to the development o f prototype networks. But the ARPANET, deployment
o f which in 1969 was sponsored by the US Defense Advanced Research Projects Agency,
was far larger and linked more diverse groups o f researchers than the prototype
networks in France and the United Kingdom. Computers attached to the ARPANET
“ backbone” communicated on the basis o f a shared set o f protocols (TCP/IP), another
outcome o f DARPA research. Later policy decisions by the National Science Founda­
tion (NSF) and other federal agencies that shared responsibility for the backbone
encouraged standardization o f Internet infrastructure. These agencies also promoted
expansion o f the Internet beyond the science and engineering communities. In 1990,
the US Defense Department transferred managerial control over the Internet infra­
structure to NSF, and five years later NSF transferred responsibility for the core
network to the private sector.
Software protocols and architectural elements critical to the Internet had been
placed in the public domain from the beginning. Open standards encouraged expan­
sion by making available the details o f core innovations and lowering entry barriers for
firms that supplied hardware, software, and networking services. State and federal
regulation o f telecommunications aided the rapid diffusion o f the Internet in the
United States by maintaining low, time-insensitive rates. The 1982 settlement of the
federal government s antitrust suit against AT8cT restructured the US telecommuni­
cations industry and encouraged entry by new service providers, spawning further
innovation. But through the late 1980s, the Internet was used m ainly by researchers
from the academic, industrial, and government communities throughout the world.
Key inventions (HTML and HTPP, developed by Berners-Lee) from CERN, the
European particle-physics installation, were used by US technology developers (a
group o f graduate students in computer science at the University o f Illinois, among
whom the best-known is Marc Andreesen, who moved to Netscape to commercialize a
browser based on the MOSAIC technology developed at the University) to produce the
first browser in 1994, which vastly expanded use o f the Internet and led to the
W orldwide Web. The large “ installed base” o f desktop computers in the United
States was a powerful impetus to the “ user-led” innovation that quickly produced a
vast array o f new applications and eventually, a speculative bubble in the equity
markets. But the radical innovation” o f the W orldwide Web in fact represented a
culmination o f nearly thirty years o f research and innovations in networking protocols
and software, as well as high-speed data transmission, routers, and computer process­
ing and memory technologies.
 INNOVATION THROUGH TIME

Com m ercial exp loitation o f the Internet proceeded m ost rapidly In the United
States during the rem aind er o f the 1990s. Like previous U S innovations in ICT, the
Internet's developm ent d rew on a m ix o f public and private R & D funding, as well as
entrepreneurial new firm s seeking to com m ercialize new applications and an active
(and entrepreneurial) c o m m u n ity o f university researchers. M oreover, the Internet's
commercial developm ent, as well as the developm ent o f desktop systems and
software in the U nited States, benefited from the enorm ous size o f the U S m arket
and the large dom estic installed base o f desktop com puters that was rapidly co n ­
nected to the Internet. In n o vatio n in the Internet benefited from the participation
o f millions o f users in d eveloping o r refining new applications, just as had been the
case with desktop co m p u ter softw are d u rin g the 1980s. In other w ords, the develop­
ment o f one o f the m a jo r industries o f the late twentieth and early twenty-
first centuries benefited fro m the large size o f the U S dom estic m arket, a central
feature o f U S econom ic developm ent that has been prom inen t since the nineteenth
century.

134.3 Innovation in Pharmaceuticals and Biotechnology

World W ar II initiated a tran sition in the U nited States to a pharm aceutical industry
that relied on form al, in -h o u se research and on stronger links w ith U S universities
that were also m o vin g to the forefront o f research in the biom edical sciences. The
post-war era in the U S ph arm aceuticals in d u stry opened w ith a w idespread expect­
ation in the in d u stry that there existed a vast potential m arket for new pharm aceut­
ical products, and that catering to this m arket, how ever costly, w ould prove to be
highly profitable.
The post-w ar perio d also w itnessed a rem arkable expansion o f federal support for
biom edical research th ro u gh the huge grow th in the budget o f the N ational In sti­
tutes o f H ealth. B etw een 1950 and 1965, the N IH budget for biom edical research grew
by no less than 18 per cent per year in real term s. B y 1965, the federal governm ent
accounted fo r alm ost tw o -th ird s o f all spending on biom edical research. A lthough
N IH funding has co n tin u ed to grow rapidly, it has been outstripped by grow th in
privately funded R & D since the 1960s. The U S Pharm aceutical M anufacturers
Association estim ated that foreign and U S pharm aceuticals firm s invested m ore
than $26 billion in R & D in the U nited States in 2002, substantially above the $16
billion R & D investm ent b y the N ation al Institutes o f H ealth in the sam e year (see
Pharm aceutical M an u factu rers A ssociation 2003, for both estim ates).
The identification of the double helix structure of DNA by Watson and Crick in
1953 eventually set off a new epoch of technological change in the pharmaceuticals
industry. Biotechnology has created new techniques for drug discovery, as well as
new techniques for production of existing drugs, such as insulin (Henderson et al.
 K R I S T I N E BR U L A N D A N D D A V I D C. M O W E R Y
372

1998). T h e bio tech n o lo g y enterprise w as su p p o rte d b y h u ge fed eral expenditures on

r & D , in clu d in g the N ixo n A d m in istratio n s W ar o n C a n c e r o f the early 1970s* As
M o w e ry and Sam p at p o in t out in their ch ap ter (8) o n u n iversities in this volume,
b io tech n o lo g y has been a key area o f u n iv e rsity -in d u stry research collaboration, as
w ell as un iversity patenting and licensing, since the 1970s. A n d as P ow ell and Grodal
(C h . 3) note, m uch o f the in n o vation process in the “ n e w ” p h arm aceu ticals industry
that has been triggered b y the rise o f b io te c h n o lo g y relies o n co llab o ratio n between
the long-establish ed m ajo r ph arm aceu ticals firm s, w h o retain stro n g capabilities in
m arketing an d m anagem ent o f the co m p le x reg u la to ry p ro cess, an d new firm s that
specialize in b io tech n o logy-b ased d ru g discovery.

13.4.4 A New “ Resource Base55 for Innovation

T h e creation o f new industries in the p o st-w a r U S e c o n o m y illu strates a fundam en­
tal change in the nature o f the U S resource e n d o w m en t an d its relationship to
tech nological in n o vation . The trajecto ry o f in n o va tio n in the U n ited States for
m u ch o f the 19 0 0 -19 4 5 p erio d was influenced b y e x p lo ita tio n o f the nation's abun­
dan t natural resource endow m ent. D u rin g the p o st-19 45 p e rio d , how ever, a com­
b in atio n o f factors shifted U S in n o vatio n fro m a n atu ral reso u rce-in ten sive path of
d evelopm ent to one that m o re inten sively exp loited a b u rg e o n in g U S “ endowm ent”
o f scientists and engineers, derived fro m b oth d o m estic an d fo reign sources. The
scale o f U S m arkets rem ained im p o rtan t, b u t even this so u rce o f n atio n al advantage
becam e less significant outside o f the IC T sector.
T h e p o st-w ar electronics and b io tech n o lo g y in d u stries d id n o t rely on domestic
endow m ents o f natural resources. But the d evelo p m en t o f these indu stries assuredly
did benefit fro m the abundance o f scientific an d en g in eerin g h u m a n capital in the
p o st-w ar U nited States, as well as an unu su al m ix o f p u b lic an d private dem and for
electronics technologies. The creation o f an in stitu tio n al in frastru ctu re during this
cen tu ry that, b y the 1940s, w as capable o f train in g large n u m b e rs o f electrical
engineers, physicists, m etallurgists, m ath em atician s, ch em ists, b iologists, doctors,
and other experts capable o f ad van cin g these new tech n o logies, m eant that the post­
w ar A m erican endow m ent o f specialized h u m an cap ital w as in itia lly m o re abundant
than that o f other indu strial nations.
This shift from natural to “ created” resource en d o w m en ts as critical factors in
international co m petition has o b v io u sly not been co n fin ed to the United States.
Indeed, N elson and W right (1994) am o n g others (A b ram o v itz 1994; see also Ch. 19 in
this volu m e by Fagerberg and G o d in h o ) argue that the early p o st-w a r economic
advantages enjoyed by the U nited States have been ero d ed b y the w idespread invest­
ments by governm ents in edu cation and the d evelo p m en t o f stro n g domestic
R8cD infrastructures. In the po st-w ar era, the resource base fo r knowledge-based
 INNOVATION THROUGH TIME 373

industries in electronics, no less than in chem icals, pharm aceuticals, or autom obiles,
has been transform ed.
Natural resources p er se now play a less central role, particularly b y com parison
with dom estic stocks o f h u m an capital that can be expanded through pu blic invest­
ments in education and train in g (see Ch. 7 b y Edquist in this volum e). In addition,
many o f the historic scale-based advantages enjoyed b y U S firm s in m anufacturing
industries were eroded d u rin g the post-1945 period by the revival o f international
flows o f trade and capital that m ade it possible for sm aller nations to achieve scale
economies through the exp ort o f their products. Expanded trade and capital flows
have also spurred grow th in cross-b ord er flow s o f technology and know h ow (see
Ch. 12 b y N aru la an d Z an fei in this volu m e). This transform ation m ade it possible
for nations such as Taiw an and South Korea, w ith low pre-1945 levels o f industrial
development (in contrast to Jap an , which had created a relatively sophisticated,
albeit m ilitarized, in d u strial base b y the 1930s) and relatively m odest endow m ents o f
natural resources, to “ catch u p ” w ith the industrial econom ies (see Ch. 19 by
Fagerberg and G o d in h o in this volu m e).

13.5 C o n c l u sio n

H istory rarely presents neat lessons for generalization, and the historical study o f
innovation is no exception. T h e p rim a ry lessons from o u r historical discussion
concern the heterogen eity o f the in n ovation process across tim e, across sectors,
and across countries. M u ch o f the su rvivin g historical evidence that guides the
historical study o f in n o vatio n tends to highlight the form al and obscures the
inform al processes o f know ledge accum ulation, learning, and dissem ination that
underpin technological change and that contribute to its econom ic benefits. A n
im portant area fo r fu rth er research is the enrichm ent o f o u r historical understand­
ing o f the in fo rm al processes for know ledge accum ulation and diffu sion that have
been neglected in h istorical research on the First Industrial R evolution in p articu ­
lar. In addition , as we p o in ted out in o u r in tro d u cto ry discussion, a key h isto rio ­
graphical m ystery rem ains— w hy w as N orthw est Europe the locus o f the first
transition to sustained, in n ovation -led grow th, rather than A sia o r som e other
region o f the global econom y? M uch o f the discussion o f this age-old question relies
heavily on sw eeping generalizations, and m ore research on the (asserted) failure of
non-European econ om ies to m ake the transition to sustained econom ic grow th
during this early p erio d is needed.
374 K R I S T I N E B R U L A N D A N D D A V I D C. M O W E R Y

O ne o f the d efin in g characteristics o f in n o va tio n th ro u g h tim e is change in the

structure o f the in n ovation system s that in flu en ce the d evelo p m en t and dissemin­
ation o f know ledge and innovations. T h e in n o va tio n system characteristic o f the
First In d u strial R evo lu tio n relied on a craft-o rien ted , tria l-a n d -e rro r process, in
w hich fam iliarity w ith basic w o o d w o rk in g an d m e talw o rk in g tech niques was valu­
able. In asm u ch as dem an d factors app ear to h ave In flu en ced the u p su rge o f innov­
ation on a b road fron t d u rin g this perio d , the in stitu tio n a l chan ges that laid the
g ro u n d w o rk fo r grow th in incom es and the e xp a n sio n o f m arkets fo r consumer
good s also w ere im po rtan t.
B y contrast, the scale and o rgan ization al co m p le x ity o f the in n o va tio n system that
characterizes the “ Second In du strial R evo lu tio n ” are va stly greater. A new system of
in n o vatio n em erged, pioneered b y G erm an and U S firm s in the electrical equipment
and chem icals industries, characterized b y organ ized in n o v a tio n activities in large
firm s interacting w ith a p u blic R & D in frastru ctu re. T h e in n o v a tio n process during
this p erio d was institutionalized w ith in large en terprises o f u n p receden ted scale, and
the Second In du strial R evo lu tio n in the U n ited States relied h e a v ily on the creation
o f a national m arket o f great size and h om ogen eity.
The “ T h ird In du strial R evo lu tio n ” defies su m m a ry d escrip tio n , b u t its character­
istic in n o vatio n system relied on the state to an even greater extent than the Second
In du strial R evolution. T h e role o f the state as R & D fu n d er an d (in m an y cases) “ first
cu sto m er” fo r the h igh -tech n olo gy indu stries that d evelo p ed d u rin g the Cold War
also contributed to another novel feature o f the in n o v a tio n system o f the Third
In d u strial R evolu tion that is h igh lighted in PavitTs ch ap ter in this v o lu m e -
increased co llab oration and interaction am o n g d ifferen t in stitu tio n s. State actions
also contributed to the spread o f in n o vatio n -led d evelo p m en t to A sia, as the military
alliances and econom ic institution s o f the p o st-19 45 p e rio d su p p o rte d the expanded
international trade and capital flow s that w ere in d isp en sab le to “ catch -u p ” In
add ition , A sian govern m en ts5 strategies fo r tech n o lo g y tran sfer and industrial
developm ent were o f great im p o rtan ce (see C h . 19 b y F ag erb erg and G odinho in
this volum e).
Recent historical research stresses the w id e d istrib u tio n o f in n o va tio n within the
in du strializin g econom ies o f the F irst In d u strial R e vo lu tio n , in m a n y cases involving
sectors overlooked b y the previou s h istorical research that has em phasized the “ key
sectors o f steam and textiles. T h is ch aracterization o f the p ro cess o f technological
in n o vation has not been sufficiently integrated in to co n cep tu al an d theoretical work
in in n o vation studies. The integration o f recent h isto rical evid en ce w ith the broader
conceptual fram ew orks em ployed in the field o f in n o v a tio n stu d ies represents an
im p o rtan t task for future research. M ore research also is needed on the factors
u n d erp in n in g in n ovation in sectors other than the “ lead in g in d u stries” o f the First
In dustrial R evolution. A lth ou gh the patent data in dicate that co n su m er goods were
an im p o rtan t focus for such in n o vation , the factors trig g erin g this upsurge remain
p o o rly understood. Sim ilarly, the d evelopm en t o f o rgan ized in n o va tio n during the
 INNOVATION THROUGH TIME

Second Industrial R evolution outside o f the electrical equipm ent and chem icals
industries has received far too little attention.
Institutional change in areas ranging from m anagerial hierarchies and enterprise
control to the training o f scientists and engineers is o f central im portance to
innovative perform an ce and to change in the structure o f innovation system s over
time. In m any cases, these institutional changes have occurred in response to
pressure from innovators and entrepreneurs, and are best characterized as “ coevol­
ving” with change in indu stries and technologies (see Engerm an and S o k o lo ff 2003
for a useful discussion o f this po in t). The transform ation o f the G erm an and US
university system s, as w ell as the m uch m ore lim ited restructuring o f British
universities, is but one exam p le o f this coevolution; the developm ent o f intellectual
property rights system s is another. But the conditions under which such political
pressure arises, as w ell as the factors contributing to the success o f failure o f such
pressure, rem ain p o o rly u nd erstood .
Although it n o w is w id ely celebrated as a hallm ark o f tw enty-first-century
“knowledge-based econom ies,” science-based innovation is in fact a relatively recent
development. Indeed, it appears w ell after the institutionalization o f R8cD within
industry in the early tw entieth cen tu ry in the U nited States and Germ any. M oreover,
even “ h igh -techn ology” sectors such as biotech nology and sem iconductors co n ­
tinue to rely o n experim ental m ethods that at their heart are “ trial and error”
approaches (see Pisano 1997; H atch and M o w ery 199S).
Our approach has been lim ited both in tim e and geographical coverage. Perhaps
the two m ost im p o rtan t o m ission s fro m this account are the spread o f industrializa­
tion beyond G erm an y w ith in nineteenth-century Europe, and a full account o f the
rise o f the A sian econom ies after 1945 (see C h. 19 b y Fagerberg and G odinh o in this
volume for a fuller accoun t o f the process o f econom ic “ catch-up” in A sia). In the
first o f these cases it is im p o rtan t to rem em ber that the sm aller European econom ies,
now am ong the w ealthiest in the w orld, have benefited from the im port and
adaptation o f tech n ology d u rin g the nineteenth-century, and that Ireland since
1970 has enjoyed rap id grow th fro m sim ilar sources. The post-1945 grow th o f
Japan and Korea was o rigin ally based on altogether different scales and types o f
innovative in du strialization — adaptation o f foreign technologies, largely in such
mature industries as au tom obiles, steel, and sh ipbuilding (see also C h. 19 in this
volume). But both o f these episodes highlight the im portance o f broad Institutional
change, rather than the “ strategic im p o rtan ce” o f any single indu stry or technology,
in m uch the sam e w ay as do the three “ Industrial R evolutions described in this
chapter.
 K R I S T I N E B R U L A N D A N D D A V I D C. M O W E R Y

N o tes _ _ _ ________________________

i. Chapter 2 by Lazo nick in this volume also discusses the role o f the institutional environ­
ment in contributing to firm-level innovative capabilities,
2 See Peer Vries (2002) for an overview and criticjue o f the literature in this field,
3. “ [Scientific explanations proved to be reliable guides to the commercial development of
new processes and products. Unlike the unrestrained inventions o f myth and fable, they
could not be ignored by industrial firms except at the risk o f being displaced by rival firms.
But to understand and apply scientific explanation required years o f training in the
theology o f an invisible pantheon o f scientific entities. That requirement professionalized
industrial science and diminished the role o f artisan invention” (Rosenberg and Birdzell
1986: 253).
4. This discussion draws on Beer (1959), Murmann (1998), and Murmann and Landau
(1998).
5. See Murmann (1998, 2003) for a more detailed discussion.

R e fer en c es

A br a m o v itz , M. (1994), “ The Origins o f the Postwar Catch-Up and Convergence Boom,” in
J. Fagerberg, B. Verspagen, and N. von Tunzelmann (eds.), The Dynamics of Technology,
Trade and Growth, Aldershot: Edward Elgar, 21-52.
B e e r , J. J. (1959), The Emergence of the German Dye Industry, Urbana, 111.: University of
Illinois Press,
B e r g , M. (1998), “ Product Innovation in Core Consumer Industries in Eighteenth Century
Britain,” in M. Berg and K. Bruland (eds.), Technological Revolutions in Europe: Historical
Perspectives, Cheltenham: Edward Elgar, 138-58.
------ and E g er , E. (2003), Luxury in the Eighteenth Century, N ew York: Palgrave.
B e y e r c h e n , A. (1988), “ On the Stimulation o f Excellence in Wilhelmian Science” in J. R.
Dukes and J. Remak, Another Germany: A Reconsideration of the Imperial Era, Boulder,
Colo,: Westview Press, 139-68.
B ra u d el , F. (1984), The Perspective of the World, Berkeley: University o f California Press.
B ru la n d , K. (1989), The Transformation o f Work in European Industrialization,” in P.
Mathias and J. A. Davis (eds.), The Nature of Industrialization: The First Industrial
RevolutionsуOxford: Basil Blackwell, 154—70.
(2004), Industrialisation and Technological C hange,” in R. Flou d and P. Johnson
(eds), The Cambridge Economic History of Modern Britain, vol. 1, Industrialisation, Cam­
bridge: C am bridge U niversity Press, 117-46.
C h a n d ler , A. D., Jr. (1977), The Visible Hand, C am b rid ge, M ass.: H arvard University Press.
C h a pm a n , S. D. (1972), The Cotton Industry in the Industrial Revolution, Lon don : Macmillan.
D e a n e , R (1965), The First Industrial Revolution, C am b rid ge: C am b rid ge University Press.
Encyclopaedia Britanmca (1995), 15th ed n „ v o l 8 , C h icago: E n cyclo p aed ia Britannica.

Asterisked items are suggestions for further reading.

 INNOVATION THROUGH TIME

Engerman, S. L., and Sokoloff, K. L. (2003), “ Institutional and Non-institutional Explan­

ations of Economic D iffere n cesN B E R Workin g Paper 9989.
Flamm, K. (1987), Targeting the Computer, Washington, DC: Brookings Institution.
------(1988), Creating the Computer>Washington, DC: Brookings Institution.
Freeman, C , and LougA, F. (2002), As Time Goes By: From the Industrial Revolution to the
Information Revolution, Oxford and New York: Oxford University Press.
Griffiths , T., H u n t , P. A., and O 'B r ie n , P. K. (1992), “ Inventing Activity in the British
Textile Industry, 1700-1800," Journal of Economic History 52: 881-906.
Haber , L. F. (1971), The Chemical Industry, 1900-1930, Oxford: Clarendon Press.
^Harris , R. (2000), Industrializing English Law: Entrepreneurship and Business Organization
1/20-1844, Cambridge: Cambridge University Press.
Hatch, N., and M o w ery , D. C. (1998), “ Process Innovation and Learning by Doing in
Semiconductor Manufacturing," Management Science 44:1471-7.
Henderson , R „ O r sen ig o , L., and P isan o , G. (1999), “ The Pharmaceutical Industry and
the Revolution in Biotechnology," in D. C. Mowery and R. R. Nelson (eds.), The Sources of
Industrial Leadership, New York: Cambridge University Press, 267-311.
Hudson, P. (1986), The Genesis of Industrial Capital: A Study of the West Riding Wool Textile
Industry c. 1/50-1850, Cambridge: Cambridge University Press.
*Jones, E. L. (2003), Growth Recurring. Economic Change in World History, Ann Arbor:
University o f Michigan Press.
Landes, D. (1998), The Wealth and Poverty of Nations, London: Little, Brown and Co.
Langton, J. (1984), “ The Ecological Theory o f Bureaucracy: The Case of Josiah Wedgwood
and the British Pottery Industry," Administrative Science Quarterly 29: 330-54.
*Lazonick , W. (2002), “ Innovative Enterprise and Historical Transformation," Enterprise
and Society 3: 3-47.
Lloyd-Jones , R., and L e w is , M. J. (1998), British Industrial Capitalism since the Industrial
Revolution, London: U CL Press.
McK endrick , N. (1961), “ Josiah Wedgwood and Factory Discipline," H isto ric a l Jo u r n a l 4(1):
^0—55,
*MacL eod, C. (1988), Inventing the Industrial Revolution: The English Patent System 1660-
1800, Cambridge: Cambridge University Press.
Malerba , F. (1985), The Semiconductor Business, Madison, Wis.: University o f Wisconsin
Press.
Mantoux, P. (1961), T h e In d u s tr ia l R e v o lu tio n in the E ig h teen th C en tu ry : A n O u tlin e o f the
Beginnings of the Modern Factory System in E n g la n d , revised ed. (English translation
originally published in 1928). New York: Harper & Row.
Mathias , P. (1983), The First Industrial Revolution: An Economic History of Britain 1 / 0 0 -19 14 >
2nd edn., London and New York: Methuen.
Meyer -T hurow , G. (1982), “ The Industrialization o f Invention: A Case Study from the
German Chemical Industry," Isis 73: 363-91.
M illard , A. (1990), Edison and the Business of Innovation, Baltimore: Johns Hopkins
University Press.
*M okyr, J. (2002), T h e G ifts o f A th e n a : H is to ric a l O rig in s o f the K n o w led g e Econom y,
Princeton and Oxford: Princeton University Press.
■-----(2003), “ Industrial Revolution," in J. Mokyr (ed.), T h e O x fo rd E n cy clo p ed ia o f E co n o m ic
History, v o l 3, Oxford: Oxford University Press.
378 K R I S T I N E B R U L A N D A N D D A V I D C. M O W E R Y

M ow ery, D. C> (1983)5 “ Industrial Research, Firm Size, Growth, and Survival, 1921—46,
Journal of Economic History 43: 953-80.
*____ (1995), “ The Boundaries o f the U.S. Firm in R&D,” in N. R. Lamoreaux and D. M. G.
Raff (eds.), Coordination and Information: Historical Perspectives on the Organization of
EnterpriseyChicago: University o f Chicago Press for NBER.
------ (1996), The International Computer Software Industry: A Comparative Study of Industry
Evolution and Structure, Oxford: Oxford University Press.
------ (1999), “ The Computer Software Industry” in D. C. M owery and R. R. Nelson (eds.),
Sources of Industrial Leadershipy New York: Cambridge University Press, 133—68.
------ and R o se n b e r g , N. (1998), Paths of Innovation: Technological Change in 20th Century
Americay New York: Cambridge University Press.
------ and S im co e , T. (2002), “ The History and Evolution o f the Internet,” in B. Steil,
R. Nelson, and D. Victor (eds.), Technological Innovation and Economic Performancey
Princeton: Princeton University Press, 229-64.
M u rm a n n , J. R (1998), “ Knowledge and Competitive Advantage in the Synthetic Dye
Industry, 1850-1914,” unpublished Ph.D. dissertation, Columbia University,
------ (2003), “ The Coevolution o f Industries and Comparative Advantage: Theory and
Evidence,” unpublished MS.
------ and L a n d a u , R. (1998), “ On the Making o f Competitive Advantage: The Development
o f the Chemical Industries o f Britain and Germany Since 1850,” in A. Arora, R. Landau, and
N. Rosenberg (eds.), Chemicals and Long-Term Economic Growthy New York: Wiley, 27-70.
N elso n , R. R., and W r ig h t , G. (1994), “ The Erosion o f U.S. Technological Leadership as a
Factor in Postwar Economic Convergence,” in W. J. Baumol, R. R, Nelson, and E, N. Wolff
(eds,), Convergence of Productive New York: Oxford University Press.
O rsen ig o , L, (1988), The Emergence of Biotechnology, London: Pinter.
P h a r m a c eu tic a l M a n u fa c tu r er s A sso ciatio n (2003), Industry Profile 2003 http://
www.phrma.org/ publications/publications/profileo2
P isa n o , G. (1997), The Development Factoryy Boston: Harvard Business School Press,
------ S h a n , W , and T e e c e , D. J. (1988), “ Joint Ventures and Collaboration in the Biotech­
nology Industry,” in D. C. Mowery (ed.), International Collaborative Ventures in L/.S.
Manufacturingy Cambridge, Mass.: Ballinger.
P o llard , $. (1965), The Genesis of Modern Management A Study of the Industrial Revolution
in Great Britiany London: Arnold,
^P o m eran z , K. R. (2000), The Great Divergence: China, Europe and the Making of the Modern
World Economy, Princeton: Princeton University Press.
R e ic h , L. S. (1985), The Making of American Industrial Research, New York: Cambridge
University Press,
R e id , T, R. (1984), The Chip, New York: Simon and Schuster.
R o sen ber g , N. (1985)» The Commercial Exploitation o f Science by American Industry,” in
K. B. Clark, R. H. Hayes, and C. Lorenz (eds.), The Uneasy Alliance, Boston: Harvard
Business School Press, 19-51.
and B ir d z e l l , L. E. (1986), How the West Grew Rich, New York: Basic Books.
S a m u el , R. (1977), The Work-shop ot the World: Steam Power and Hand Technology in
mid-Victorian Britain,” History Workshop 3: 6-72.
Sch m o o kler , J. (1957), Inventors Past and Present,” Review of Economics and Statistics 57:
 INNOVATION THROUGH TIME 379

___ (1962),uChanges in Industry and in the State of Knowledge as Determinants of Indus­

trial Invention” in R. R. Nelson (ed.), The Rate and Direction of Inventive Activity;
Princeton: Princeton University Press for NBER.
Sto kes , R. (1994)* Opting for Oil: The Political Economy of Technological Change in the West
German Chemical Industry, 1945-бу New York: Cambridge University Press.
Su lliv a n , R. (1990), “ The Revolution o f Ideas: Widespread Patenting and Invention During
the Industrial Revolution,” Journal of Economic History 30(2): 340-62,
T ilton , J, E. (1971), The International Diffusion of Technology: The Case of Transistors,
Washington, DC: Brookings Institution.
T o yn b ee , A, (1969; orig, 1908), Lectures on the Industrial Revolution in England, New York:
A. M. Kelley.
U glow , J. (2002), The Lunar Ment New York: Farrar, Straus & Giroux.
U ssel m a n , S. (1993), “ IBM and its Imitators: Organizational Capabilities and the Emergence
of the International Computer Industry,” Business & Economic History 22:1-35.
von T u n z e l m a n n , G. N. (1978), Steam Power and British Industrialization to i860, Oxford:

Clarendon Press.
*V r ie $, P, (2002), “Are Coal and Colonies Really Crucial? Kenneth Pomeranz and the Great
Divergence,” Journal of World History 12(2): 407-45.
W a d e , R. (1999), Governing the Market: Economic Theory and the Role of Government in East
Asian IndustriaizationyPrinceton: Princeton University Press.
W a l l e r s t e in , I. (1974), Capitalist Agriculture and the Origins of the Modern World Economy;
New York: Academic Press.
W h it e h e a d , A. N. (1925), Science and the Modern World, New York: Macmillan.
 CHAPTER 14

SECTORAL SYSTEM S
HOW A N D W H Y
INNOVATION DIFFERS
ACROSS SECTORS

FRANCO MALERBA

14.1 I n t r o d u c t i o n

I nnovation greatly differs across sectors in terms o f characteristics, sources, actors

involved, the boundaries o f the process, and the organization o f innovative activ­
ities. A focus on a representative firm as the main actor; on narrowly, well defined,
and static boundaries as a sector delimitation; on R8cD and learning-by-doing as the
on у two sources of innovation, on competition and formal R & D joint ventures as
e on у kind of interaction among firms; and on the patent system and public
support for R & D as the only relevant institutions and policies that matter for
innovation, would capture only part of the action that takes place in sectors and
would identify only a few of the key variables that matter for innovation and
performance.
i ff comPar‘son °f actors, sources, institutions, and policies for innovation in
erent sectors (e.g inpharmaceuticals and biotechnology, chemicals, software,
computers, semiconductors, telecommunications, ormachinetools) showsstriking
 S E C T O R A L S Y S T E M S OF I N N O V A T I O N 38і

differences. The role o f in n o vation in the dynam ics and transform ation o f these
sectors is highly diverse.
How is it possible to analyze consistently these differences and their effects on
sectoral growth and perform ance? The industrial econom ics approach pays a lot o f
attention to differences across sectors in R8cD intensity, m arket structure, the range
of viable R & D strategies and R & D alliances, the intensity o f the patent race, the
effectiveness o f patent protection , the role o f com petition policy and the extent o f
R&D support. But, w hile these are very im p o rtan t factors, they are not the only ones
nor are they the m ost relevant for a full understanding o f the differences in in n o v­
ation across sectors.
A rich and heterogeneous trad ition o f sectoral studies has clearly shown both that
sectors differ in term s o f the know ledge base, the actors involved in innovation, the
links and relationships am o n g actors, and the relevant institutions, and that these
dimensions clearly m atter fo r u nderstandin g and explaining innovation and its
differences across sectors. H ow ever, these case studies are quite different in terms
of methodology, variables, and countries exam ined.
This chapter w ill b riefly discuss the previous literature on differences across
sectors in innovation (Section 14.2) and then propose the concept o f sectoral systems
of innovation (14.3). In the next sections, the basic building blocks o f sectoral
systems will be discussed: know ledge, technological dom ains, and sectoral b o u n d ­
aries (14.4); actors, relationships and netw orks (14.5); and institutions (14.6). Then
the dynam ics and tran sfo rm atio n o f sectoral system s (14.7) is exam ined. Finally,
some policy im plication s (14.8) and the challenges ahead (14.9) are discussed.
The chapter w ill discuss a large n u m ber o f sectors that are highly innovative and
technologically advanced and have strong links w ith science, w hich nevertheless
organize innovation v e ry differently: com puters, sem iconductors, telecom m unica­
tion equipm ent and services, softw are, chem icals, pharm aceuticals and biotechnol­
ogy, and m achine tools. M ost o f the sectoral exam ples in this paper are draw n from
M owery and N elson (1999) and M alerb a (2004).

14.2 P r e v i o u s L i t e r a t u r e on S e c t o r a l
D i f f e r e n c e s in In n o v a t io n

The literature has advanced some distinctions among sectors in innovation and
diffusion based on different dimensions. The simplest one, widely used in inter­
national studies by the OECD, EU, and international organizations, refers to sectors
ЗВі FRANCO MALERBA _____________ ___________________________________ —

that are high R&D-intensive (such as electronics or drugs) and low RfkD-intemive
(such as textiles or shoes).
A n oth er d istin ction , co m in g fro m the Sch u m p eterian legacy, focuses on differ­
ences in m arket structure and in d u strial d yn am ics a m o n g sectors. Sch u m p eter Mark
I sectors are characterized b y "creative destru ctio n , w ith tech n o lo g ical ease o f entry
and a m ajo r role played b y entrepreneurs and n ew firm s in in n o vative activities.
Sch u m p eter M a rk II sectors are characterized b y 'c re a tiv e a c c u m u la tio n ” (in Keith
Pavitt s w ords) w ith the prevalence o f large establish ed firm s an d the presence o f
relevant barriers to en try for new in n ovators. T h is regim e Is characterized by the
d om in an ce o f a stable core o f a few large firm s, w ith lim ited entry. T h e distinction
refers to the early Schu m p eter o f Theory o f Economic Development (19 11, "Sch u m p­
eter M a rk I” ) and to the later one o f Capitalism > Socialism and Democracy (1942,
"S ch u m p e ter M a rk II” ). M ach in ery o r b io tech n o lo g y are exam p les o f Schumpeter
M a rk I sectors, w hile the sem icon d u ctor in d u stry o f the 1990s (th in k o f m icro­
processors and d yn am ic m em ories) o r m a in fra m e c o m p u te rs in the period
19505-19905 are exam ples o f Schu m peter M a rk II sectors.
O ther differences across sectors have been related to tech n o lo g ical regimes, a
n o tio n introduced b y N elson and W inter (1982), referrin g to the learning and
know ledge en viron m en t in w h ich firm s operate. A sp ecific tech n o logical regime
defines the nature o f the prob lem firm s have to so lve in th eir in n o vative activities,
affects the m odel fo rm o f tech nological learn in g, shapes the incen tives and con­
straints to particu lar b eh avio r and o rgan ization , an d in flu en ces the basic processes
o f variety generation and selection (and therefore the d y n a m ics and evolution o f
firm s). M ore generally, M alerb a and O rsenigo (1996 an d 1997) have p ro p o sed that a
technological regim e is com posed b y o p p o rtu n ity and a p p ro p ria b ility conditions,
degrees o f cum ulativeness o f tech nological kn ow ledge, an d ch aracteristics o f the
relevant know ledge base. M o re specifically, tech n o logical o p p o rtu n itie s reflect the
likelih ood o f in n o vatin g for any given am o u n t o f m o n e y invested in search. High
o p p o rtu n ities p ro vid e p ow erfu l incentives to the u n d erta k in g o f in n ovative activ­
ities and denote an econom ic en viron m en t that is n o t fu n c tio n a lly constrained by
scarcity. In this case, potential innovators m ay com e up w ith freq u en t and im portant
tech nological innovations. A p p ro p ria b ility o f in n o vatio n s su m m arizes the possibil­
ities o f protectin g in n ovation s fro m im itation and o f reap in g p ro fits fro m innovative
activities. H igh ap p ro p riab ility m eans the existence o f w ays o f su ccessfu lly protect­
ing in n o vation from im itation. L o w ap p ro p riab ility co n d itio n s d en o te an economic
en viron m en t characterized b y the w idespread existence o f extern alities (Levin et al.
1987)* C um ulativeness con dition s capture the p ro p erties that to d a y ’s innovations
and innovative activities form the starting p o in t fo r to m o rro w in n o vation s, More
broadly, one m ay say that high cum ulativeness m eans that to d a y ’s in n o vative firms
are m o re likely to innovate in the future in specific tech n o logies an d along specific
trajectories than n o n -in n ovative firm s. C u m u lativen ess m a y be d u e to knowledge/
cognitive factors, organizational factors, or m arket factors o f the "su ccess breeds
 S E C T O R A L S Y S T E M S OF I N N O V A T I O N 383

success” type. The properties o f the know ledge base relate to the nature o f knowledge
underpinning firm s innovative activities. Technological know ledge involves various
degrees o f specificity, tacitness, com plem entarity, and independence and m ay
greatly differ across sectors and technologies (W inter 1987). Differences in techno­
logical regimes affect the organization o f innovative activities at the sectoral level and
may lead to a fundam ental d istin ction between Schum peter M ark I and Schum peter
Mark II m odels. H igh technological opportu nities, low appropriability, and low
cumulativeness (at the firm level) conditions lead to a Schum peter M ark I pattern.
By contrast, high a p p ro p riab ility and high cum ulativeness (at the firm level) co n d i­
tions lead to a Schum peter M ark II pattern: think again o f the sem iconductor
industry o f the 1990s (i.e. m icroprocessors and dynam ic m em ories) and m ainfram e
computers in the p erio d 19505-19905.
Technological regim es and Schum peterian patterns o f innovation change over
time (Klepper 1996). A cco rd in g to an in d u stry life-cycle view, a Schum peter M ark I
pattern o f innovative activities m ay turn into a Schum peter M ark II. E arly in the
history o f an in d u stry— w hen know ledge is changing very rapidly, uncertainty is
very high, and barriers to en try very low — new firm s are the m ajor innovators and
are the key elem ents in in du strial dynam ics. W hen the indu stry develops and
eventually m atures and technological change follow s w ell-defined trajectories,
economies o f scale, learning curves, barriers to entry, and financial resources becom e
important in the com petitive process. Thus, large firm s w ith m onopolistic pow er
come to the forefron t o f the in n o vation process (U tterback 1994; G o rt and K lepper
1982; Klepper 1996). In the presence o f m ajor know ledge, technological, and m arket
discontinuities, a Schu m peter M ark II pattern o f innovative activities m ay be
replaced by a Schum peter M a rk L In this case, a rather stable organization charac­
terized by incum bents w ith m o n o p o listic pow er is displaced b y a m ore turbulent
one with new firm s using the new technology or focusing on the new dem and
(Henderson and C la rk 1990; C hristensen and R osen bloom 1995)* A lthough rather
archetypical, these analyses p o in t to the direction o f placing a lot o f attention to
differences across sectors in som e key factors related to knowledge and learning
regimes. A s the exam ples discussed above suggest, change over tim e also reflects
institutional change and the co evolu tion o f industries and institutions.
Other distinctions refer to sectors that are net suppliers o f technology and sectors
that are users o f technology. O n the bases o f the R & D done by 400 A m erican firm s
and o f intersectoral flow s in the A m erican econom y, Scherer (1982) identifies
sectors that are net sources o f R & D for other sectors (such as com puters and
instrum ents), and sectors that are net users o f technology (such as textiles
and m etallurgy). A sim ilar analysis is done b y R obson et al. (1988) who, on the
basis o f 4,378 in n o vation s in the U K between 1945 and 1983, identify (a) “ core
sectors” (such as electronics, machinery, instruments, and chemicals) which gener­
ate m ost o f in n o vation s in the econ o m y and are net sources of technology, (b)
secondary sectors (such as auto and metallurgy) which play a secondary role in
384 FR A N CO MALERBA

term s o f sources o f in n ovation for the econ om y, an d (c) u ser sectors su ch as services
w h ich m a in ly absorb technology,
A key difference am o n g sectors refers to the so u rces o f in n o vatio n and the
ap p ro p ria b ility m echanism s. Pavitt (1984) p ro p o ses fo u r typ es o f sectoral pattern
fo r in n o vative activities. In su p p lier-d o m in ated (e.g. textile, services) sectors, new
technologies are em bod ied in n ew co m p o n en ts an d eq u ip m en t, an d the diffusion o f
n ew technologies and learning takes place th ro u g h le a rn in g -b y -d o in g and by-using.
In scale-intensive sectors (e.g. autos, steel), process in n o v a tio n is relevant and the
sources o f in n o vation are both internal (R & D an d le a rn in g -b y -d o in g ) and external
(eq u ipm en t p ro d u cers), w hile a p p ro p ria b ility is o b tain ed th ro u g h secrecy and
patents. In specialized suppliers (e.g. eq u ip m en t p ro d u ce rs), in n o va tio n is focused
on perfo rm an ce im provem en t, reliability, an d cu sto m izatio n , w ith the sources o f
in n o vatio n being b oth internal (tacit know ledge an d exp erien ce o f skilled techni­
cians) an d external (u se r-p ro d u c e r in teractio n ); a p p ro p ria b ility com es m ain ly from
the localized and interactive nature o f know ledge. Finally, scien ce-b ased sectors (e.g.
ph arm aceuticals, electronics) are characterized b y a h igh rate o f p ro d u ct and process
in n o vation s, by internal R & D , and b y scien tific research d o n e at universities and
pu b lic research laboratories; science is a sou rce o f in n o va tio n , an d appropriability
m eans are o f variou s types, ran gin g fro m patents, to lead -tim es an d learning curves,
and to secrecy. The Pavitt taxo n o m y has been trem en d o u sly successful in empirical
research and has guided the identification o f firm s an d c o u n try advantages. Refine­
m ents and enrichm ents o f the ta x o n o m y h ave been p ro p o sed in the following
decades. A v e r y interesting and relevant w o rk in this d irectio n is the on e by Marsili
(2001).
D ifferences across sectors in a p p ro p ria b ility co n d itio n s h ave been exam ined by
Levin et al. (1987), PA C E (1996) and C o h en et al. (2002) u sin g su rve y questionnaires
fo r R & D m anagers in the United States, E u rop e, an d Ja p an , fo llo w in g the pioneering
Yale survey. H ere, m ajo r differences across sectors h ave been id en tified in terms o f
a p p ro p ria b ility m eans— patents, secrecy, lead -tim es, learn in g cu rves, and com ple­
m en tary assets. A ll these surveys have fo u n d m a jo r d ifferen ces acro ss sectors in the
use o f patents.

14*3 S e c t o r a l S y s t e m s o f I n n o v a t i o n

The co n trib u tio n s exam ined above focus on a specific d ifference a m o n g sectors. In
this and the fo llo w in g sections, a m u ltid im en sio n al, integrated , an d d yn am ic view o f
in n o vation in sectors is proposed , related to the fram ew o rk o f sectoral systems o f
 S E C T O R A L S Y S T E M S OF I N N O V A T I O N 385

innovation, which pro vid es a m eth o d o lo gy for the analysis and com parison o f
sectors.
A sector is a set o f activities that are unified by som e linked product groups for a
given or em erging dem and and w hich share som e com m on knowledge. Firm s in a
sector have som e co m m on alities and at the sam e tim e are heterogeneous. A sectoral
system fram ew ork focuses on three m ain dim ensions o f sectors:
(a) Knowledge and tech nological d om ain
( b) Actors and netw orks
(c) Institutions

(a) Knowledge and technological domain. A n y sector m ay be characterized by a

specific knowledge base, technologies and inputs. In a dynam ic way, the focus on
knowledge and the technological d om ain places at the centre o f the analysis the issue
of sectoral boundaries, w hich u su ally are not fixed, but change over time.

(b) Actors and networks. A sector is com posed o f heterogeneous agents that are
organizations or in d ivid u als (e.g. consum ers, entrepreneurs, scientists). O rganiza­
tions m ay be firm s (e.g. users, producers, and input suppliers) or n on-firm s (e.g.
universities, financial institutions, governm ent agencies, trade-unions, or technical
associations), and include subunits o f larger organizations (e.g. R M ) or production
departments) and grou p s o f organizations (e.g. indu stry associations). Agents are
characterized b y specific learning processes, com petencies, beliefs, objectives, o r­
ganizational structures, and beh aviors, w hich interact through processes o f co m ­
munication, exchange, co op eration , com petition, and com m and.
Thus, in a sectoral system fram ew ork, innovation is considered to be a process
that involves system atic interactions am on g a w ide variety o f actors for the gener­
ation and exchange o f know ledge relevant to innovation and its com m ercialization.
Interactions include m arket and non -m arket relations that are broader than the
market for technological licensing and know ledge, interfirm alliances, and form al
networks o f firm s, and often their ou tcom e is not adequately captured b y our
existing system s o f m easu rin g econ om ic output.

(c) Institutions. A gen ts’ cognition , actions, and interactions are shaped by insti­
tutions, which include n o rm s, routines, co m m on habits, established practices, rules,
laws, standards, and so on. In stitu tions m ay range from ones that bind or im pose
enforcements on agents to ones that are created b y the interaction am ong agents
(such as contracts); fro m m ore b in d in g to less binding; from form al to inform al
(such as patent laws o r specific regulations vs. traditions and conventions). A lot o f
institutions are national (such as the patent system ), w hile others are specific to
sectors (such as sectoral lab o r m arkets o r sector specific financial institutions).

Over tim e, a sectoral system undergoes processes o f change and transform ation
through the co evolu tion o f its variou s elem ents.
386 FR A N C O MALERBA

The notion of sectoral system of innovation and production complements other

concepts within the innovation system literature (Edquist 1997) such as national
systems of innovation delimited by national boundaries and focused on the role of
non-firm organizations and institutions (Freeman 19875 Nelson 19935 Lundvail 1993)»
regional/local innovation systems in which the boundary is the region (Cooke et aL
19 9 7) , technological systems, in which the focus is on technologies and not on
sectors1 (Carlsson and Stankiewitz i 9 9 5 >Hughes 19845 Gallon 1992) >and distributed
innovation system (in which the focus is on specific innovations— Andersen et al.
2 0 0 2 )*
W hat are the m ain differences betw een a sectoral in n o v a tio n system and a
n ation al in n o vatio n system perspective? W h ile n atio n al in n o v a tio n system s take
in n o vatio n system s as d elim ited m ore o r less clearly b y n atio n a l boundaries, a
sectoral system app roach w o u ld claim that sectoral system s m a y h ave local, national,
an d /or global dim ensions. O ften these three d ifferen t d im e n sio n s coexist in a sector.
In add ition , national in n o vatio n system s result fro m the d ifferen t com position o f
sectors, som e o f w h ich are so im p o rtan t that they d rive the gro w th o f the national
econom y. Fo r exam ple, Japanese grow th in the 1970s and 1980s w as d riven by specific
sectors, w hich w ere different fro m the sectors b eh in d the A m e ric a n “ resurgence”
d u rin g the 1990s. Sim ilarly, Italian econ om ic grow th is d rive n b y specific sectors.
T h us, un d erstan d in g the key d rivin g sectors o f an e c o n o m y w ith their specificities
greatly helps in un d erstan d in g n ation al grow th and n ation al p attern s o f innovative
activities.
The theoretical and analytical ap p ro ach o f sectoral system s is g ro u n d ed in the
e vo lu tio n ary theory. E vo lu tio n ary th eo ry places a k ey em p h asis o n dynamics,
in n o vation processes, and econom ic tran sfo rm atio n , L earn in g and know ledge are
key elem ents in the change o f the econ om ic system . “ B o u n d e d ly ra tio n a l” agents act,
learn, an d search in uncertain and ch an gin g en viron m en ts, A gen ts k n o w h o w to do
different things in different ways. T h u s learn in g, kn o w led ge, an d b eh avio r entail
agents' heterogen eity in experience and organ izatio n . T h e ir d ifferen t competences
affect their persistent differential p erform an ce. In ad d itio n , e v o lu tio n a ry theory
places em phasis on cognitive aspects such as beliefs, o bjectives, and expectations,
w hich are in turn affected b y previou s learning and exp erien ce an d b y the environ­
m ent in w hich agents act, A central place in the e v o lu tio n a ry a p p ro a ch is occupied by
the processes o f variety creation (in technologies, p ro d u cts, firm s, and organiza­
tions), replication (that generates inertia and c o n tin u ity in the system ), and selec­
tion (that reduces variety in the econ om ic system and d isco u rages the inefficient or
ineffective utilization o f resources). Finally, for e v o lu tio n a ry theory, aggregate
ph enom en a are em ergent properties o f fa r-fro m -e q u ilib riu m interactions and
have a m etastable nature (N elson 1995; D osi 1997; M etcalfe 1998). H ere, the environ­
m ent and co n d itio n s in w hich agents operate m a y d rastically differ. Evolutionary
theory stresses m ajo r differences in o p p o rtu n ities related to science an d technolo-
gies, I he sam e holds for the know ledge base u n d erp in n in g in n o vative activities, as
 S E C T O R A L S Y S T E M S OF I N N O V A T I O N 387

well as for the institutional context. Thus the learning, behavior, and capabilities of
agents are constrained and “ bounded” by the technology, knowledge base, and
institutional context. Heterogeneous firms facing similar technologies, searching
around similar knowledge bases, undertaking similar production activities, and
“embedded” in the same institutional setting, share some common behavioral and
organizational traits and develop a similar range of learning patterns, behavior, and
organizational forms.
One last remark regards the aggregation issue regarding products, agents or
functions. For example, sectoral systems may be examined broadly or narrowly (in
terms of a small set of product groups) . 2 A broad definition allows us to capture all
the interdependencies and linkages in the transformation of sectors, while a narrow
definition identifies more clearly specific relationships. Of course, within broad
sectoral systems, different innovation systems related to different product groups
may exist. The choice of the level of aggregation depends on the goal of the analysis.
In the following pages we will concentrate on each block of a sectoral system of
innovation and production:
• Knowledge, technological domain, and boundaries
• Agents, interaction and networks
• Institutions

14.4 Kn o w led ge, T echnological

D o m a in , and Sectoral Boundaries

Knowledge plays a central role in innovation. Knowledge is highly idiosyncratic at

the firm level, does not diffuse automatically and freely among firms, and has to be
absorbed by firms through their differential abilities accumulated over time. The
evolutionary literature has proposed that sectors and technologies differ greatly in
terms of the knowledge base and learning processes related to innovation. Know-
ledge differs across sectors in terms of domains. One knowledge domain refers to the
specific scientific and technological fields at the base of innovative activities in a
sector (Dosi 1988; Nelson and Rosenberg 1993b while another comprises applica­
tions, users, and the demand for sectoral products. Recently a major discontinuity
has taken place in the processes of knowledge accumulation and distribution with
the emergence of the knowledge-based economy which has redefined existing
sectoral boundaries, affected relationships among actors, reshaped the innovation
process, and modified the links among sectors (Nelson i 9 9 5 >Dost x9 9 7 i Metcalfe
199В; Lundvall 1993; Lundvall and Johnson 1994)*
3§8 FRANCO MALERBA

What d o we know about the main dimensions of knowledge? First, knowledge may
have different degrees of accessibility (Malerba and Orsenigo 2000), i.e. opportunities
of gaining knowledge external to firms, which in turn may be internal or external to
the sector. In both cases, greater accessibility of knowledge m a y decrease industrial
concentration. Greater accessibility internal to the sector implies lo w er appropriabil­
ity: competitors may gain knowledge about new products and processes and, if
competent, imitate those new products and processes. Accessibility of knowledge
that is external to the sector may be related to the levels and sources of scientific and
technological opportunities. Here, the external environment may affect firms through
human capital with a certain level and type of knowledge or through scientific and
technological know ledge developed in firm s o r n o n -firm organizations, such as
universities o r research laboratories (M alerba and O rsenigo 2000).
The sources of technological opportunities differ markedly among sectors. As
Freem an (1982) and R osen berg (1982), am o n g oth ers, h ave sh o w n , in som e sectors
opportunity conditions are related to m ajo r scientific breakthroughs in universities;
in others, opportunities to innovate may often come fro m advancements in R&D,
equipment, and instrumentation; while in still other sectors, external sources of
knowledge in terms of suppliers or users may play a crucial role. Not all external
knowledge may be easily used and transformed into new artifacts. If external
know ledge is easily accessible, tran sform able into n ew artifacts an d exposed to a
lot o f actors (such as custom ers o r su p pliers), then in n o vative e n try m a y take place
(W in ter 1984). I f advanced integration capabilities are n ecessary (C ohen and
Levin th al 1989), the in d u stry m ay be concen trated and fo rm e d b y large, established
firm s.
Second, know ledge m ay be m ore o r less cu m u lative, i.e. the degree b y which the
generation o f new know ledge bu ilds u p o n cu rren t k n o w led ge. O n e can identify
three different sources o f cum ulativeness.

(1) C o gn itive. The learning processes an d past k n o w led ge co n strain current

research, but also generate new questions and n ew kn ow led ge.
(2) The firm and its organizational capabilities. O rg an izatio n al capabilities are
firm -sp ecific and generate know ledge w h ich is h ig h ly path -d ep en d en t. They
im p licitly define what a firm learns and w hat it can h o p e to achieve in the
future.
(3) Feedbacks from the m arket, such as in the “ su ccess-b reed s-su ccess” process.
In novative success yields profits that can be reinvested in R & D , thereby
increasing the p ro b ab ility to innovate again.

H igh cum ulativeness im plies an im plicit m echanism leading to h igh appropriability

o f innovations. In the case o f know ledge spillovers w ithin an industry, however, it is
also possible to observe cum ulativeness at the sectoral level. C um ulativeness m ay also
be present at the local level. In this case, high cum ulativeness w ith in specific locations
is m ore likely to be associated with low ap p ro p riab ility co n d itio n s and spatially
 S E C T O R A L S Y S T E M S OP I N N O V A T I O N 389

localized knowledge spillovers. Finally, cum ulativeness at the technological and firm
levels creates first m over advantages and generates high concentration. Firm s that have
a head start develop a new know ledge based on the current one and introduce
continuous innovations o f the increm ental type.
Accessibility, o p p ortu n ity, and cum ulativeness are key dim ensions o f knowledge
related to the notion o f tech nological and learning regim es (N elson and W inter 1982;
Malerba and O rsenigo 1997), w hich, as seen above, m ay differ across sectors. Other
dimensions o f know ledge could be related to its tacitness, codificability, com plexity,
systemic features, scientific base, and so on (W inter 1987; C ow an, D avid, and Foray
2000).
The boundaries o f sectoral system s are affected by the knowledge base and
technologies. H ow ever, the type and dynam ics o f dem and represent a m ajor factor
in the processes o f tran sfo rm ation o f sectoral system s. The sam e holds for links and
complementarities am o n g artifacts and activities. These links and com plem entar­
ities are, first o f all, o f the static type, as are in p u t-o u tp u t links. Then there are
dynamic com plem entarities, w h ich take into account interdependencies and feed­
backs, both at the dem and and at the p ro d u ction levels. D ynam ic com plem entarities
among artifacts and activities are m ajo r sources o f transform ation and growth o f
sectoral systems, and m ay set in m o tio n virtu ou s cycles o f innovation and change.
This could be related to the concept o f filiere and the notion o f developm ent blocks
(Dahmen 1989). Links and com plem entarities change over tim e and greatly affect a
wide variety o f variables o f a sectoral system : firm s’ strategies, organization, and
performance; the rate and d irection o f technological change; the type o f com peti­
tion; and the netw orks am o n g agents. T h u s the boundaries o f sectoral systems m ay
change more or less rap id ly over tim e, as a consequence o f dynam ic processes related
to the transform ation o f know ledge, the evolution and convergence in dem and, and
changes in com petition and learning b y firm s.
In general, the features and sources o f know ledge affect the rate and direction o f
technological change, the o rgan ization o f innovative and production activities, and
the factors at the base o f firm s’ successful perform ance.
Great differences am o n g sectors in the dim ensions discussed above exist. Let us
compare, fo r exam ple, ph arm aceu ticals and m achine tools. In the pharm aceutical
industry, the know ledge base and the learning processes have greatly affected
innovation and the organ ization o f innovative activities. In the early stages (18 50 -
i945)> the in d u stry w as close to chem icals, w ith little form al research until the 193°$
and a m ajor use o f licenses. T h e fo llo w in g perio d (19 45-early 1980s) was character­
ized by the in tro d u ction o f ran d o m screening o f natural and chem ically derived
com pounds. T h is led to an exp losio n o f RScD and, although few blockbusters were
discovered in each p erio d , nevertheless, each period enjoyed high grow th. The
advent o f m olecular b io lo g y since the 1980s led to a new learning regim e based on
molecular genetics and rD N A technology, w ith tw o search regim es: one regarding
specialized technologies, the other generic technologies. N ow adays, no individual
390 FR A N CO MALERBA

firm can gain control o n m ore than a subset o f the search space* In n o v a tio n increas­
in gly depends on stron g scientific capabilities and o n the a b ility to interact with
science and scientific institution s in o rd er to exp lo re the search space (McKelvey,
O rsenigo, and P am m o lli 2004; H en derson , O rsen igo, an d P isan o 1 999 )*
In m achine tools, innovation has been m a in ly in crem en tal an d n o w is increas­
in gly systemic* K n ow led ge ab ou t ap p lication s is v e ry im p o rta n t, an d therefore user-
p ro d u cer relationships as well as partn ersh ip s w ith cu sto m ers are com m on . The
know ledge base has been em bod ied in skilled person n el o n the sh op flo o r level (with
applied technical q ualification) and in design engineers (n ot n ecessarily with a
u n iversity degree but w ith lo n g -term em p lo ym en t in the co m p an y ). Internal
train in g (p articu larly apprenticeships) is qu ite relevant. In sm all firm s, R8cD is not
done extensively and R8cD co o p eratio n is n ot co m m o n . R ecently, the knowledge
base has shifted fro m pu rely m echan ical to m ech an ic, m icro e lectro n ic and infor­
m atio n intensive, w ith an increasing co d ificatio n and an in creasin g use o f formal
R8cD. Prod ucts have increasingly being m o d u larized an d stan d ard ized . A key role is
also played b y in fo rm atio n flow s ab ou t co m p o n en ts c o m in g fro m producers o f
different technologies, such as lasers, m aterials, m easu rem en t, an d con trol devices.
N ow ad ays, m an y large m ach in e tool com pan ies o p erate alre a d y o n an international
basis m akin g use o f specific know ledge sources at th eir d ifferen t firm sites (Wengel
and Shap ira 2004; M azzoleni 1999).

14.5 A c t o r s , R e l a t i o n s h i p s ,
a n d N etworks

Sectoral system s are co m p o sed o f heterogen eous actors. In general, a rich, m ulti­
d isciplinary, and m u ltisou rce know ledge base an d rap id tech n o lo g ical change im ­
plies a great heterogen eity o f actors in m ost sectors.
Firm s are the key actors in the generation, ad o p tio n , and use o f n ew technologies,
are characterized b y specific beliefs, expectations, goals, co m peten ces, an d organiza­
tion, and are co n tin u o u sly engaged in processes o f learn in g an d know ledge accu­
m ulation (N elson and W inter 1982; M alerb a 1992, Teece an d P isan o 1994, Dosi,
M aren go, and Fagiolo 1998, M etcalfe 1998). T h e extent o f firm heterogen eity is the
result o f the o p p o sin g forces o f variety creation, rep lication , an d selection (Nelson
1995) M etcalfe 1998). Selection increases h om ogen eity, w h ile e n try an d technological
and o rgan ization al in n ovation s are fu n d am en tal sources o f heterogeneity. Firm
heterogeneity is also affected by the characteristics o f the k n o w led ge base, specific
experience and learning processes, and the w o rk in g o f d y n a m ic com plem entarities.
 S E C T O R A L S Y S T E M S OF I N N O V A T I O N
391

Actors also include users and suppliers w ho have different types o f relationships
with the innovating, p ro d u cin g, or selling firm s, Users and suppliers are character­
ized by specific attributes, know ledge, and com petencies, with m ore or less close
relationships w ith p rodu cers (V onH ippel 1988, Lundvall 1993). As previously m en ­
tioned, in a dynam ic and innovative setting, suppliers and users greatly affect and
continuously redefine the bou n daries o f a sectoral system.
Other types o f agents in a sectoral system are n on -firm organizations such as
universities, financial organizations, governm ent agencies, local authorities, and so
on. In various ways, they su p p o rt innovation, technological diffusion, and pro d u c­
tion by firm s, but again their role greatly differs am ong sectoral systems. In several
high technology sectors, universities play a key role in basic research and hum an
capital form ation, and in som e sectors (such as biotechnology and software) they are
also a source o f start-u ps and even innovation. In sectoral system s such as software
or biotech nology-pharm aceu ticals, new actors such as venture capital com panies
have emerged over tim e. These financial organizations have played a different role
according to the stage o f the in d u stry life-cycle. W hen indu stry m atures or large
firms are relevant, capital constraints becom e lighter and m uch investm ent is self-
financed. B y contrast, fo r start-ups in em erging or new high-tech sectors, capital
constraints are v ery high and specific financial interm ediaries such as venture capital
firms are im portan t (R ivau d -D an set 2001; D ubocage 2002).
Often the m ost app ro priate units o f analysis in specific sectoral systems are not
necessarily firm s b u t in d ivid u als (such as the scientist w ho opens up a new biotech ­
nology firm ), firm s’ subunits (such as the R & D or the production departm ent), and
groups o f firm s (such as in d u stry consortia).
The focus on users, govern m en t agencies, and consum ers puts a different em ­
phasis on the role o f dem and. In a sectoral system , dem and is not seen as an
aggregate set o f sim ilar buyers o r atom istic undifferentiated custom ers, but as
com posed o f heterogeneous agents w ho interact in various ways w ith producers.
Demand then becom es com posed b y in divid u al consum ers, firm s, and public
agencies, w hich are in tu rn characterized b y know ledge, learning processes, and
competences, and w hich are affected b y social factors and institutions. The em er­
gence and tran sfo rm ation o f dem and becom e then a very im portant part in the
dynamics and evolu tion o f sectoral system s. In addition, dem and has often proven
to be a m ajor factor in the redefin ition o f the boundaries o f a sectoral system, a
stimulus for in n o vation , and a key factor shaping the organization o f innovative and
production activities.
W ithin sectoral system s, heterogeneous agents are connected in various ways
through m arket an d n o n -m arket relationships. It is possible to identify different
types o f relations, linked to different analytical cuts. First, traditional analyses o f
industrial organizations have exam ined agents as involved in processes o f exchange,
com petition, and co m m an d (such as vertical integration). Second, in m ore recent
analyses, processes o f fo rm al co op eration or in form al interaction am ong firm s or
392 FRANCO MALERBA

a m o n g firm s an d n o n -firm organ ization s h ave been exa m in e d in d ep th (as one may
see fro m the literature o n tacit o r exp licit co llu sio n , o r h y b rid go vern an ce form s, or
fo rm a l R & D co o p eratio n ). T h is literature has an alyzed firm s w ith certain market
pow er, su p pliers o r users facing o p p o rtu n istic b e h a v io r o r asset specificities in
tran saction , an d firm s w ith sim ilar know ledge h av in g a p p ro p ria b ility and indivis­
ib ility problem s in R & D . Finally, the e vo lu tio n a ry a p p ro a c h an d the innovation
system s literature have also paid a lot o f atten tion to the w id e ran ge o f form al and
in fo rm al co o p eratio n an d interaction am o n g firm s. H ow ever, acco rd in g to this
perspective, in uncertain and ch an gin g en viron m en ts n etw o rk s em erge n o t because
agents are sim ilar, b u t because they are different. T h u s, n etw o rk s integrate com ple­
m entarities in know ledge, capabilities, and sp ecializatio n (see Lu n dvall 1993;
E dq u ist 1997; N elson 1995; Teubal et al. 19 9 1). R elatio n sh ip s b etw een firm s and
n o n -firm o rgan ization s (such as universities and p u b lic research centers) have
been a source o f in n o vatio n and change in several sectoral system s: pharm aceuticals
and biotech nology, in fo rm atio n technology, an d tele co m m u n ica tio n s have been
relevant (N elson an d R osenberg 1993).
O ne final o b servation needs to be m ade: the k ey role p layed b y netw orks in a
sectoral system leads to a m ean in g o f the term “ sectoral stru ctu re ” d ifferent from the
o n e used in in d u strial econom ics. In in d u strial eco n o m ics, stru ctu re is related
m ain ly to the concept o f m arket structure and o f vertical in tegratio n an d diversifica­
tion. In a sectoral system perspective, o n the contrary, stru ctu re refers to links am ong
artifacts and to relationships am o n g agents: it is th erefo re far b ro a d e r than the one
based on e xch a n g e -co m p e titio n -co m m a n d . T h u s w e can say th at a sectoral system
is com posed o f w ebs o f relationships a m o n g h etero gen eo u s agents w ith different
beliefs, goals, com petencies, and behavior, and that these relatio n sh ip s affect agents'
actions. T h ey are rather stable over tim e.
In su m m ary, the types and structures o f relation sh ips an d n etw o rk s d iffer greatly
fro m sectoral system to sectoral system , as a co n sequ en ce o f the features o f the
know ledge base, the relevant learning processes, the b asic tech n o logies, the charac­
teristics o f dem and , the key links, an d the d yn am ic co m p lem en tarities. A gain, let's
p ro vid e som e exam ples.
A gain , the co m p ariso n o f fo u r quite d ifferent sectoral system s, such as chemicals,
com pu ters, sem icond u ctors, and softw are, illustrates this p o in t. In chemicals,
the structure o f the sectoral system has been centered a ro u n d large firm s, which
have been the m ajo r source o f in n o vatio n o ver a lo n g p erio d o f tim e. Large
R & D expenditures, econom ies o f scale and scope (C h an d ler 19 9 0 ), cum ulativeness
o f technical advance, and co m m ercialization cap ab ilities have given these firms
m ajo r in n ovative and com m ercial advantages (A rora, et al. 1998). W ith the diffusion
o f the synthetic d yestu ff m odel, firm s scaled up their R & D d ep artm en ts and the
role o f universities increased. T h e in tro d u ction o f p o ly m e r ch em istry (1920s)
affected the structure o f the in d u stry because kn ow led ge ab o u t the characteristics
of different m arket segm ents becam e im p o rtan t, so that firm s had to
 S E C T O R A L S Y S T E M S OF I N N O V A T I O N
393

develop extensive linkages w ith dow nstream markets. The other m ajor change
related to the developm ent o f chem ical engineering and the concept o f unit o f
operation led to an increasing d ivisio n o f labor between chem ical com panies and
technology suppliers, w ith the rise o f the specialized engineering firm s (SEFs), which
developed vertical links w ith chem ical com panies. In this period, university research
continued to be im p o rtan t fo r the developm ent o f innovations, and links between
universities and in d u stry increased. In addition, advances in chem ical disciplines
and the separability o f know ledge increased the transferability o f chem ical technolo­
gies. Thus, there has been a greater role o f licensing also by large firm s, which in turn
increased know ledge diffusion.
In computers, the different stages o f the evolution o f the industry (related to
different products) have been characterized by different actors and networks. H aving
been a typical Schum peter M ark II sector for m ost o f its history (until very recently),
mainframe com puters have always been dom inated by large firm s, with high cum ula­
tiveness o f technical advance. In particular, during the 1960s and 1970s, m ainfram es
were produced and integrated b y vertically integrated firm s, and IB M was the typical
example. IB M was produ cing both com ponents and systems and was active in the
development, m anufacturing, m arketing, and distribution o f large systems and o f
the key com ponents. W hen m inicom puters were introduced, the com puters sector
experienced the entry and grow th o f firm s specialized in com ponents or in systems
(with the early years characterized b y a Schum peter M ark I pattern). The sam e holds
for the early years o f m icrocom puters. Later on, however, com petition became
characterized b y groups o f specialized firm s related to different platform s. Each
platform was characterized b y divid ed technical leadership o f several disintegrated
firms. Innovation becam e decentralized, and the control over the direction by a single
firm became very difficult. Recently, in com puter networks, m odularity and connect­
edness increased the role o f netw orks o f firm s with local developm ent and local
feedbacks (Bresnahan and Greenstein 1999; Bresnahan and M alerba, 1999).
In sem iconductors, the in d u stry has been characterized b y a quite different set o f
actors, ranging fro m m erch ant sem iconductor m anufacturers to vertically inte­
grated producers. T h e types o f actors have been quite different from period to
period and fro m co u n try to co u n try d u rin g the evolution o f the industry. N ew
entrants and specialized p ro d u cers w ere quite relevant in the U nited States, w ith
entrants particu larly high either early on in the h istory o f the indu stry or during
phases o f tech nological discontinu ities (and giving the indu stry a typical Sch u m p ­
eter M ark I fashion in these p eriods o f rapid and radical change). Large, vertically
integrated produ cers w ere m o re co m m on in Japan and Europe (M alerba 1985;
Langlois and Steinmueller 1999). T h u s, in these countries a Schum peter M ark
II m ode characterized the industry. In sem iconductors, other m ain actors have
played a m ajo r role. T h e m ilitary w as one o f the m ajo r factors responsible fo r the
growth o f the A m erican industry, com pared to Europe and Japan, because it
supported the en try of new firms and pro vid ed competent firms w ith a large and
394 FRANCO MALERBA

in n o vative dem and . D u rin g the 1970s in Jap an , M IT I w as a m a jo r facto r in allowing

the Japan ese in d u stry (com p o sed o f large pro d u cers) to close the gap w ith Am erican
p ro d u cers in som e pro d u ct ranges (such as m e m o ry devices).
In softw are, specialization o f b oth global players an d lo cal p ro d u ce rs is present. In
ad d itio n , the ch an gin g know ledge base has created an e v o lv in g d iv isio n o f labor
am o n g users, “platform” developers, an d specialized so ftw are v en d o rs (Bresnahan
an d G reenstein 1998). The sectoral system o f in n o v a tio n in so ftw are, however, is
incom plete w ith o u t the ad d ition o f com pan ies that utilize these p la tfo rm s to deliver
en terprise-critical app lication s. M a n y o f these ap p licatio n s co n tin u e to be produced
in -h o u se b y o rgan ization s using the tools p ro v id e d as p a rt o f the platform or
available from the d evelopm ent tools m arkets (Stein m u eller 2004).

14.6 In s t i t u t i o n s

In all sectoral system s, institution s p lay a m a jo r role in affectin g the rate o f techno­
logical change, the organ izatio n o f in n o vative activity, an d p erfo rm an ce. T h ey may
em erge either as a result o f deliberated plan ned d ecision b y firm s o r oth er organiza­
tions, o r as the unpredicted consequence o f agents' in teraction .
Som e institution s are sectoral, i.e. specific to a sector, w h ile oth ers are national.
T h e relationship betw een national in stitu tion s and sectoral system s is quite im port­
ant in m ost sectors. N ation al in stitu tion s have d ifferen t effects o n sectors. For
exam ple, the patent system , p ro p e rty rights, o r an titru st regu latio n s have different
effects as a consequence o f the different features o f the system s, as surveys and
em p irical analyses have sh ow n (see fo r exam p le L evin et al. 1987). H ow ever, the same
in stitu tion m a y take d ifferent features in d ifferent cou n tries, an d thu s m ay affect the
sam e sectoral system differently. F o r exam ple, the w e ll-k n o w n d iv ersity between the
first-to -in ven t and the first-to -file rules in the patent system s in the U nited States
and in Jap an had m ajo r consequences o n the b eh avio r o f firm s in these two
countries. O ften, the characteristics o f n ation al in stitu tio n s fa v o r specific sectors
that fit better the specificities o f the national in stitu tion s. T h u s, in certain cases,
som e sectoral system s becom e p red o m in an t in a c o u n try becau se the existin g insti­
tu tions o f that co u n try p ro vid e an en viron m en t m o re su itab le fo r certain types o f
sectors and not for others. For exam ple, in France, sectors related to pu b lic dem and
have grow n co n sid erab ly (C hesnais 1993). In o th er cases, n atio n al in stitu tion s m ay
constrain d evelopm ent o r in n o vation in specific sectors, o r m ism atch es between
national and sectoral institutions and agents m ay take place. T h e exam p les o f the
different types o f interaction betw een n ation al in stitu tion s and sectoral evolution in
various advanced countries in D osi and M alerb a (1996) are cases in poin t.
 S E C T O R A L S Y S T E M S OF I N N O V A T I O N
395

The relationship between national institutions and sectoral systems is not always
one-way, as it is in the case of the effects of national institutions on sectoral variables.
Sometimes, the direction is reversed, and goes from the sectoral to the national level.
In fact, it may occur that the institutions of a sector, which are extremely important
for a country in terms of employment, competitiveness, or strategic relevance, end
up emerging as national, thus becoming relevant for other sectors. But in the process
of becoming national, they may change some of their original distincti ve features.
Again, m ajor differences em erge across sectors, as in the case o f pharm aceuticals,
software, m achine tools, and telecom m unications, for exam ple. In pharm aceuticals,
national health system s and regulations have played a m ajor role in affecting the
direction o f technical change, in som e cases even blocking or retarding innovation.
In addition, patents have played a m ajo r role in the app ropriability o f the returns
from innovations. In softw are, standards and standard setting organizations are
important, and IP R play a m ajo r role in strengthening appropriability. However, the
emerging open source m ovem en t aim s to create a new segm ent o f the software
industry which is characterized by new d istribution m ethods and b y cooperative
production activities based on vo lu n ta ry association. This has reduced the possibil­
ity o f m aintaining p ro p rietary control over data structure, thus inducing entry and
more com petition (Stein m u eller 2004). In m achine tools, internal and regional
labor markets and local institution s (e.g. local banks) have played a m ajor role in
influencing international advantages o f specific areas. Trust-based, close relation­
ships at the regional level have over a long tim e ensured a sufficient financing o f the
innovation and o f the exp an sion plans o f fam ily businesses in G erm any and Italy
(Wengel and Shap ira 2004). Finally, in telecom m unications, the roles o f regulation,
liberalization/privatization, an d standards have been o f m ajor im portance in the
organization and perfo rm an ce o f the sector. A s discussed in D alum and V illum sen
(2001), liberalization and p rivatization have had m ajor effects on the behavior and
performance o f incum bents and have transform ed the structure o f the industry. A n
example o f the role o f institution s is given b y G S M in Europe.

14*7 T h e D y n a m i c s a n d T r a n s f o r m a t i o n
of S e c t o r a l S y s t e m s

As m entioned above, at the base o f the dynam ics and tran sform ation o f sectoral
systems lies the in terp lay am o n g evo lu tio n ary processes (such as variety creation,
replication, and selection) that d iffer fro m sector to sector (N elson 19955 M etcalfe
1998). Processes o f variety creation may refer to products, technologies, firms, and
institutions, as well as firm strategies and behavior and could take place through
entry, R8cD, innovation, and so on (Cohen and M alerba 2002). Sectoral systems
differ extensively in the processes o f variety creation and o f heterogeneity among
agents. The creation o f new agents— both new firm s and non-firm organizations—
is particularly important for the dynamics o f sectoral systems. As examined by
Audretsch (1996) and Geroski (1995), am ong others, the role o f new firms differs
drastically from sector to sector (in terms o f entry rates, com position, and origin),
and thus has quite different effects on the features o f sectoral systems and their
degree o f change. Sectoral differences in the level and type o f entry seem to be closely
related to differences in the knowledge base; level, diffusion and distribution of
competences; the presence o f non-firm organizations (such as universities and
venture capital); and the working o f sectoral institutions (such as regulations or
labor markets) (Audretsch x996; M alerba and Orsenigo 1999; M cKelvey 1997; Ger­
oski 1995). Processes o f selection play the key role o f reducing heterogeneity among
firm s and may drive out inefficient or less progressive firm s. T h ey m ay refer to
products, activities, technologies, and so on. In addition to market selection, in
several sectoral systems non-m arket selection processes are at w ork, as in the cases of
the involvement o f the military, the health system, and so on. In general, selection
affects the growth and decline o f the various groups o f agents and the range o f viable
behaviors and organizations. Selection m ay greatly differ across sectoral systems in
terms o f intensity and frequency. Theoretical w ork (see M etcalfe 1998) and empirical
work on “ competence destroying” innovation, industrial dynam ics, firm s’ entry and
exit, and mergers and acquisitions have shed light on several aspects o f selection.
Changes in sectoral systems are the result o f coevolutionary processes o f their
various elements, involving knowledge, technology, actors, and institutions. Nelson
(1994) and Metcalfe (1998) have discussed these processes at the general level by
focusing on the interaction between technology, industrial structure, institutions,
and demand. These processes are sector-specific and often path-dependent. Here,
local learning, interactions among agents, and networks may generate increasing
returns and irreversibilities that may lock sectoral systems into inferior technolo­
gies.3 In addition, the interaction between knowledge, technology firm s, and insti­
tutions are also shaped by country-specific factors.
In general, one could say that changes in the knowledge base and in the relevant
learning processes of firms induce deep transform ations in the behavior and struc­
ture of the agents and in their relationships between one another. Overall market
competition and market structure depend on the strategies and fortunes of individ­
ual companies, which are linked to different national contexts o r to the international
scene. Firms have diverse reactions in order to try to increase their fit and to survive
in their particular environment. These environm ents keep changing, not least due to
innovations and choices made by all the constituent competitors: some of these
environm ents are national, others increasingly international.
 S E C T O R A L S Y S T E M S OF I N N O V A T I O N
397

Over the past decades, com puters have had m ajor r e v o lu t io n a r y processes, quite
different from one another. In m ainfram es, coevolution has been characterized by
large systems requiring u se r-p ro d u c e r relationships, centralization o f user in fo r­
mation systems, and extensive sales and service efforts b y large vendors. M arket
structure was h igh ly concentrated and suppliers were vertically integrated. A d o m ­
inant design (IBM /360) em erged in the grow th phase o f the segment and a market
leader (IBM ) dom inated the in d u stry early on, with a coordinating role over the
platform and the ability to steer the direction o f technical change. The U S govern­
ment played a role in early su p p o rt for technological exploration and was a m ajor
buyer o f early com puters. In m in icom pu ters and m icrocom puters, coevolution has
been characterized b y tech nological change focused on dedicated applications in the
case o f m inicom puters o r on system s that increased ease o f use and a lower price/
performance ratio (in the case o f m icrocom pu ters). The relationships w ith custom ­
ers have required m uch less post-sales effort and service. M arket structure was
characterized by high entry early on, and then by increasing concentration in
platforms in both m in icom p u ters and m icrocom puters. In com puter networks,
connectivity and co m p atib ility led to m odular, open, and m ultiform client/server
platforms. Technical change follow s a variety o f directions w ith an upsurge in the
number o f potential technologies associated w ith the relevant platform s. Interde­
pendencies and externalities have increased. D ivided technical leadership has
emerged, in that no single firm has been able to govern change and coordinate
platform standards.
This exam ple is quite different from coevolution in other sectoral systems. In
pharmaceuticals, the nature o f the process o f drug discovery (discussed in Section
14.4) had im portant consequences on the patterns o f com petition and on m arket
structure. U ntil the m o lecu lar b io lo g y revolution, dom inant firm s persisted as
leaders. The m o lecu lar b io lo g y revolution induced deep changes in the incentive
structures w ithin firm s and universities, w ith the advent o f university spin-offs and
the emergence o f the specialized new b iotech n ology firm s. In this process o f
adaptation and change, d ifferent dynam ic processes led to different patterns
o f com petition and p erform an ces (M cKelvey, O rsenigo, and Pam m olli 2004).
In telecom equipm ent and services, the early separation o f the radio spectrum for
use in one-w ay broadcastin g and tw o -w ay telephony gave rise to an oligopolistic
structure that persisted fo r quite a long tim e (D alum and V illum sen 2001). The
convergence w ithin IC T and betw een IC T and b roadcasting-audio-visual and the
emergence o f the Internet origin ated a m ore fluid m arket structure w ith a lot o f
different actors w ith different specializations and capabilities, and new types o f
users. This in turn greatly expanded the boundaries o f the sector b y creating new
segments and new o p p ortu n ities, and also by creating national differences in the
organization o f in n ovation . M oreover, the em ergence o f the Internet has generated
more pressure in favor o f open standards and has led to the rise o f new actors (such
as ISP and content p ro vid ers). In softw are, since the early 1980s, the spread o f
39^ FRANCO MALERBA

netw orked co m p u tin g, em bedded softw are, the In tern et, th e d evelo p m en t o f open-
system architecture and o pen source, and the gro w th o f w e b -b ased netw ork com-
p u tin g has led to the decline o f large co m p u ter p ro d u ce rs as d evelo p ers o f integrated
h ard w are an d softw are system s and to the em ergence o f a lo t o f specialized software
com panies. A lso, softw are d istrib u tio n has greatly ch an ged , fro m licensing agree­
m ents in the early days, to the rise o f in d ep en d en t so ftw are ven d o rs, to price
d isco u n ts fo r package softw are, and, w ith the d iffu sio n o f th e C D -R O M and the
Internet, to sh arew are and freew are (this last on e p a rtic u la rly relevan t w ith Linux)
(D ’A d d erio 20 0 1). In m ach in e too ls, a m a jo r d riv in g fo rce fo r coevolutionary
processes is the d em an d fro m advanced cu sto m er sectors, n a m e ly the autom otive,
aeronautics, and defense indu stries, and the in creasin g use o f electron ic devices.
The em ergence o f new clusters that span several sectors, su ch as internet-soft­
w a re-tele co m , b io tech n o lo g y -p h a rm a ceu ticals, an d n ew m aterials, is one o f the
m ost relevant current tran sfo rm atio n processes in sectoral system s. H ere a great role
is played b y the integration an d fu sio n o f p re v io u sly sep arated know ledge and
tech nologies an d b y the new relations in v o lv in g u sers, co n su m ers, firm s with
d ifferent specializations and com petences, an d n o n -firm o rg an izatio n s and insti­
tutions grou n d ed in p revio u sly separated sectors.

14.8 P o l ic y Im p l i c a t i o n s

A sectoral system o f in n o vatio n ap p ro ach p ro vid es a d esign fo r in n ovation and

tech n ology policies. W ithin a system o f in n o va tio n fra m e w o rk , id en tifyin g defi-
ciencies in the fu n ctio n in g o f a system is the sam e as id e n tify in g those systemic
d im en sio n s that are m issin g o r in ap p ro p riate o r n o t w o rk in g an d w h ich lead to a
p ro b lem in term s o f co m p arative perfo rm an ce. W h en w e k n o w the causes behind
a certain p ro b lem ” — fo r exam ple, w eak tech n o logical tran sfer b etw een universities
and in d u stry we have identified a “ system fa ilu r e ” N o t u n til they know the
character o f the system failure can p o licy m akers k n o w w h eth er to influence or to
change organ ization s or institution s o r the in teractio n s betw een them . Therefore, an
id en tification o f a prob lem sh ou ld be su p plem en ted w ith an an alysis o f its causes as
part o f the analytical basis for the design o f an in n o v a tio n policy. Benchm arking is
not enough.
1 hus a sectoral system app roach pro vid es the id en tificatio n o f “ system failures”
and the related variables w hich sh ou ld be p o licy targets. S ecto ral analyses should
fo cu s on system ic features in relation to know ledge an d b o u n d a rie s, heterogeneity o f
actors and netw orks, institution s and tran sfo rm atio n (th ro u g h coevolutionary
 S E C T O R A L S Y S T E M S OF I N N O V A T I O N
399

processes), A s a consequence, the understanding o f these dim ensions becom es a

prerequisite for any p o licy addressed to a specific sector.
Given the m ajo r differences am on g sectoral system s exam ined in this chapter, the
impact o f general or h orizontal policies m ay drastically differ across sectors, because
the channels and w ays policies have their effects differ from sector to sector. For
example, cooperation and netw orks or n on -firm organizations and institutions
could have different relevance in different sectors. Therefore, policies affecting
networks or n o n -firm organizations, such as transfer agencies, have to take these
differences into account.
In addition, a sectoral system approach em phasizes that, for fostering innovation
and diffusion in a sector, tech nology and innovation policies m ay not be enough.
A wide range o f other policies m ay be necessary. Innovation and technology policy
could be supplem ented b y other types o f policies, such as science policy, industrial
policy, policies related to standards and IP R , and com petition policy. This point
highlights the im portan ce o f the interdependencies, links, and feedbacks am ong all
o f these policies, and their com bined effects on the dynam ics and transform ation o f
sectors.
Relatedly, a sectoral system approach em phasizes that p olicy m akers being w ithin
a variety o f netw orks are an active internal (part) o f sectoral system s at different
levels. In fact, the p o licy m akers intervene actively in know ledge creation, IPR,
corporate governance rules, tech n ology transfer, financial institutions, skill fo rm a­
tion, and public procurem ent. A s a consequence, they have to develop advanced
competences and create an institutional setting in order to be effective and consist­
ent at the variou s different levels.
Finally, p o licy has to consider the coexistence o f different geographical d im en ­
sions o f sectoral system s. D evelopm ents in the local, national, regional, and global
levels influence the articu lation o f technological capabilities. Policies that focus on
only one level are likely to m iss constraints or opportu nities that are influential in the
innovative b eh avio r o f in d ivid u al organizations.
The em phasis on the d iversity o f sectoral system s highlights also different p o licy
measures for different sectors. In fact, p o licy needs are closely related to the problem s
faced b y the v ario u s actors operatin g in the sectoral contexts and to the sectoral
specificity o f know ledge, b ou n daries, actors, and networks.
In sum , trad ition al in n o vation policies have been form ulated as p ro vid in g public
resources for R & D and changing the incentives for firm s to innovate. Tax breaks fo r
R&D , innovation subsidies, and patents are typical exam ples o f these policies.
A sectoral system perspective does not deny the significance o f this approach. It
recognizes, how ever, that the effects m ay run rapidly into d im inishing returns. To
offset this, it is necessary that in n ovation opportu nities be enhanced. Im provin g the
organization o f an in n o vation system w ithin a sector is an alm ost certain route to
im proving the co m p lem en tary payoffs from pu blic and private R & D . The sectoral
perspective pro vid es a too l for p o licy m akers to com prehend the differences in
400 F R A N C O MALERBA

In n o vatio n system s and fo r id en tifyin g the sp ecific acto rs that sh o u ld b e influenced

b y policy. T h e q u id pro q u o, how ever, is that p o lic y m ak ers need to invest much
m o re effort in un d erstan d in g the id iosyn crasies o f the sp ecific sectors that they use
to channel the influence o f po licy (E d q u ist et al. 2004).

14*9 T h e C h a l l e n g e s A h e a d

T h is chapter has claim ed that in n o vatio n greatly d iffers acro ss sectors in terms
o f sources, actors, features, b ou n d aries an d o rg an izatio n . It has p ro p o sed an inte­
grated an d co m p arative w ay to lo o k at sectors based o n the sectoral systems
fram ew ork.
Som e rem arks have to be advanced here in w ay o f co n clu sio n . T h e discussion o f
sectoral system s has show n that there cou ld be several levels o f sectoral aggregation,
an d that the choice o f one depends on the goal o f the analysis. W h ile the discussion
here has been very b ro ad in term s o f sectors in o rd e r to em ph asize linkages,
interdependencies, and tran sfo rm ation , fo r d ifferen t research goals the level o f
d isaggregation co u ld be m uch higher, at the level o f p ro d u c t g ro u p s. Still, we may
talk ab ou t system s o f in n o vation in this respect.
G eograp h ical bou n daries are a key d im en sio n to be co n sid ered in analyses o f
sectoral system s. N ation al b ou n d aries are n ot alw ays the m o st ap p ro p riate ones
fo r an exam in atio n o f structure, agents, an d co evo lu tio n . O ften , sectoral systems
are h igh ly localized and frequently define the sp ecializatio n o f lo cal areas (as in the
case o f m achinery, som e trad ition al indu stries, and even in fo rm a tio n technology).
For exam ple, m ach in ery is concentrated in specialized region al areas. Sim ilarly
sectoral specialization and local agglo m eratio n have o verlap p ed in R ou te 128 (for
m in icom p u ters) and in Silicon V alley (fo r p erson al co m p u te rs, softw are, and
m icroelectronics) (Saxenian 1994). M oreover, in the co n text o f transnational
econ om ic integration, the sector m ay m atter as m u ch o r m o re th an the national
system .
D ifferences across countries in sectoral system s have been relevant and have
affected countries international p erform an ce. In general, on e co u ld claim that
those coun tries that did not have effective sectoral system ch aracteristics did not
p erfo rm w ell in international m arkets. T h e sam e h old s fo r th o se co u n tries that tried
to replicate the success o f w orld leaders b y m im ick in g so m e o f the features o f the
sectoral system s o f the leading countries, w ith o u t h avin g the ap p ro p riate set o f
actors, linkages, and institution s. B y contrast, those co u n tries that have tried to
specialize in subsectors w ith products, know ledge, an d in stitu tio n al requirem ents
 S E C T O R A L S Y S T E M S OF I N N O V A T I O N 401

that match their specific institutional fram ew ork have been successful (Coriat,
Malerba, and M o n to b b io 2004),
Finally, this chapter has tried to show h ow relevant a sectoral system approach is
for an understanding o f the features, determ inants, and effects o f innovation, in
terms o f research and policy. The p olicy aspect has been discussed in the previous
section and will not be repeated here, but research on sectoral systems m ay prove
very fruitful and has to m ove along several lines o f advancem ent.

(1) A sectoral system fram ew o rk m ay allow for detailed analyses o f innovation in

sectors in term s o f know ledge and learning processes, structure (where
structure is seen here as a netw ork o f relationships), and institutions. In
addition, a sectoral system approach provides a w ay to exam ine the dynam ics
o f sectors due to in n o vation and technological change and the r e v o lu t io n a r y
processes taking place am o n g know ledge, technology, actors, and institu­
tions. D ifferent sectoral system s m ay be com pared along sim ilar dim ensions
(in order to try to iden tify sim ilarities across sectors), and the sam e sectoral
system m ay be exam ined across different countries (in order to focus on the
interplay between sectoral and national variables).
(2) The specific m echanism s, causal relationships, and interactions am ong the
variables co m po sin g a sectoral system have to be studied in great depth both
em pirically and theoretically. T his requires the developm ent o f quantitative
analyses, econom etric studies, and form al m odels. D riven b y em pirical an ­
alyses, appreciative and form al theoretical w ork has to be carried out
regarding the basic relationships am ong the elem ents o f a sectoral system,
the em ergence and persistence o f networks, the basic processes o f variety
creation and selection, and coevolution. Here, both theoretical m odels o f
industry dynam ics and h isto ry-frien d ly m odels can be useful. In the best
evolutionary and in n o vatio n system traditions, this w ork should go hand in
hand w ith, and be co n tin u ou sly confronted by, em pirical w ork.
(3) Research sh ou ld focus o n som e key variables that are still rather unexplored.
In particular:
* the extent and features o f w ithin-sector firm s heterogeneity and the related
processes o f variety creation and selection;
* dem and, in term s o f em ergence, structure, and role in the innovation
process;
* netw orks, in term s o f em ergence, com position, structure, and evolution;
* coevolution o f the vario u s elem ents o f a sectoral system;
* institutions, both in term s o f em ergence and role o f sectoral institutions
and in term s o f the sectoral effects o f national institutions.
(4) Taxonom ies o f sectoral system s have to be constructed. Here, com parative
w ork is p articu larly relevant. These taxonom ies should group sectoral
system s in term s o f elem ents, structure, and dynam ics, so that regularities
402 FR A N C O MALERBA

ш а у be identified am o n g sectors* P avitt s ta x o n o m y (P avitt 1984) and the

Sch u m p eter M a rk I and Sch u m p eter M a rk II d istin c tio n co u ld be useful
startin g points.
{5) A n alyses o f the relationship betw een the presen ce an d stren gth o f elem ents o f
sectoral system s and the international p e rfo rm an ce o f co u n tries have to be
developed (see e.g. Coriat, Malerba, and Montobbio 2004)*
In co n clu sion , as stressed above, a full u n d erstan d in g o f the determ inants,
features, an d effects o f in n o vatio n in sectoral system s req u ires the integration o f
vario u s types o f co m p lem en tary analyses: d escriptive, q u an titative, econom etric,
and theoretical.

N otes

1. In fact sectoral systems often have more than one technology, while the same technology
(as in the general purpose technology case) may be used by many different sectors.
2. Sim ilarly in addition to firm and non-firms organizations, also agents at lower and higher
levels o f aggregation such as individuals or consortia o f firms may be the key actors in a
sectoral system.
3. For example, sectors with competing technologies such as nuclear energy (Cowan 1990),
cars (and their power sources— Foreman-Peck 1996), metallurgy (ferrous casting— Foray
and Grubler 1990) and multimedia (VCR— Cusumano et al. 1992) show interesting
examples o f path-dependent processes.

R eferen c es

A n d e r se n , B., M e t c a lfe , J, S., and T eth er , B, S. (2002), “ Distributed Innovation Systems

and Instituted Economic Processes,” Working Paper ESSY, http://wwwxespri.it/ ricerca/
metcalfeetal.PDF
A ro ra , A., L a n d a u , R., and R o sen ber g , N. (eds.) (1998), C h e m ic a ls a n d L o n g -T e rm G row th:
In sig h ts fr o m the C h e m ic a l In d u stry , New York: John Wiley.
A u d retsch , D. (1996), In n o v a tio n a n d In d u s try Evolution-, Cambridge, Mass.: M IT Press.
B r e sc h i , S., and M a l e r b a , F. (1997), “ Sectoral Systems o f Innovation,” in Edquist 1997*
130—55.
B r e sn a h a n , T., and G r e e n st e in , S. (1998), “ Technical Progress in Com puting and the Uses
o f Computers, B ro o k in g P a p ers on E c o n o m ic a c tiv ity : M ic ro e c o n o m ic s 1 : 1 —78.
" (1 999)> Technological Competition and the Structure o f the Computer Indus­
try,” Jo u r n a l o f In d u s tr ia l E co n o m ics 47:1-40.
and M a l e r b a , F. (1999), Industrial Dynamics and the Evolution o f Firms’ and
Nations Competitive Capabilities in the World Computer Industry,” in Mowery and
Nelson 1999: 79-132.
 S E C T O R A L S Y S T E M S OF I N N O V A T I O N 403

Gallon, М. (1992)* "The Dynamics o f Techno-Economic Networks,” in R. Coombs, R Saviotti,

and V. Walsh (eds«), Technical C h a n g e a n d C o m p a n y Strategiesy London: Academy Press
* Carlsson , B., and S t a n k ie w it z , R. (1995), "O n the Nature, Function and Com position o f
Technological System s, in B. Carlsson (ed. )> T echnological System s a n d E co n o m ic P erfo rm -
ancCy Dordrecht: Kluwer.
Casper , S., and K ettler , H. (2002), “ N ational Institutional Fram eworks and the H ybrid­
ization o f Entrepreneurial Business M odels: The Germ an and U K Biotechnology Sectors,”
Working Paper ESSY, http://www.cespri.it/ricerca/ESSY.htm
---- and S o s k ic e , D. (2004), "Patterns o f Innovation and Varieties o f Capitalism : Explain­
ing the D evelopm ent o f H igh-Technology Entrepreneurialism in Europe,” in F. M alerba
(ed.), Sectoral System s o f In n o v a tio n . C o n cep t , Issues a n d A nalyses o f S ix M a jo r Sectors in
Europe , Cam bridge: C am bridge U niversity Press.
Cesaroni, E , G a m ba r d ella , A., G a r c ia -F ontes , W., and M ariani , M. (2004), "T h e
Chemical Sectoral System : Firm s, M arkets, Institutions and the Processes o f Knowledge
Creation and D iffusion,” in M alerba 2004.
C handler , A. (1990), S c a le a n d Scope: the D y n a m ic s o f In d u s tria l C a p ita lism , Cam bridge,
Mass.: Beliknap Press.
Chesnais, F. (1993), "T h e French N ational System o f Innovation,” in Nelson 1993:19 2-229,
Christensen , С M ., and R osenbloom , R, S. (1995), "E xplain in g the A ttackers Advantage:
Technological Paradigm s, O rganizational D ynam ics, and the Value Network,” R esearch
Policy 24: 233-57.
* C oriat , B., M a l e r b a , E , and M onto bbio , F. (2004), "T h e International Perform ance o f
European Sectoral System s,” in M alerba 2004.
Cohen, W., G oto , A ., N ag at a , A ., N elson , R., and W a lsh , J. (2002), “ R & D Spillovers,
Patents and the Incentives to Innovate in Japan and the United States,” R esearch P o licy 31:
1349-67.
------ and L e v in th a l , D. (1989), "In n o vation and Learning: The Two Faces o f R & D ,”
Econom ic Jo u r n a l 99: 569-96.
------ and M a l e r b a , E (2002), "Is the Tendency to Variation a C h ief Source o f Progress?”
In d u stria l a n d C o rp o ra te C h a n g e 10: 587-608.
Cooke R , U ra n g e , М. G., and E x t e b a r r ia , G. (1997), "R egional Innovation Systems:
Institutional and O rganizational D im ensions,” R esearch P o lic y 26: 475~9i*
C oriat , B „ and W e in s t e in , O. (2004), "T h e O rganization o f and the D ynam ics o f In n o v­
ation: A 'Sectoral5 V ie w ” in M alerba 2004.
C owan, R. (1990), "N u clear Power Reactors: A Study o f Technological Lock-In,” Jo u r n a l o f
E conom ic H is to ry s o : 541-6 6 .
------ D a v id , R , and F o ray , D. (2000), "T h e Explicit Econom ics o f Codification and the
Diffusion o f Knowledge,55 In d u s tr ia l a n d C o rp o ra te C h a n g e 9: 211-53,
C usumano , М . (1991), Ja p a n S o ftw a re F a cto ries , O xford: O xford U niversity Press.
------ M ylo n ad is , Y „ and R osenbloom , R, (1992), "Strategic M aneuvering and M ass-
Market D ynam ics: The Trium ph of VHS over Beta,55 B u sin ess H isto ry R e v ie w 66: 51—94.
D'A dderio , L. (2001), In s id e the V irtu a l P ro d u c t: T h e In flu e n c e o f In teg ra ted S o ftw a re System s
on O rg a n isa tio n a l K n o w le d g e D y n a m ic s , SPRU , Brighton: University o f Sussex.
------ (2002), "T h e D iffusion o f Integrated Software Solutions: Trends and Challenges,55
W orking Paper ESSY, http://www.cespri.it/ ricerca/ESSY.htm

* Asterisked items are suggestions for further reading.

404 FRANCO MALERBA

D a h m e n , E. (1989)» “ D evelopm ent Blocks in Indu strial E co n o m ics,” in B. Carlsson (ed.),

In d u s tr ia l D y n a m ic s, Boston: Kluwer.
D atu m , B. (2002), “ D ata C o m m u n ication — the Satellite an d T V Subsystem s, Working
Paper ESSY, http://www.cespri.it/ ricerca/ESSY.htm
_____and V i l l u m s e n , G. (2001) “ Fixed D ata C o m m u n icatio n s— C hallenges Fo r Europe”
W orking Paper ESSY, http://www.cespri.it/ ricerca/ESSY.htm
D o s i, G. (1988), “ Sources, Procedures and M icro eco n o m ic Effects o f I n n o v a t i o n Jo u rn a l o f
E c o n o m ic L ite ra tu re 2 6 :112 0 -7 1.
-— (1997), “ O pportunities, Incentives and the C ollective Patterns o f Technological
Change,” E c o n o m ic J o u r n a l 10 7 :15 3 0 -4 7 .
-------and M a l e r b a , F. (1996), O rg a n iz a tio n a n d S tra te g y in th e E v o lu t io n o f the Enterprise ,
Lon don: M acM illan.
------ M a r en g o , L „ and F agxolo, G. (1998), “ Learnin g E v o lu tio n a ry En viron m en ts ” IIASA
W orking Paper, W P-96-124.
D u bo c a g e , E. (2002), “ The Financing o f In n ovation b y V enture C ap ital in Europe and in the
U SA : A C o m p arative and Sectoral A pproach,” W orking Paper ESSY, http://www.cespri.it/
ricerca/ESSY.htm
E d q u ist , C. (ed.) (1997), System s o f in n o v a tio n , Lo n d o n : Pinter.
-------(2004), “ Telecom m unication Equipm ent and Services,” in M alerb a 2004.
* ---- M a l e r b a , E , M e t c a l fe , S., M o nto bbio , E , and S t ex n m u eller , E. (2004), “ Sectoral
System s: Im plications for European Technology P olicy”, in M alerb a 2004.
* E SSY (2002), “ European Sectoral System and Eu ropean G ro w th an d Com petitiveness,
European Targeted Socio Econom ics Research,” in ESSY, http://w w w .cespri.it/ ricerca/
ESSY.htm
F o ray , D., and G r u b l e r , A. (1990), “ M o rp h o lo gical A n alysis, D iffu sio n and Lock-out of
Technologies: Ferrous Casting in France and the FR G ,” R ese a rc h P o lic y 19: 535-50.
F o rem a n -P e c k , J. (1996), “ Technological Lock-in and the P ow er Sou rce fo r the M otor Car,”
U n iversity o f O xford, D iscussion Paper in Econom ics.
F r e e m a n , C. (1968), “ C hem ical Process Plant: In n o vatio n an d the W orld M arket,” N a tio n al
In stitu te E c o n o m ic R e v ie w 45: 29—51.
-------(1982), T h e E co n o m ics o f In d u s tr ia l In n o v a tio n , Lon don : Pinter.
-------(1987), T ech n ology P o lic y a n d E c o n o m ic P e rfo rm a n c e : L esso n s fr o m J a p a n y London:
Pinter.
G a m b a r d ella , A. (1995), S cien ce a n d In n o v a tio n : T h e U .S. P h a r m a c e u t ic a l In d u s try du rin g
the 1980s , C am bridge: C am bridge U n iversity Press.
G e r o sk i , P. ( і 9 9 5 )> W hat do we K n ow about E n try?” In t e r n a t io n a l Jo u r n a l o f In du strial
O rg a n iz a tio n 4: 4 2 1-4 0 .
G o r t , M ,, and K l e p p e r , S. (1982), “ T im e Paths in the D iffu sio n o f P rod u ct Innovations,”
E c o n o m ic Jo u r n a l 92: 630-53.
H en d erso n , R ., and C l a r k , K. (1990), “A rchitectural In n o vatio n ,” A d m in is t r a tiv e Science
Q u a rte rly 35: 9-30.

O r sen ig o , L., and P isan o , G. (1999), “ The Pharm aceutical In d u stry and the Revolu-
tion in M olecular Biology,” in M o w ery and N elson 1999: 26 7-312.
H u g h es , E P. (1984), I he Evolution o f Large T echnological System s,” in W. Bijker,
I. Hughes, and T. Pinch (eds.), T h e S o c ia l C o n stru c tio n o f T e ch n o lo g ica l System s, Cam ­
bridge, M ass.: M IT Press.
 S E C T O R A L S Y S T E M S OF I N N O V A T I O N 405

Klepper , S. (1996), “ Entry, Exit, G row th and Innovation over the Product Life Cycle”
A m erican E co n o m ic R e v ie w 86: 562-83.
Langlois, R«, and S t ein m u e ller , E. ( 1 9 9 9 )> I h e Evolution of Competitive Advantage in
the Worldwide Sem iconductor In d u stry ” in M ow ery and Nelson 1999:19-78.
Levin , R., K lev o ric k , A ., N elso n , R., and W in t e r , S. (1987), “A ppropriating the Returns
from Industrial R & D ,” Brookings Papers on Economic Activity 3: 783-831.
Lundvall, R. A. (1993), National Systems o f Innovation, London: Pinter.
----- and J ohnson , B. (1994), “The Learning Economy,” Journal of Industry Studies 1 : 23-42.
McK elvey , M . (1997)» “ Using Evolutionary T h eory to Define Systems o f Innovation,” in
Edquist 1997: 200-22.
----- O rsenigo , L , and P a m m o l l i , F. (2004), “ Pharm aceuticals as a Sectoral Innovation
System,” in M alerba 2004,
----- A lm , EL, and R ic c a bo n i , M . (2002), D oes C o -lo ca tio n m a tter?K n o w le d g e collabo ration
in the Swedish Biotechnology-pharmaceutical Sector, Working Paper ESSY, http://www.ce-
spri.it/ ricerca/ESSY.htm
Malerba , F. (1985). T h e S e m ic o n d u c to r B u sin ess , M adison, Wis.: University o f W isconsin
Press.
------ (1992), “ Learning by Firm s and Incremental Technical Change,” E co n o m ic Journal 102:
845-59*
------ (2002), “ Sectoral System s o f Innovation and Production,” R esearch P olicy 31: 247-64.
* ----(ed.) (2004), S ectoral System s o f In n o v a tio n : C on cept , Issues and Analyses o f S ix M a jo r
Sectors in E u ro p e , C am bridge: Cam bridge U niversity Press.
------ and O rsen igo , L. (1996), “ Schum peterian Patterns o f Innovation,” C a m b rid g e Jo u r n a l
o f Econom ics 19: 47-65.
* ----------- (1997), “ Technological Regim es and Sectoral Patterns o f Innovative Activities,”
In d u stria l a n d C o rp o ra te C h a n g e 6: 83-117.
------------- (1999), “ Technological Entry, Exit and Survival: An Em pirical Analysis o f Patent
Data,” R esearch P o lic y 28: 643-60.
----- ------ (2000), “ Knowledge, Innovative Activities and Industry Evolution,” Industrial
a n d C o rp o ra te C h a n g e 9: 289-314.
------ and T o r r isi , S. (1996), “ The D ynam ics o f M arket Structure and Innovation in the
Western European Softw are Industry,” in M ow ery 19 9 6 :16 5-9 6 .
* M a r sili , O. (2001), The Anatomy and Evolution o f Industries: Technological Change and
Industry Dynamics, Cheltenham: Edward Elgar.
------ and V ersp a g e n , B. (2002), “ Technology and the D ynam ics o f Industrial Structures: An
Empirical Mapping o f Dutch Manufacturing,” Industrial and Corporate Change 11(4)*
791-815.
Mazzoleni , R. (1999), “ Innovation in the M achine Tools Industry: A Historical Perspective
o f the D ynam ics o f C om parative Advantage,” in M ow ery and Nelson i 9 9 9 : 169-216.
M etcalfe , S. (1998), E v o lu tio n a ry E co n o m ics a n d C re a tiv e D estru ctio n , London: Routledge.
M owery , D. (ed.) (1996), T h e In te rn a tio n a l C o m p u te r S o ftw a re In d u stry : A C o m p a ra tiv e
S tu d y o f In d u s tr y E v o lu tio n and Structure, Oxford: Oxford University Press.
* ----and N e l s o n , R. (1999), The Sources o f Industrial Leadership, Cambridge: Cambridge
University Press.
N e l s o n , R. (1993), National Innovation System s : A Comparative Study, Oxford: Oxford
University Press.
40 6 FRANCO MALERBA

*____ (i994), “ The Coevolution of Technology, Industrial Structure and Supporting Insti­
tutions,” Industrial and Corporate Change 3: 47-64.
N elso n , R. (1995), “ Recent Evolutionary Theorizing about Economic Change, Journal of
Econom ic Literature 33: 48-90.
— - —and R o s e n b e r g , N. (1993)» “ Technical Innovation and National Systems, in R, Nelson
(ed.), N ation al Innovation Systems, Oxford: Oxford University Press, 3-22.
------and W i n t e r , S. (1982), A n Evolutionary Theory o f Econom ic Change, Cambridge, Mass,:
Belknapp Press.
O w en -S m it h , J., R ic c a bo n i , M., P a m m o l l i , F., and P o w ell , W. W. (2002), A Comparison
of US and European University-Industry Relations in the Life Sciences,” Working Paper
ESSY, http://wwwxespri.it/ricerca/ESSY.htm
*Р а ѵ іт т , K. (1984), “ Sectoral Patterns of Technical Change: Towards a Taxonomy and a
Theory,” Research Policy 13: 343-73.
PACE (1996) (Policy, Appropriability and Competitiveness of European Enterprises), Brus­
sels: European Commission.
R iv a u d -D a n s e t , D. (2001), “ The Financing of Innovation and the Venture Capital, the
National Financial and Sectoral Systems,” Working Paper ESSY. http://www.cespri.it/
ricerca/ESSY.htm
R obson , M „ T o w nsend , J., and P a v it t , K. (1988), “ Sectoral Patterns of Production and
Use of Innovation in the U.K.: 1943-1983,” Research Policy 17 :1- 14 .
R o se n b e r g , N. (1982), Inside the Black Box, Cambridge: Cambridge University Press.
------(1998), “ Technological Change in the Chemicals: the Role of University-Industry
Relationships,” in Arora, Landau, and Rosenberg 1998:193-230.
S a x e n ia n , A. (1994), Regional Advantages, Cambridge, Mass.: Harvard University Press.
S c h e r e r , M. (1982), “ Interindustry Technological Flows in the U.S.,” Research Policy 11:
227-46.
S t e i n m u e l l e r , W. E. (2002) “ Embedded Software: European Markets and Capabilities,”
Working Paper ESSY, http://wwwxespri.it/ ricerca/ESSY.htm
------(2004), “ The Software Sectoral Innovation System,” in Malerba 2004.
T e e c e , D., and P isan o , G. (1994), “The Dynamic Capabilities of Firms: An Introduction”
Industrial and Corporate Change , 3: 537-56.
T e u b a l , M., Y inn o n , X, and Z u sco v itc h , E. (1991), “ Networks and market creation,”
Research Policy 20: 381-92.
F i r o l e , J. (1988), The Theory o f Industrial O rganization , Cambridge, Mass.: MIT Press.
T orrisi , S. (1998) Industrial Organisation and Innovation: A n International Study o f the
Software Industry, Cheltenham : Edw ard Elgar.
T u sh m an , M. L., and A n d e r s o n , P. (1986), “Technological Discontinuities and Organiza­
tional Environments,” A dm inistrative Science Q uarterly 14: 311-47.
U t t e r b a c k , J. (1994), M astering the D ynam ics o f Innovation, Boston: Harvard Business
School Press.
V on H ippel , E. (1988), The Sources o f Innovation, Oxford: Oxford University Press.
Wengel , and S h a p i r a , R, (2004), “ Machine Tools: The Remaking of a Traditional
Sectoral Innovation System?” in Malerba 2004.
W inter , S. (1984), Schumpeterian Competition in Alternative Technological Regimes,”
Journal of Econom ic Behaviour and Organisation 5: 287— 320,
_ _ ( 19g7y "Knowledge and Competence as Strategic Assets,” in D. J. Teece (ed.), The
C o m p e titiv e C h a lle n g e : Strateg ies f o r In d u s t r ia l In n o v a tio n a n d R e n e w a l, Cambridge,
Mass,: Ballinger, 159-84.
 C H A P T E R 15

I N N O V A T I O N IN
« Т Л ТДT . T p Г U ”
ЛЬшвЯШ ▼ V iJm im MR ^ m Jb * *Ш я*

INDUSTRIES

NICK VON T U N Z E L M A N N

V IR G IN IA A C H A

15.1 In tr o d u c tio n

T title o f this chapter is inherently paradoxical— low-tech industries are not

h e

supposed to be characterized by any significant amount o f innovation ex definitione.

We intend to resolve this conundrum by arguing that there are few if any
industries in present-day advanced countries which conform to the general under
standing o f what constitutes a “ low-tech” industry. In our view, this is more than just
a matter o f semantics, and it is crucial for understanding where the comparati
advantages o f countries at varying levels o f development may lie. We believe that a
policy obsession with purported “ gaps” in “ high-tech industries has istracte t e
attention o f both policy makers and academics away from making more positive
efforts to develop and sustain development in other sectoral directions w іс
countries might find more viable. In the OECD, high-tech in ustries as t e
itself defines them account for only about 3 per cent o f va ue a e irsc
Kreinsen et al. 2003), rising to 8.5 per cent if medium-high-tech industries 1 e
4 08 NICK VON TUNZELMANN AND VIRG IN IA A C H A

m otor vehicles are included (O ECD 2003), so even if they could be expanded the
impact on GD P would be quite small. Governments need to give m ore thought to
the activities which generate most o f the output and em ploym ent o f their countries
and the best targets for “ dynamic comparative advantage for growth.
In this chapter we will consider not just tradition al4low-tech industries but also
those classified by the OECD as “ m edium -tech” We combine them hereafter as
“ low- and medium-tech” (LM T) industries. Our reason for banding the two to­
gether is that both are being driven by similar factors which somewhat distinguish
them from the high-tech industries. These are frequently “ m ature” industries, where
technologies and market conditions may change m ore slowly. They can include non­
m anufacturing activities, such as exploration for new resources in the oil and gas
industry. We naturally avoid much discussion o f services as being the subject o f the
following chapter, but this ignores the growing interrelationships o f service and
production/manufacturing activities in many relevant areas, so we include some
consideration o f the services sector.
Section 15.2 untangles the relationship between the technologies and markets
which comprise an “ industry” — effectively we reject O ECD -type classifications in
favor o f alternative sectoral taxonomies. This permits a m ore constructive analysis of
the key drivers o f change at the industry/sector level as between dem and and supply
factors in Section 15.3. Section 15.4 elucidates the roles o f firm strategies and
structures in the LM T industries. The implications o f this discussion for the evolu­
tion o f industry structure, especially entry, are considered in Section 15.5, and
illustrations from some “ low-tech” industries are given in Section 15.6. Section 157
provides implications for government policy, which we believe requires radical
rethinking, and Section 15.8 concludes with a call for revising the academic agenda.

15.2 T he “ T ech no lo gy P ro file”

of In d u st r ia l S ec t o r s

Sectors are generally taken to be identifiably similar aggregations o f productive

activities. Conventionally, sectors o f all types were supposed to be recognizably
different from one another not only in the goods and services they produced but
also in the technologies and processes they used to produce them. However, the
boundaries have blurred over historical time in both dimensions. Technologies
originally developed for one set o f products spill over into use in the production
or “ architecture” o f other sets o f products. Moreover, new technologies more often
tend to supplement and complement old technologies rather than replace them.
 I N N O V A T I O N IN “ L O W - T E C H ” I N D U S T R I E S 409

One simple consequence is that even “ old” products can be produced by, or partly
consist of, elements drawn from what had previously been a totally different set o f
activities. Equally, markets have become more blurred through the bundling of
goods and services (e.g. sales o f music products via the Internet).
As a result, conventional classifications o f sectors as high- or low-tech (etc.), as
long practised by the OECD, are becoming less and less useful for academic analysis,
though their sway still prevails in government policy making (see Section 15.7
below). To be fair, the OECD (2003) is coming— rightly— to place greater emphasis
on the “ knowledge-intensity” o f industries, based on criteria such as their use of
capital inputs from R&D-intensive industries, This necessitates rethinking the kinds
of taxonomies that help us to comprehend structural change (see Ch. 6 by Smith, this
volume).

15.2.1 Technologies vs. Products

The conventional definition o f the technological profile o f industrial sectors put
forward by the OECD claims to measure the direct plus indirect technology content
of particular industries. The m ajority o f manufacturing industries are defined
mostly according to their product range but a good number have in common
their technologies rather than their products, e.g. biotechnology. Whether, say,
plant biotechnology is regarded as part o f the biotechnology industry (technol­
ogy-defined) or o f agriculture (product-defined) makes a big difference to the
inferences drawn.
Allowing for these difficulties, others then arise in connection with the measures
of technology content. The key issues are outlined by Smith (Ch. 6 in this volume),
and need not be reiterated here. However we would emphasize that a good part o f
the innovation activities in LM T industries may fall outside the Frascati definition o f
R&D (OECD 1994), for example oil and gas exploration (see Box 15,1). Knowledge
search, identification and proof, rather than basic research, are likely to be of
particular importance to innovation in the non-manufacturing activities o f LMT
industries. Most importantly, we have to ask what part o f each “ industry"' we are
characterizing as high- or low-tech when considering their growth potential
Meliciani (2001) found that in the 1980s the ICT industries figured prominently
among the fastest-growing industries in advanced industrial countries, but in the
1970s they did not, whereas many LM T industries were present among the fastest
growing group at this time. For the years after 1994, the OECD high-tech sectors
show first rapid growth then rapid contraction. Governments have been attracted
towards the high-tech industries for their potential for growth and structural
change. The benchmarks here are however confused by the definitional problems
already highlighted— for instance Denmark appears to have a comparative
4Ю NICK VON T U N Z E LM A N N AND V IR G IN IA A C H A

Box 15*1 Planetary science in the North Sea

There are many who believe that aU of the science and technology needed to dig oil and
gas out of the ground has been already discovered and deployed. Relatively low R&D
intensity figures attest to the same; over the period 1995 to 1997, the top teii oil majors
worldwide averaged an RBcD intensity o f 0.52 per cent However the accounting
category of R&D does not capture the full scope of investment into the search for
novelty and the applications thereof in the upstream petroleum industry Мшу
exploration activities that involve, in the words of the Frascati Manual, an Appreciable
element of novelty and the resolution of scientific and/or technological uncertainty^
and “ contribute to the stock of knowledge'" are captured under different budgetary
headings, notably exploration costs (Acha 2002). Mansfield (1969: 53) was also
concerned that geological and geophysical exploration were excluded from the R&D
definitions for precisely these reasons.
In fact, the annually growing numbers (in their hundreds per company) o f technical
and academic papers produced by the leading oil operating and service companies
reflect a substantially developed research program for the evolution of the science and
technologies underpinning this industry. This fact is better reflected in the number of
Fh.D. qualified staff from these firms actively working on a better understanding of the
composition and dynamics of this planet. This scientific endeavor in planetary science
does not go unnoticed even outside the industry. NASA has enlisted the help of
upstream oil industry companies in developing new technologies for drilling on
Mars (Babaev 2000). Likewise space scientists are collaborating with the geologists
and geophysicists who have been delineating marine space impact craters on Earth,
including one off the UK whose only natural analog exists on Jupiter"s ice moon,
Europa.

advantage in low-growth sectors (such as processed foods), but its production lies
towards the high-tech end o f these low-growth sectors (like applying biotechnology
to food processing).

15*2*2 Factor intensity

1 he LM T industries are usually regarded as providing m any points o f entry for
developing countries, in view of their relative labor-intensity. N ot all are o f this
nature, however. Some branches o f such “ low-tech” industries as food processing are
highly capital-intensive (e.g. tobacco and m any beverages), as are som e branches of
building materials (e.g. cement). M any more take on varying shades o f labor- or
capital-intensity depending on the economic environm ent in which they find
themselves the same industry may be capital-intensive in the USA and labor-
intensive in China. Areas like software production have been favored points of
entry into high-tech production in countries like India in the 1990s and beyond
 I N N O V A T I O N IN “ L O W - T E C H ” I N D U S T R I E S 411

precisely because they are “ labor-intensive” (though intensive in cheap skilled labor
rather than cheap unskilled labor).
Peneder (2001) provides a tripartite classification o f manufacturing industries, at
the 3-digit level o f disaggregation. One o f his taxonomies rests on factor intensity
(mainstream Le. average; labor-intensive; capital-intensive; marketing-driven; tech­
nology-intensive), another on labour skills (low-skill; medium-skill blue-collar;
medium-skill white-collar; high-skill), and the third on external service inputs
(from knowledge-based services; from retail and advertising services; from transport
services; and from other industries). Only one o f the ninety-nine manufacturing
industries he lists (aircraft and spacecraft) has the classic high-tech profile of being
technology-intensive and predominantly using high-skill and knowledge-based
services; conversely there are labor-intensive industries which utilize high skills
(e.g. machine tools) and others utilizing knowledge-based service inputs (some
branches o f metallurgy). His classification underlines the great variety o f observable
combinations.

15.2.3 Pavitfs taxonomy

In contrast to classifying industrial sectors according to product range, Pavitt (1984)
arranged them according to technology characteristics (see Tidd et al. 2001). The
suggested categories are: “ supplier-dominated,” “ scale-intensive,” “ information-
intensive,” “ science-based,” and “ specialized-supplier” (for further discussion of
this taxonomy see the Introduction and Chapter 6 by Smith, in this volume). A
modified taxonomy o f this kind that is more explicitly geared to the kinds o f
technological paradigms (chemical, mechanical, etc.) in different sectors is given
by Marsili (2001). Both Pavitt and Marsili deliberately aimed at means o f classifica­
tion that brought together characteristics which certain groups o f technologies
appeared to share, even though they might pertain to different “ sectors.” Generally
this taxonomy does a better job o f explaining technological performance than factor
content, but since this is its intention that may not be too surprising. However, again
the LMT industries resist easy classification, precisely because many o f them are not
very distinctive or singular in technological terms.
An attempt to use a modified Pavitt taxonomy to analyze changes in world export
shares is given in Table 15.1. The industries given special attention in this chapter are
sprinkled across the first four categories, as already implied (e.g. food products, oil
and gas, paper in agricultural products and raw materials; textiles and clothing, glass
in traditional manufactures; vehicles, steel in scale-intensive; machinery for many of
these in specialized suppliers). The inroads made into European and US export
shares during the 1970-95 period by Japan and the Asian NICs in science-based and
specialized suppliers are evident, though the share o f the Asian NICs also in a eases in
412 NICK VON T U N Z E LM A N N AND VIRGINIA A C H A

Table 15.1 Market shares, 1 9 7 0 -1 9 9 5 (percentage ratio o f national exports to

world exports)

A g ric u ltu ra l T ra d itio n a l S cale- Specialized Science- Total

p ro d u cts and in d u strie s intensive suppliers based
raw materials

Europe 1970 24.1 57.0 55.7 61.2 4 8 ,6 ,' 44.6:

1995 31.6 40.1 47.3 47.6 3 3 .8 39.6
Change + 7 .5 — 16.9 -8 .4 “ 13,6 -1 4 .8 - 5 .0

USA 1970 13.1 7.4 14.5 22.3 29.5 14,8

1995 11,0 6.7 10 .3 13.7 17,9 11,0
Change - 2.1 “ 0.7 “ 4.2 ~ 8,6 “ 11.6 “ 3.0

Japan 1970 1.2 9.3 13.8 6.4 7,7 6J

1995 1.4 3.2 12.8 15.7 14.3 9.0
Change + 0.2 “ 6.1 “ 1.0 + 9 .3 + 6.6 + 2 ,3

Asian 1970 2.0 6.1 1.0 0.8 1.0 2.1

NICs 1995 3.4 16.2 8.7 8.8 17,8 10.8
Change + 1.4 + 10.1 !77 + 8.0 + 1 6 .8 + 8 .7

Source: Fagerberg et al. 1999:12.

traditional manufactures. The US actually loses least, in percentage points, in

traditional manufactures from 1970 to 1995, though its share in such trade was
very low throughout the period compared with Europe. Nevertheless the table
should warn us against making oversimplified statements about technological
patterns o f development.

15.2,4 Sutton's taxonomy

Yet another way o f classifying industries can be derived from the work o f Sutton
(1991,1998), who demonstrated that what firms are prepared to spend on marketing
their products on the one hand, and on developing their technologies on the other,
depend on factors that were partly under a firm’s control and partly beyond it. The
latter means that if the firms belong to a certain industry they are committed to a
certain level o f sunk costs, e.g. for their production processes, regardless o f what
strat egies they then adopt. Ih e nature o f technologies as they relate to products sets
bounds to market concentration in a particular line o f activity. The tire industry,
for example, is very capital intensive, bound by production technologies that have
 I N N O V A T I O N IN LOW -TEC H ” INDUSTRIES
40

not essentially changed since the first large-scale use o f rubber; this promotes a global
oligopolistic structure (Box 15.2 discusses the tire industry more fully). But what
firms* in tires or elsewhere* choose to do as a result of strategic decision making—the
endogenous element— would rest on the profitability o f their strategic expenditures
in the face o f similarly strategic investment behavior by their competitors.

Box 15.2 Whatfs so clever about a rubber tube?

The modern tire industry has its origins in the nineteenth century and* by and large* its
development has mirrored that of the automotive industry. The world tire industry has
evolved to address several markets which each have different characteristics* direct
customers, and potential for growth. Over a century of development, it has segmented
across a wide number of markets, including automotive, aerospace, bicycle, and
locomotive tires. The industry is relatively highly concentrated; in 2000, the top ten
global tire manufacturers accounted for 83 per cent of the sales of the top seventy-five
tire manufacturers globally (www.tirebusiness.com/statistics).
Technology is applied by the tire manufacturers to reduce costs, to differentiate the
product line and to focus on greater value-adding activities (Acha and Brusoni 2003).
Facing a global market where it is more and more difficult to make a profit, the leading
manufacturers are continuously focusing on reducing costs through reducing
throughput and labor costs (including the long-awaited introduction of robotics),
innovations in processing technologies and source product (a new polyurethane tire
polymer), and in the method itself.
Beyond influencing the cost and ease of production, tire manufacturers have
invested in research and technology to also help them to move away from the
“commodity trap/5where products can only compete on price. Product differentiation
has occurred as companies offer a variety o f tire profiles and even colours to match cars
(the latter was led by Kumho, a Korean tire manufacturer that has successfully broken
into the top ten manufacturers). Such differentiation is certainly much more than
cosmetic; manufacturers have successfully incorporated new sensors into tire assem­
blies and developed run-flat tires. Leading tire manufacturers are now looking to move
up the value chain by manufacturing entire tire assembly systems rather than simply
supplying the tire itself
O f course all o f this could be achieved through the support o f suppliers from the
high-tech sectors o f electronics and chemicals, yet the tire manufacturers themselves
patent in these areas and lead developments as applied to their business. Patenting in
technologies related to tires has increased, and most dramatically so for the tire
manufacturers (Acha and Brusoni 2003). The top ten tire manufacturers worldwide
had an average RScD intensity o f 4 per cent in 2000. Moreover, their patents are
applicable to a large number o f International Patent Classification (IPC) subclasses
(an average of forty-four subclasses over the period 1990 to 2000), indicating some
complexity in the nature o f these firms5knowledge bases. These firms have broadened
their focus to address the crucial interface between their chemical knowledge base (Le.
rubber and other chemical compounds) on the one side, and mechanical engineering
and electronics (e*g. sensors) on the other.
 NICK VON T U N Z E LM A N N AND V IR G IN IA A C H A
414

Drawing on this reasoning, Davies and Lyons (1996) classified country strengths
on the basis o f dividing industries into four categories: those with high R8cD (i.e.
technology) expenditures; those with high advertising (i.e. sales) expenditures;
those with both; and those with neither. They showed that Western Europe was
relatively strong in the second and third o f these categories, i.e. those where both
R8cD and advertising were high (like pharmaceuticals) or advertising alone was
high. The latter includes particularly some industries norm ally regarded as “ low-
tech,” especially food processing (the industry selected by Sutton in his 1991 book to
validate his theory). The Sutton approach and associated taxonom y can be especially
useful for analyzing LM T industries, because supply (technology) is combined with
demand (product) aspects in a rigorous way.

15.2.5 Summary
Attempts to appraise innovation through adopting conventional sectoral classifica­
tions can be quite misleading. Innovation is rapid in particular segments o f both
high-tech and LM T “ se c to rs e v e n if more segments o f the high-tech sectors display
such rapid innovation (for evidence see Ch. 6 by Smith in this volume). It is
admittedly possible to detach the high-tech segments from LM T “ industries” and
label them as new high-tech sectors, as was done for artificial fibers in textiles when
they arose to compete with natural fibers in the early twentieth century (see Section
15.6), but the final products remain very similar so this looks specious. Approaches
that blend the technology dimension and the product dim ension, such as those by
Sutton or by Peneder, appear to be not only m ore analytically satisfying but better
able to account for observed empirical differences between countries and regions. As
suggested below, they are also better able to account for dynam ic paths o f industrial
evolution over historical time. Furthermore, they are in our view a m ore advisable
platform for policy than simple OECD-style definitions o f high- or low-tech. They
need however to be supplemented by technology-oriented distinctions among
sectors such as the Pavitt taxonom y to provide a better grasp o f the nature of
structural change and competitiveness.

15.3 T h e K e y D r i v e r s

I he drivers o f change as they affect low, medium, and high-tech sectors can be
amilarly envisaged from the side of the products or from the side o f the technologies,
 I N N O V A T I O N IN “ L O W - T E C H " I N D U S T R I E S 415

which again give rise to significant differences in interpretation and understanding*

Firms hold different interpretative "fram es” (see below), and in LMT industries
firm-level differences in the interpretation o f demand drivers are particularly im ­
portant because their well-established markets necessitate a broader variety of
strategic choices for differentiation. Demands change sometimes slowly but some­
times rapidly and unpredictable negating attempts to routinize operations and
generating turbulence.

15.3.1 Demand Differentiation

15.3.1.1 Q uality Innovation
An important way in which even older industries can bounce back is by producing
for new markets. Producing the same type o f goods for untouched regions can work
for well-known brands, like Coca-Cola, but producing different types o f the same
categories o f good (“ product differentiation” ) is generally necessary for such resur­
gence. Within given markets, product growth is heavily determined by income
elasticities o f demand, i.e. the responsiveness o f consumers to the particular product
as their incomes rise. It is usually the case that low-tech industries face somewhat
inelastic demands because many produce comparative “ necessities,” and as con­
sumers attain higher income levels they have satisfied most o f their needs for
necessities. To stave o ff this “ satiation o f wants,” producers in LM T industries have
to find new products to attract the custom o f higher earners. The availability of
advanced technologies m aybe an important factor for innovation strategies in LMT
firms through dictating the scope for such new products, and even then may not
result in products that customers find attractive, as has been the case for genetically
modified (GM ) foodstuffs in some countries (see below).

15.3.1.2 N ew Tastes
In addition to quality upgrading, consumers may switch their demand patterns to
goods which have new characteristics. While high-tech sectors may have greater
innate capacity to spawn product innovations, LM T industries may be faced with a
greater necessity to do so. Sectors such as food, energy generation and automobiles
have to confront intense pressure from communities and from governments to
produce safer and more environmentally friendly items. The same pressures extend
to the processes by which they produce these outputs. Less essential but often more
lucrative for LM T industries have been shifts in their product mixes to reflect the
changing composition o f consumers, for example the implications o f demographic
change (gender relations, ageing, etc.). These create new niches in which firms in
416 NICK VON TUNZELMANN AND VIRG INIA A C H A

low-tech industries can experience some resurgence. For exam ple, a leading Japan­
ese toy manufacturer, Bandai, recently launched a new doll product range ( Primo-
puel") targeted not at children but at “ em p ty nesters" (women without children at
home) by embedding sophisticated electronic sensors and program m ing within this
“ real baby5' doll in an effort to develop a higher value market.
Both the potential for developing higher-quality products (“ quality ladders") and
new products therefore offset the seemingly inevitable m aturation o f older indus­
tries, and give rise to new production and trade patterns (Grossm an and Helpman
1991). Indeed, over time, through adding value in processing and in new products,
the declines in demand for the products o f these industries have been less marked
than might be expected. In the case o f tires, companies have developed new model
profiles (e.g. “ fat" tires) to shift demand options from simple requirements (four
tires per vehicle) to a series o f options (different tires for different occasions), and to
differentiated product lines o f increasing quality (e.g. “ run-flat" tire systems).
The challenge for innovation strategies becomes how and how well firms in LMT
industries can alter their products and services and leverage the outcomes through
introducing better products or new products, as explored in Sections 15.4 to 15.6
below.

15.3.2 New Technological Paradigms

15.3.2.1 G eneral Purpose Technologies a n d L ea rn in g in L M T F irm s
Certain new technologies can spill out o f their industry o f origin and be recruited
by older industries. Key technologies often have the property o f being able to become
“ pervasive," through their take-up in one industry after another (Freeman and
Perez 1988; Freeman and Lou<~a 2001). Industrial revolutions are generally consti­
tuted o f several o f these “ general purpose technologies" (Helpm an 1998); for
example machinery, steam power, and iron in the First Industrial Revolution
o f the late eighteenth century; chemicals, internal com bustion, electricity and
steel in the Second Industrial Revolution o f the late nineteenth century; and infor­
mation and communication technologies, biotechnology and smart materials in
the Third Industrial Revolution o f the late twentieth century. In our view, the
general-purpose technologies associated with the Third Industrial Revolution create
new opportunities for LM T industries to enhance their innovative and economic
performance through the effective adoption and application o f ICT, biotechnology,
and smart materials.
The p r o p e r t ie s o f th e se t e c h n o lo g ie s a r e n o t o f c o n c e r n h e re , b e in g th e h ig h -te c h
s e c to r s o f th e ir d a y , b u t th e ir s p r e a d to a d o p t i n g s e c t o r s is o f g r e a t s ig n ific a n c e . T h ey
 I N N O V A T I O N IN t O W - Т Е С н ” INDUSTRIES 4 17

arise most commonly (though not invariably) in “ upstream” activities— in the cases
mentioned, in equipment and capital goods, in motive power and in basic mater­
ials— from which they trickle down to user industries.
In LMT industries there is usually relatively little formal learning by science and
technology, at least at the firm level. Instead, innovation- and adoption-related
learning activities operate in practical and pragmatic ways by doing and using, The
bulk of the new technologies on which they draw, including these general-purpose
technologies, are developed by separate companies (rarely subsidiaries), specialized in
the relevant technological fields. However the downstream LMT industries need to
have “absorptive capacities” to make productive use of these upstream developments.
As shown by Cohen and Levinthal (1989, l 99o)> absorptive capacities are best incul­
cated by doing some o f the innovative activities oneself, even if this means replicating
what has already been done elsewhere. Thus food-processing companies involved in
advanced (“ third-generation” ) biotechnology do not carryout much ofthe associated
research themselves but are often prominent in patenting in less advanced (“ second-
generation” ) biotechnology— this seems to provide them with the necessary absorp­
tive capabilities. In addition, the national systems of innovation literature has drawn
attention to how formal science may congregate in national or regional laboratories in
such industries, instead o f being internalized within firms (Nelson 1993). The applic­
able science produced by these organizations (e.g. public laboratories) may be generic,
but the applications themselves will require firm-specific absorptive capacities, gener­
ated not just from formal R8cD but from broader-based innovative activities that
include engineering, continuous improvement processes, and organizational innov­
ations such as integrated service and supply

15.3.2.2 C a rrier Industries

Describing one particular industry supplying technologies, namely machine tools,
Rosenberg (1963) showed how the number o f different types o f tools was quite
limited and as a result their principles could readily be stretched to being applied in
industries other than where they were first deployed. The counterpart o f this process
was the equally significant role o f “ carrier industries,” which incorporated these
proliferating machine tools into making the machines they produced or used. Based
on the notion o f general purpose technologies, any industry can act as a carrier if its
demand for the new capital good is large enough or growing fast enough. Thus even
low-tech sectors can act as receptors for new process-oriented technologies, and
many examples are given below.
4X8 NICK VON T U N ZE LM A N N AND VIRG IN IA ACHA

15.4 F i r m s a n d C o r p o r a t e C h a n g e
i n LMT I n d u s t r i e s

This section raises issues concerning strategies and structures at the firm level, issues
o f scale and scope in both Carge firms and SM Es, and discusses different kinds of
integration.

15.4л Strategy and Structure

15.4.1.1 Strategy in L M T Industries
According to Porter (1985), there are three main types o f corporate strategy, namely
cost leadership, differentiation, and focus. If viewed from the aspect o f the orthodox
product lifecycle, cost leadership is the likely choice o f firm s in a “ mature” LMT
industry that rely on process innovation to pare down costs even though the
innovative spark has largely ceased. From the alternative points o f view outlined
above, however, the other two strategies also are plausible candidates for companies
wishing at least to survive if not to thrive in competitive lower-tech environments.
Branding is often crucial to the choice o f a differentiation strategy, as consumers
V

choosing between a Saab and a Skoda can be influenced by product reputations

V

(Skoda is nowadays using advertising o f its new technologies to change deep-seated

customer prejudices).

15.41.2 Functions a n d Structures o f the F irm

LM T firms and industries do not differ from any others in the tasks that need to be
carried out, but clearly they may place less emphasis on technology functions and
more on product/marketing functions than, say, a science-based industry. Building
implicitly on this presumption, Chandler (1990) argued that the multidivisional
( M -form ) company was an appropriate organizational structure for diversified
firms in industries characterized by “ branded, packaged goods” (as well as many that
were science-based), such as many branches o f the food industry. By organizing
divisions entrusted with specific product lines in m ultiproduct firms, and giving
middle management the job o f running these divisions, there could be greater focus
on the targeting o f markets, as all the functions necessary for the product line were
incorporated within the division (and the responsibility o f lower management
tiers). However this raised several problems, for example if there were inherent
spillovers between the same functions located in different divisions— this could
make it awkwai d for M -form companies to introduce inform ation systems and
 I N N O V A T I O N IN ^ L O W - T E C H ” I N D U S T R IE S 419

other radical technological changes, which were bound to affect all their carefully
separated divisions (Mowerу 1995).

15.4.1.3 “ Fram es” fo r Each Function

In previous work, we suggested that large firms interpret competitive challenges and
problems through mental “ frames,” organizational cognitive maps through which
they view their horizons and functions, including technology (Acha 2002). In
contrast to the Porter concept o f corporate strategy, the frame is the filter through
which strategies are conceived and chosen. This filter, held by individuals and
mediated at the organizational level by senior management, comprises the main
variables, functions, and contingencies o f a particular phenomenon. In practice,
managers consider particular activities rather than the whole at once. A technology
frame, therefore, is the interpretative system o f managers to understand the firm s
technological position and opportunities as well as the expectations o f the dynamics
of their relevant innovation system(s) (Orlikowski and Gash 1994; Orlikowski
2000).1
For the oil and gas industry we found that frames significantly influenced
performance differences among companies with similar technological competen­
cies. That is, there is little direct correlation between technological performance (say
patenting and scientific publications) and business performance (say expansion or
profitability), unless this intervening variable o f the frame is accounted for.
The frame concept is relevant to the study o f innovation in LM T sectors. Firms in
LMT industries are basically using rather than selling technology, and therefore tend
to adopt technology frames that are quite different from those in high-tech indus­
tries (where technology itself is a key selling point), and often quite different from
other LMT firms even in the same industry. In general, the market characteristics o f
LMT industries (where, over time, markets become segmented and competitive
advantage depends upon product differentiation, cost efficiency and control o f
complementary assets) lead firms to different interpretations about the role for
technology for commercial success.
In “ high-tech” firms, by contrast, the role for technology is more explicitly central
to commercial success, and there m aybe greater tendencies for consensus (general or
by groups) on aspects o f technology frames across these competing firms because:
(x) the market is relatively “ thin,” with relatively fewer product and technology
options as yet; (2) appropriation o f technology through IPR is more aggressively
pursued and this structures the role o f technology for the industry; and (3) regula­
tory environments more frequently play a structuring role (e.g. biotech). To gener­
alize, variation in technology frames across “ high-tech” firms derives more from a
focus on how the technology (broadly stated) should develop, whereas variation in
technology frames amongst LM T firms pertains more to what the role for technol-
ogy {broadly stated) should be.
420 NICK VON TU N ZELM A N N AND VIRGINIA ACHA

15.4.2 Scale and Scope: Large Firms and SMEs

Although they are now mature, such “ L M T ” industries as meatpacking, automobile
production, and consumer durables were im portant sources o f production innov­
ation in their early years, essentially developing the technologies o f mass production.
The driving force o f mass production was to reap economies o f scale in the produc­
tion processes, and the M -form company was a very suitable organizational form for
doing so (Chandler 1977). Beyond being labour-saving and capital-intensive, the
main direction o f technical change was to be “ time-saving.” This was achieved
through raising throughput, reducing downtime, and im proving the m achinery so
as to ensure the speedy throughflow in all stages o f its operation. These benefits can
be thought o f as arising from “ dynamic economies o f scale,” as contrasted with the
usual static scale economies arising, for instance, out of having large plant and
equipment. In some medium-tech industries like steel, the adoption o f new pro­
duction technologies (the electric furnace) has led to a decline in average establish­
ment size (e.g, huge “ integrated plants” have been replaced by “ m inim ills” ), in
which the losses in terms o f static scale economies have been more than offset by
increases in dynamic scale economies through specialization and throughput,
Because o f the high capital costs o f large-scale assembly lines in automobiles, it
was necessary to produce very long runs o f the product in order to cover them.
Com petition in this industry increasingly took the form o f m aintaining the same
m ajor components (body, engine, etc.) for as long as possible, while giving the
semblance o f regular updating o f the model by adding on inexpensive “ frills,” e.g.
associated with annual product launches. In essence the assembly line, to pay off,
required producing products that were as standardized as possible.
This approach was challenged by the Toyota system o f “ lean production”
(Womack et al. 1990), which was a response to the needs o f customers for variety
and specialization while sacrificing as little as possible o f the benefits o f high
throughput— in effect obtaining “ dynamic economies o f scope.” Best known o f
the constituents o f lean production was just-in-tim e production, accelerating the
speed o f response to changes in customer needs, but other changes worked in the
same direction. The lean production system was intended to supplement the
American system o f interchangeable parts in the product with interchangeable
parts in the process. The economies o f scope obtained in the efficient production
o f a variety o f products (the Toyota Crown car fam ily allegedly came in over 100,000
variants) more than compensated for abandoning standardization.
At the other end o f the spectrum, small and medium-sized enterprises (SMEs)
have reappeared on government technology policy agendas. One o f the reasons for
renewed emphasis on them as sources o f innovation is the perceived advantage o f
SM Es in responding quickly to technological change, because o f the absence o f
complex management structures within smaller enterprises. Against that, SM Es may
lack the financial clout to undertake the kinds o f investments in new technologies
 I N N O V A T I O N IN “ L O W - T E C H ” IN D U S T R IE S 421

necessitated by radical change (D o d gso n and Rothw ell 1995). Q uestions o f access to
new technologies and on w hat term s then becom e param ount. This becom es sig­
nificant in those L M T indu stries w hich are dom inated b y SM Es, which include many
of the more traditional areas. W hilst SM E s are frequently seen as a cause o f hope in
high-tech sectors, they are often (perhaps unfairly) seen as a m atter for despair in
LMT sectors.

15.4.3 Vertical and horizontal integration

The LMT industries have been characterized by a variety o f patterns of vertical
integration and disintegration through their development. The low-tech industries
(plus some medium-tech ones) for the most part long pre-date the onset o f indus­
trial revolutions and some date back to prehistoric times. When industries
first emerge, there is usually a high degree o f vertical stratification, i.e. with different
firms responsible for each stage o f production (the “ chain” ). Some go on to acquire
additional technological capabilities which may eventually spin off into newly
created companies (as Rosenberg (1963) showed for machine tools). However the
segments interact in systemic fashion with one another through the vertical
“chains ” In textiles, there were “ imbalances” between each stage o f production—
the early mechanization o f weaving speeded up the production o f cloth, so pressure
was placed on spinners to speed up their own production o f yarns to go into the
cloth, thereby giving rise to the mechanization o f spinning; this in turn became
so successful that renewed pressure was put on the weaving segment to develop
power-based weaving, and so on. The other stages o f textile production— fiber
production (e.g. cotton growing), preparation and finishing-—were similarly
affected along the “ value chain.” Similar “ imbalances” arose in the stages o f produc­
tion of iron.
In the second phase o f industrialization there were pressures to link the segmented
processes more directly through moves towards vertical integration. The pressures
of throughput that gave rise to “ mass production” in the USA in the late nineteenth
century also promoted vertical integration, because this could ensure smooth
production flows throughout the value chain.
More recently, the rise o f the steel minimill, partly from technical change, at the
expense o f the large integrated mill has been one o f the more dramatic demonstra­
tions o f a retreat from vertical integration. Tasks that corporations previously
undertook themselves have been “ outsourced” where possible, thereby returning
to the traditional low-tech model o f vertical disintegration even in some high-tech
industries. To retain and intensify the economies o f scope required in this unfolding
of circumstances, companies also chose to limit the range o f their horizontal
diversification. Firms, including some o f the largest corporations, practised
422 NICK VON TUNZELM ANN AND VIRGINIA ACHA

“ d o w n sizin g ” an d in m an y cases strip ped o u t large n u m b ers o f m id d le m anagem ent

in the b e lie f that this fu rth ered “ lean p ro d u ctio n .
W h ile fin an cial consideration s had often en cou raged d iv ersifica tio n into unre­
lated activities, studies dem onstrated that “ co n g lo m erate” firm s b ased o n unrelated
d iversificatio n w ere n o t v e ry profitab le (R u m elt 1974)- T ak in g o n b o a rd the techno­
lo gical and p ro d u ctio n aspects, an d thereby tak in g in to acco u n t the issues o f
synergies an d econ om ies o f scope, m an y firm s reo rien ted th eir stru ctu re to lim it
them selves to “ related ” d iversificatio n . Yet m an y o f the larger co m p an ies in low-tech
in du stries like fo o d m an u factu rin g con tin u e to p u rsu e a p p a re n tly u nrelated diver­
sification. T h e m ost likely reason fo r this is that lo w tech n o lo g ical o p p o rtu n ity in
trad ition al segm ents o f low -tech in du stries m a y also go w ith relatively h igh app ro­
priability, especially th rou gh b ran din g. T h is generates a p o o l o f resources (Penrose
1959), w h ich can be used to invest in o th er areas in w h ich b ra n d in g p ro vid e s a secure
finan cial base. T h e scale and scope o f econ om ies thus arise v ia m ark etin g rather than
th ro u gh technologies.

15.5 C hange at the In d u stry L evel

15.5.1 Vertical Alignment and Networks

These changes at the firm level in terms o f size, integration and diversification carry
strong implications for the structure o f the industries in which they are embedded.
Firms have been driven to develop closer relationships with upstream suppliers and
downstream customers. Am ong the best known o f these developments are those in
the m otor vehicle industry. One implication o f the Toyota system o f flexible and
just-in-time manufacture was the need for high reliance on suppliers to deliver
equipment and components on time and o f high quality. Rather than controlling
this process hierarchically, Toyota undertook joint development with the suppliers,
spending long periods o f time negotiating the exact specifications and costings
required. Some o f the most critical suppliers became “ first-tier” suppliers and
worked in close association with Toyota wherever the production was located.
Others were seen as less critical or supplied the suppliers: these became “ second-
tier and third-tier suppliers and were involved less and less directly in negoti­
ations with Toyota itself.
Many of these firms saw themselves as developing into “ system integrators,”
acting as the hub for operations immediately connected to their central activity.
Depending on how far they outsourced these related activities, they might end up as
 i n n o v a t i o n in “ l o w -t e c h ” in d u s t r ie s 423

directly producing only a m inority o f the value added by their “ system/5 Moreover
system integrators could emerge at varying points on the value chain, so that vertical
links might increasingly be occurring between various system integrators, That is,
each integrator would be surrounded by a network o f suppliers and related activities'
Within these complex structures o f interrelationships, new power balances could
emerge, In the lower-tech industries especially, like textiles and some branches of
food, the manufacturing stages o f the “ chain55were squeezed as power tended to shift
downstream to the final stages and even to the retailers, The growing competition in
diversified products in conjunction with the rising incomes and influence of con­
sumers, often dissatisfied with what they are getting, is driving this trend. It remains
to be seen whether this pattern will proceed further in the direction o f buyer-driven
rather than producer-driven chains (Gereffi 1999).

15.5.2 Industry Differences

These observations raise the issue o f how industries differ in their behavior; in
particular, can one observe different patterns in LM T industries than high-tech
industries at the industry and sector levels? Many studies have sought to classify
industry behavior by resorting to the two broad ideal types usually identified with
the name of Schumpeter (see Ch. 14 by Malerba in this volume).
One group o f such studies identifies the key conditions as opportunity, appro­
priability, cumulativeness, and knowledge base (Malerba and Orsenigo 1996,1997).
Using these criteria, the great m ajority o f industries in Europe are classified as Mark I
or Mark II using objective indicators. Thus among the LM T industries, clothing falls
obviously into the category o f M ark I, characterized by low technological oppor­
tunity, weak appropriability o f any innovations, small firms and rapid entry and exit,
and a practical rather than scientific knowledge base. Motor vehicles as a medium-
tech industry have greater technological opportunities, greater appropriability and
the persistence o f large firms. High-tech industries naturally have high technological
opportunities and a scientific knowledge base, though they may be variously typified
by large or small firms.
From the viewpoint o f LM T industries, where demand plays such a large role (see
Section 15.3.1 and case studies below), market opportunity can be as important as
technological opportunity, and may be very different in extent as well as in nature.
Fast-growing areas o f consumption are not always the same as fast-growing areas o f
technology, as noted above. Moreover, for our immediate purposes, some o f the low-
tech industries are Mark I but others are nearer to Mark II, while food-processing
resists any easy classification since its sub-branches operate in a whole variety o f ways.
We have indicated that technological opportunities may be enlarging again for
low-tech” industries (see Sectio n 15.3.2 above), although firm s in such industries
424 NICK VON T U N Z E L M A N N AND V IR G IN IA A C H A

w ill fo r the m o st part outsource the d evelo p m en t o f these n ew technologies.

O u tso u rcin g m ay lim it o p p o rtu n ities fo r u ser firms to a p p ro p ria te the returns
from in n o vatio n that relies on this ap p ro ach (recall that a p p ro p ria b ility w as a key
factor in the Sutton analysis discussed above). In the su p p lie r-d o m in a te d low -tech
indu stries (as defin ed b y Pavitt) the a p p ro p ria b ility o f tech n o lo g ies rests u p on the
d iv isio n o f po w er betw een the tech n o lo gy d evelopers the u p stream suppliers—
and the users, like food or clothing companies. These activities are rarely vertically
integrated, because the suppliers usually wish to supply a variety o f users both within
the same activity and outside. The appropriability o f the products depends on both
the m arketin g en deavou rs o f the co m p an ies con cern ed, an d the p o w e r balan ce vis-a-
vis d ow n stream d istrib u to rs an d retailers. L in ks w ith tech n o lo g y su p p liers tend to
be m u ch m ore distant than in h igh -tech sectors.

15.5.3 Industrial Dynamics

15.5.3.1 Entry, Exity a n d Technological A ccu m u lation
Models o f industrial dynamics in the Schumpeterian tradition often contrast Mark I
and Mark II industries in respect o f the technological advantage o f incumbent firms
over possible entrants (M arsili 2001). The deciding factor can be the extent to which
individual firms in an industry— incumbents or entrants— can earn economies of
scope in combining or differentiating technological trajectories (Sutton 1998).
M any LM T industries are however characterized by high levels o f uturbulence,"
with a churning o f entry and exit. These pose the issue o f learning in turbulent
environments— if the individuals cannot be retained in the industry when exit
occurs, new entrants may simply replicate their predecessors' mistakes. However,
in North American environments at least, the individuals concerned do tend to go
on to form another firm (Baldwin 1995). Alternatively the continuity can be main­
tained by technological dependence on a large supplier or an industrial district, as is
often fo u n d in clothing. In co m p lex p ro d u cts system s, w h ere allian ces are reconsti­
tuted for each new project, learning is achieved by the flux o f Interactions o f the
constituent firms, although a high level o f “ forgetting” also seems to be common
(H o b d a y 1998). The dichotomous M ark I and II categorizations m ay be too restrict­
ive to p o rtra y the m ain patterns o f evo lu tio n o f d ifferen t in d u stries, and m ore
complex schema such as that o f Pavitt (1984) or M arsili (2001) m ay be preferable
for understanding the impact o f differing technologies.

15.5.3.2 D yn a m ic C om petition in T im e
A cross the full range o f industries, the m o d ern era is su p p o sed to be characterized by
co m p etition that has intensified because o f glo balization an d becau se o f m ore rapid
 I N N O V A T I O N IN “ L O W - T E C H ” I N D U S T R I E S 425

change m m arket dem and. In L M T industries the pace o f change and com petition
may be less intense, as m arket leaders seek to retain their predom inance by brand
loyalty. In the tobacco industry, the trajectory o f a new product launch is norm ally to
begin with a new bran d m arketed at low prices and with high tobacco content. Once
loyalty is obtained the produ cer relaxes the prices and/or the m aterial content until
eventually consum ers lose faith, w hereupon a new brand is launched on the same
terms to eclipse it. Equally, brand loyalty can be extended through globalized
marketing.
Thus m arketing-based ap p ro p riab ility offsets som e o f the pressures to develop
new products. H ow ever firm s in L M T industries are arguably alm ost as susceptible
as those in higher-tech industries to innovations that accelerate the developm ent
times (“ cycle tim e” ) an d rate o f application and diffusion for new technologies. In
the clothing industry, Benetton com bined its production netw ork o f small suppliers
with a sophisticated IC T system for feedback from custom ers, achieving rapid global
growth in a relatively m ature in d u stry (Belussi 1987). For the m ore basic production
technologies, the L M T industries are distinguished less by the pressure to innovate
than by the difficulties encountered in app lyin g those basic technologies to purposes
that m ay be very different fro m their origin al intentions. H aving to confront such
difficulties am plifies rather than dim inishes the im portance o f speed to m arket for
competitiveness even in L M T industries. The key issue here is surely the real-tim e
achievability o f econom ies o f scope— do integrated firm s or outsourced networks
effect this better?

15.6 In n o v a t io n in S o m e L o w -Tech
In d u st r ies

15.6.1 Textiles and Clothing

As the original “ leading sector” o f the First Industrial R evolution in Britain, textiles
have often been regarded as the quintessential low-tech industry in the m odem era.
First, however, most low-tech industries today are not structured like textiles and
clothing. Second, textiles themselves have shown a repeated ability to update by
bringing in contemporary developments in technology. In the early twentieth
century the chemicals technologies at the forefront o f the Second Industrial Revolu­
tion were recruited to launch “ artificial fibers” (rayon etc.); in the middle o f the
century “ synthetic fibers” (nylon etc.) drew upon synchronous advances in plastics.
Roth artificial and synthetic fibers remained branches o f the chemicals industry in
426 NICK VON TU NZELM A N N AND VIRG INIA A C H A

terms o f corporate structures as well as technologies and are thus often thought o f
not as textiles proper, but this hardly seems reasonable. In m ore recent times
computerized technology is making inroads into the still m ore fragmented clothing
sector, while genetics is assisting advances in textiles. M icrofibers o f exceptional
fineness have evolved for particular market niches and have begun penetrating
standard clothing segments.
It is indeed the case that the recent developments have been slow to diffuse, but this
probably has less to do with technological limits than with organizational aspects. The
textile-clothing industry is still largely based on a pre-industrial vertical structure that
is highly segmented and allows countries with low wage rates to enter through the use
o f relatively simple technologies, such as sewing machines. To leverage new textile
technologies in the way we have described, change in production technologies may
have to be linked to product changes, as for instance with microfibers.
The demand side normally offers greater prospects for leverage than the supply
(technologies) side in the textiles and clothing sector. As incomes rise, consumers are
willing to pay extra for fashionable brand-names. M any o f the observed changes in
what is fashionable rest on styling rather than technological innovation, though
naturally that does not imply any lesser need for creativity. Even here, however, new
technology may help bridge the gap between designers, producers, and their markets.
An important innovation by Benetton was the firm s use o f IC T to learn quickly about
changes in market tastes in order to instruct their suppliers (Belussi 1987).

15*6.2 Food-processing
A second industry frequently, and often inappropriately, classified as low-tech is
food manufacturing. The “ industry” is characterized by a huge variety o f organiza­
tional forms, related principally to the extent to which marketing can be leveraged
(Sutton 1991). It has been custom ary to see the industry as “ supplier-dom inated”
according to Pavitfs taxonomy, resting on a dependence on suppliers o f the ma­
chinery for production, but this is changing nowadays in some critical ways. As such,
the industry is becoming more market-driven, but one cannot reasonably leave
technology out o f the picture.
Like all industries, the technical efficiency o f this industry comes from the
knowledge bases it draws upon, but here there exists an unusual am ount o f variety
to go with the variety o f organizational structures. The range o f expertise from
science (e.g. food microbiology) through to engineering has to span not just
production conditions but sanitation, quality assurance, environmental acceptabil­
ity, and so on. The range of new technologies currently being drawn upon in
this industry covers almost the whole spectrum. The traditional reliance on suppliers
o f machinery is being overtaken by needs for technologies from advanced
 I N N O V A T I O N IN 4L OW - T E C H >J I N D U S T R I E S 427

instrumentation (e.g. lasers), electronics and computing (problematic because of

the irregular shape o f some o f the leading products), biotechnology (for both
materials and production processes), pharmaceuticals, and smart materials (espe­
cially in packaging), supplied by high-tech firms or public laboratories, The seem­
ingly simple packaging o f readymade and microwavable foods for sale in
supermarkets in fact required very sophisticated analyses o f smart materials to
combine heat responsiveness, gas release (controlled oxygenation), ease of produc­
tion, ease in filling during processing, as well as ease o f consumer use.
Many of these new techniques remain controversial, particularly in the field o f
biotechnology (genetic modification, GM ). At the heart o f these controversies lies an
issue already noted— is a GM food the same product or a different one? Around such
points trade wars rage. As already implied, the nature o f the product as a staple o f
human existence means that safety and quality procedures play a substantial role.
The difficulty with prevailing methods for testing food safety (wet chemistry) is that
they involve cutting off a piece o f the product— which may or may not be represen­
tative— taking it to the laboratory and waiting about three weeks for the results. By
this time the rest o f the product could long since have been sold and consumed.
Biotechnological methods and other procedures are therefore being sought to effect
testing in 'Teal tim er be. synchronously with the processing. Again the main
motivation for technological innovation is time-saving— reducing downtime and
waste and increasing throughput in the interests o f systemic efficiency.
Changes on the demand side (from socioeconomic factors like rising wealth,
growing female employment, ageing o f the population) and the above technological
changes on the supply side are channeled in part through a changing vertical
structure o f the industry, in which growing power is accruing to giant retailers vis-
a-vis large processors. At the time o f writing, the US food retailer Wal-Mart has
become the world's largest company according to the F o r t u n e list, superseding the
high-tech and medium-tech companies o f the recent past. Similarly in Europe,
supermarket chains have been proactively utilizing information technologies for
robust expansion o f their power base. These developments make the predominance
of demand-side influences on innovation here abundantly clear,

15.7 R ole of G o v e r n m e n t P o licy

G o v e rn m e n t te c h n o lo g y p o lic ie s at n a tio n a l a n d s u p ra n a tio n a l levels h a v e o n th e

w h ole te n d e d to g iv e th e h ig h e st p r io r it y to h ig h -te c h sec to rs a n d a c tiv itie s, O n e o f
the m o st c o m m o n ly u se d b e n c h m a rk s fo r th e su ccess o f a g o v e rn m e n t s te c h n o lo g y
428 NICK VON T U N Z E LM A N N AND VIRG INIA A C H A

strategy is the proportion o f total output emanating from high-tech sectors, in

which dimension East Asian countries tend to score very highly (see Ch. 19 by
Fagerberg and Godinho in this volume). Such a benchm ark falls into the trap,
emphasized throughout this chapter, o f confusing the sector with the technology
level. Success in low-end activities in high-tech industries, such as “ screwdriver”
assembly activities, m ay or may not lead on to success in high-end activities in those
sectors— often it m ay sim p ly generate co m p etitio n fro m n ew riv als in the low -end
activities, based o n p ro vid in g still cheaper labour.
A s n oted in C h ap ter 22 b y Lu n d vall and B o rras in this v o lu m e , go vern m en t policy
tow ard s in n o vatio n w as fo r lo n g d om in ated b y the so -called lin ear m o d el” o f
tech n o logy-p u sh . A lth o u g h this ap p ro ach does n o t p reclu d e ap p licatio n s o f high
tech n o lo gy to m ore trad ition al fields, the bias in p ractice u su a lly veers tow ards new
fields as w ell as new technologies. F o r m an y decades the b ias to w ard s such technolo­
gies w as also coupled to one tow ards co n so lid atio n o f firm s an d in d u stries, seen as
“ n ation al ch am p io n s” fo r tech nological and co m m ercial success. T h is v ie w has been
w eaken ed in recent tim es as sm all and m ed iu m -sized en terprises (S M E s) have been
in creasin gly th ou gh t o f as p ro gen ito rs o f h igh tech n ology, in fields such as biotech­
n o lo g y and genom ics, softw are, advanced in stru m en tatio n , an d so o n . Yet in low-
tech industries, SMEs are most widely seen as dragging down overall performance.
The antithesis o f the “ supply-push” contribution o f the linear model is the
“ dem and-pull” approach to innovation, in which the causal sequence is roughly
reversed. Governments, though, tend to be loath to “ leave it all to market forces,” in
an arena in which “ market failures” are so evidently present as they are in innovative
activities. An intermediate position is for governments to foster diffusion rather
than upstream invention and innovation (Stoneman and David 1986). While this
m ay not overcome all the market failure shortcomings, there is m uch to be said for
rebalancing policy in such a way as to place greater stress on the diffusion aspects. O f
particular relevance to this chapter would be the diffusion o f high-tech activities into
supposedly low-tech sectors. As we have observed, there is a m uch wider scope for
such migration o f technologies across sectoral boundaries than is often supposed.
M a n y coun tries have in fact based their d evelo p m en t o n startin g in supposedly
low -tech activities like the fo o d -p ro cessin g o r textile in d u stries su rveyed a b o v e -
D en m ark, Sw itzerland, and A u stralia are evident exam p les (vo n T u n zelm an n 1995).
A t the sam e tim e, a perhaps even larger n u m b er o f co u n tries have becom e locked
into the low -tech activities and never adequ ately escaped, as is o ften alleged for Latin
A m erica. T h e d ifficu lty rem ains that co u n tries’ co m p arative advantages often
rem ain in d issolu b ly linked to the low -tech sectors. O u r v ie w that there are no true
low-tech sectors in the modern world however overrides this. It is perfectly possible
to diffuse high technologies into the “ low-tech sectors,” as our illustrations o f chosen
low-tech sectors aimed to underline. In that way intermediate and developing
countries do not need to face a dilemma when choosing between static comparative
 I N N O V A T I O N IN “ l O W - T E C u ” I N D U S T R I E S
429

advantage in traditional fields and dynamic comparative advantage from techno­

logical opportunity— they can have their cake and eat it.
This would involve moving away from usual defeatist views of traditional sectors
as “sunset industries ” As Ernest Hemingway observed, “ the sun also rises." The sun
can rise both from downstream diffusion o f high-tech activities into “ low-tech
sectors," and from the lateral diffusion o f old technological fields into new ones.
Countries like Switzerland developed, consciously or unconsciously, a pattern of
evolution that went— in this case— from textiles to dyeing and chemicals and then
into strength in pharmaceuticals on the one hand, and into machinery and thence to
advanced engineering on the other.
A clear implication o f these policy considerations is that strength in low-tech
sectors does not have to act as a “ block to development,” although national and
supranational governments frequently embody this error in their technology pol­
icies. The EU, for instance, counterposes the Framework Programs and the Struc­
tural Funds as means for overcoming development-oriented market failures, but it is
conceivable to think o f reconciling the two, and using the Structural Funds as means
for implementing innovation in catching-up regions (Fagerberg et al. 1999). The
“block to development” view can be replaced by a “ development block” view, in
which LMT sectors act as “ carrier industries” for diffusing the gains from new
technologies across the industrial spectrum, as they have so often managed to do
in the past (Rosenberg 1976).
We do not pretend that this has yet happened on any major scale, though it is
significant in some services and a number o f older branches o f manufacturing. On
the contrary, on analogy with past “ long waves” o f technological innovation and
industrial growth, it is the key challenge for the next few decades.

15*8 C o n c l u s i o n s

The underlying argument o f this chapter is that low-tech sectors do not lack for
technological opportunities, nor indeed for appropriability and other factors asso­
ciated with benefiting from technological innovation. Indeed, our principal analyt­
ical conclusion is that in the modern world there are no true low-tech sectors.
Instead what we observe is a varying degree o f permeation o f high technologies into
low-tech and medium-tech as well as into high-tech sectors.
The primary issue for further research is thus the need to explore much more
critically how industries and sectors ought to be classified. The OECD itself appears
 NICK VON T U N ZE L M ANN AND V IR G IN IA A C H A

to be m oving towards replacing its conventional accounting o f direct and indirect

R & D with a m ore subtle assessment o f the knowledge intensity o f industries (OECD
2003; see Ch. 6 by Smith, for further discussion). This will have substantial implica­
tions for government policy making but also for analytical understanding o f struc­
tural change and the sources o f long-term development. We w ould suggest going
further and aiming to think not just o f im proving our conception o f how technolo­
gies can be clustered (as for instance in the Pavitt taxonom y) but also o f how
technologies map into products. Some pioneering w ork has been published by
individual scholars in this regard (e.g. Piscitello 2003) but this w ork needs support
from national and supranational agencies to instil the notion that demand as well as
supply factors drive technological change and long-term prosperity. The LMT
industries bulk far too large in total output to be relegated to the scrap-heap of
stagnation.

N otes

1. Fram es should not be confused w ith organization al “ cu ltu re” ; the latter com prises the
values established at the group level and interpreted by in d ivid u als, w hereas frames
com prise the interpreted relationships between ph en o m en a (and their associated values)
and are com posed o f individuals, m ediated at a group le v e l

R eferen c es

A c h a , V. L. (2002), “ Framing the Past and Future: The Development and Deployment of
Technological Capabilities in the Upstream Petroleum Industry,” Unpublished D.Phil.
Thesis, SPRU, University of Sussex.
and B r u so n i , S. (2003), “ C o m p lexity and In d u stry E vo lu tion : N ew in sights from an
O ld Industry, in Best Papers Proceedings o f 2002, E u ropean A sso ciatio n for Evolutionary
Political E co n o m y Conference, 2003.
B a b a e v , H . (2000),
Scientists, Engineers Team ing up to D evelop M ars D rillin g Technol­
ogy,” reprinted in O il a n d G a s Jo u r n a l 98 (17).
B a ld w in , J. R. (1995I T h e D y n a m ic s o f In d u s t r ia l C o m p e titio n : A N o r th A m e r ic a n Perspective ,
Cambridge: Cambridge University Press.
and S a b o u r in , D. (2002), A dvanced Technology Use and Firm Perform ance in
C anad ian M an u facturin g in the 199ns, In d u s t r ia l a n d C o rp o ra te C h a n g e 11: 761—90.
B e l u s s i , F. (1987), Benetton, In form atio n le c h n o lo g y in P ro d u ctio n an d Distribution:
A Case Stu d y o f the Innovative Potential o f T raditional Sectors,” S P R U O ccasional Papers
25, SPRU , U niversity o f Sussex.
 I N N O V A T I O N I N " L O W - T E C H ” I N D U S T R IES 431

Chandler , A. D, jr. (1977)) T h e V isib le H a n d : T h e M a n a g e ria l R evo lu tio n in A m erica n

Business, Cam bridge M ass,: Belknap Press.
-— - (1990), Scale a n d S c o p e . T h e D y n a m ic s o f In d u s tria l Capitalism ^ Cambridge, Mass,:
Belknap Press.
C W. M ., and L e v in t h a l , D. A. (1989), “ Innovation and Learning: The Two Faces o f
o h e n ,

R W f E conom ic Jo u r n a l 99: 569-96.

------------ (1990), “A bsorptive C apacity: A New Perspective on Learning and In n ovation ”
A dm inistrative Scien ce Q u a rte rly 3 5 :12 8 -5 2 .
* D S., and L yo ns , B. (199b), In d u s tr ia l O rg a n iz a tio n in the E u ro p ea n U n io n : S tructu re ,
a v i e s ,

Strategy a n d the C o m p e titiv e M e c h a n is m , Oxford: Oxford University Press.

Dodgson, M ., and R oth w ell , R. (eds.) (1995), T h e H a n d b o o k o f In d u stria l In n o v a tio n ,
Cheltenham: Edw ard Elgar.
*Fagerberg , J., G u e r r ie r i , P , and V e r sp a g e n , B. (eds.) (1999), T h e E co n om ic C h a llen g e o f
Europe: A d a p tin g to In n o v a tio n -B a s e d G r o w t h Cheltenham : Edward Elgar.
^Freeman , C., and Lou^A, E (2001), As T im e G oes B y : Fro m the In d u stria l R evo lu tio n s to the
Inform ation R e v o lu tio n , O xford: O xford U niversity Press.
----- and P erez , C. (1988), “ Structural Crises o f Adjustm ent: Business Cycles and Investment
Behaviour,” in G. D osi, C. Freem an, R. N elson, G. Silverberg, and L. Soete (eds.), Technical
Change a n d E co n o m ic T h e o ry , Lon don: Pinter, 38-66.
G ereffi , G. (1999)» “ International Trade and Industrial U pgrading in the Apparel C o m ­
modity Chain,” Jo u r n a l o f In te rn a tio n a l E co n o m ics 4 8 :37-70 .
G rossman , G. M ., and H e l p m a n , E. (1991), In n o v a tio n a n d G ro w th in the G lo b a l E co n om y ;
Cambridge, M ass.: M IT Press.
Helpman , E. (ed.) (1998), G e n e ra l P u rp o se Technologies a n d E co n o m ic G r o w th Cam bridge,
Mass,: M IT Press.
* Hirsch -K r ein sen , H., J acobso n , D., L a est a d iu s , S., and S m ith , K. (2003), “ Low-Tech
Industries and the Know ledge Econom y: State o f the A rt and Research Challenges,”
mimeo, EU 5th Fram ew ork project, “ Pilot: Policy and Innovation in Low-tech.”
Hobday, M. (1998), “ P roduct Com plexity, Innovation and Industrial Organisation,”
Research P olicy 26: 689-710.
Malerba , R, and O rsen ig o , L. (1996), “ Schum peterian Patterns o f Innovation,” C a m b rid g e
Journal o f E co n om ics 19: 47-65.
* --------- (1997), “ Technological Regimes and Sectoral Patterns of Innovative Activities,”
Industrial a n d C o rp o ra te C h a n g e 6: 8 3-117.
Mansfield , E. (1969), In d u s tr ia l R esearch a n d T echnological In n o v a tio n : A n E co n om etric
Analysis , New Haven: Yale University Press.
*Marsili , O. (2001), T h e A n a to m y a n d E vo lu tio n o f In d u stries: Technological C h a n g e a n d
In du strial D y n a m ic s , Cheltenham: Edward Elgar.
Meliciani , V. (2001), T echn ology ; T ra d e a n d G ro w th in O E C D C o u n tries: D oes Sp ecia lisa tio n
M atter? London: Routledge.
Mowery, D. C. (1995), “ The Boundaries o f the US Firm in R8cD, in N. R. Lam oreaux and
D. M. G. R a ff (eds.), C o o rd in a tio n a n d In fo rm a tio n : H isto rica l Perspectives on the O rg a n iz a ­
tion o f E n terp rise , C hicago: U niversity o f Chicago Press, pp. 147-82. *

* Asterisked items are suggestions for further reading.

432 NICK VON TU N ZELM A N N AND V IRG IN IA ACHA

N elso n , R, R. (ecL ) (1993), National Innovation Systems: A Comparative Analysis, New York:
Oxford University Press.
OECD (1994). The Measurement of Scientific and Technical Activities: Proposed Standard
Practice for Surveys of Research and Experimental Development (Frascati Manual: 1993)»
Paris: OECD,
------ (2003), Science, Technology and Industry Scoreboard 2003 - Towards a Knowledge-Based
Economy Paris: OECD.
Q r l i k o w s k i , W, J. (2000), “ Using Technology and Constituting Structures: A Practice Lens
for Studying Technology in Organizations,” Organization Science 11: 404-28.
------ and G a s h , D. C. (1994), “ Technological Frames: Making Sense of Information Tech­
nology in Organizations,” ACM Transactions on Information Systems 2 : 174-207.
*P a v it t , K. (1984), “ Sectoral Patterns of Technical Change: Towards a Taxonomy and a
Theory” Research Policy 13: 343-73.
*P e n e d e r , M. (2001), Entrepreneurial Competition and Industrial Location, Cheltenham:
Edward Elgar.
P en ro se , E. T. (1959), The Theory of the Growth of the Firm, 3rd edn., Oxford: Oxford
University Press, 1995.
P isc it e l lo , L. (2003), “ Generation and Applicability of Technological Competencies: An­
other way of Measuring Coherence in Corporate Diversification,” paper for Conference in
Honour of Keith Pavitt, Politecnico di Milano.
P o r t e r , M. E. (1985), Competitive Advantage: Creating and Sustaining Superior Performance,
New York: Free Press.
R o sen ber g , N. (1963), “ Technological Change in the Machine Tool Industry, 1840-1910”
Journal of Economic History 23: 414-46.
------(1976), Perspectives on Technology, Cambridge: Cambridge University Press.
R u m elt , R. (1974), Strategy; Structure and Economic Performance, Boston: Harvard Business
School Press, rev. edn., 1986.
S to n em a n , Р. L., and D a v id , P, A. (1986), “Adoption Subsidies vs. Information Provision as
Instruments of Technology Policy”, Economic Journal, Conference papers, 96:142-50.
S u tto n , J. (1991), Sunk Costs and Market Structure, Cambridge, Mass.: MIT Press.
*------(1998), Technology and Market Structure, Cambridge, Mass.: MIT Press.
T id d , J., B e ssa n t , J., and P a v it t , K. (2001), Managing Innovation: Integrating Technological,
Market and Organizational Change, Chichester: Wiley.
* v o n T u n zelm a n n , G. N. (i995)> Technology and Industrial Progress: The Foundations of
Economic Growth, Cheltenham: Edward Elgar.
W o m ack , J. P., J o n es , D. T. and Roos, D. (1990), The Machine that Changed the World, New
York: Rawson.
 C H A P T E R 16

I N N O V A T I O N IN
SERVICES

IAN M ILES

16.1 In tro d uctio n

Se r v ic e s innovation is a topic o f growing interest for innovation researchers and

policy makers alike. This has particularly been so in the last decade, as services have
grown to constitute the larger part o f employment and output in most industrial
countries. The services sectors o f these economies are important for their product­
ivity, economic competitiveness, and quality o f life. But innovation in services is
important for other reasons beyond the economic importance o f the sector. Innov­
ation in services extends beyond the services sectors to affect service activities in all
sectors o f the economy. Second, some services play central roles in innovation
processes throughout the economy, as agents o f transfer, innovation support, and
sources o f innovations for other sectors.

16.2 Services— G row th,

C h a r a c t e r is t ic s , and In n o v a tio n

Output and employment in the service sectors have grown significantly throughout
the industrial world since the 1950s. Table 16 л shows that as early as the 1970s, services
434 IAN M ILES

Table 16.1 Share ( % ) of gross value added Inservices in total GDP in EU countries

EC-9 Ec- 1 0 Ec- 1 2 EU-15 EU-15

(1973) (1981) (1986) (1995) (2001)

EC/EU 50.7 56.3 59 §3,1 >' 65,3

Ireland 44.5 49,4 50,4 ■ 48.6
Finland n/a 49.8 53 56,4 57,1
Portugal n/a 48.8 54,1 59.1 61
Austria n/a 55.5 ;■ 58.6 63.3 63,6
Denmark 58.8 60.9 60,3 64 64
Spain 47,8 57.3 58 63.7 64,2
Germany 48.1 54.2 55,9 62.5 64.9
Ita ly 51.1 56.4 60.6 62.8 649
Greece 40.3 47.7 48.9 62.6 65,4
Sweden n/a 60.1 59.7 62.2 65.5
N etherlands 52.2 58.8 60.4 63.9 65.5
France 50.8 57,4 61.1 65.3 66,8
UK 54.7 54.5 57.1 62.1 67.2
Belgium 54.6 60.7 62,8 66.1 67.3
Luxembourg 46.9 65.2 76.4 83.7 83.8

Sourer, Eurostat (2003), 50 years o f figures on Europe, Luxembourg: European Communities.

constituted more than half o f the value added in European Union countries, and by
the new century they contributed over two-thirds.

16*2.1 Services Diversity— and Distinctiveness

The rates o f growth in services output and employment vary am ong the nations in
Table i6n> and in the case o f Ireland, the share o f output attributable to services
declined during 1986—2001. These differences reflect contrasts in the composition of
the services sector in different countries, highlighting the fact that the category
service sector comprises a huge range o f different activities with very different
characteristics. At one extreme are personal services like hairdressing that involve
basic technologies and often are organized on a small-scale basis. In contrast, the
FIRE (finance, insurance, and real estate services) sector is dom inated by very large
firms using advanced information technologies intensively. Other technologies are
used in distributive services, which include transport in all its varieties and retail and
wholesale trade and, in some classifications, telecommunications and broadcasting.
 I N N O V A T I O N IN S E R V I C E S

HORECA (hotels, restaurants and catering) is dom inated by food preparation and
delivery, and includes other elem ents o f hospitality, entertainm ent, and com fort.
Social and collective services such as public adm inistration and health and educa­
tional services are delivered largely or entirely through the state, though patterns o f
organization v ary a great deal over tim e and across countries: the back-office
operations o f such bodies can be h igh ly IT-intensive. Business services include
practical support w ith logistics and office and building services, as well as support
for adm inistrative m atters (such as law and accountancy) and technology
support, such as com pu ter and engineering services. This sum m ary by no m eans
exhausts the variety o f activities and technologies included in the services sector.
Services m arkets also are diverse, spanning consum ers, businesses, and the public
sector and its clients. The tran sform ation s effected by these services industries
operate on such diverse “ raw m aterials” as hu m an clients (as well as som e other
biological organism s, e.g. veterinary services), physical artifacts (they m ay be
repaired, m aintained, stored, tran sported, tested, integrated into larger system s. . . ) ,
and data, sym bols and in fo rm atio n (that m ay be processed, stored, telecom m uni­
cated, etc. b y services like financial industries as well as b y com puter and com m un i­
cations services, etc.). T h e technical skills dem anded o f the w orkforce range from the
minimal ones used in fast fo o d outlets and office cleaning to the professional
qualifications o f m arket researchers and architects, and the scientific and engineer­
ing credentials o f sta ff in specialized R & D firm s.
This diversity m eans that any generalizations about the nature o f services and
innovation in services m ust be qualified b y num erous exceptions. Som e services are
more like m an ufactu rin g in term s o f som e param eters— som e are technology­
intensive (e.g. m edia, telecom m unications), som e w ork w ith m aterial artifacts
(e.g. rapid pro to typ in g, repairs). A n d the operations o f m any m anufacturing
firms include a great m an y “ services” activities (e.g. transport and logistics, office
work, m arketing and aftersales). N evertheless, a set o f com m on features charac­
terizes m any services and differentiates them from m anufacturing.
For exam ple: m an y services products are intangible ones, w hich m akes it harder to
store, transport, and exp ort them than is true o f m anufacturing products. H istoric­
ally, m any service in n o vation s have been difficult to protect via patent m echanism s,
although this situation has begun to change in som e sectors such as com puter
software and services, or F IR E (See H all 2003; G raham and M ow ery 2003, FhG . 151
2003). Services are typ ically interactive, involvin g high levels o f contact between
service supplier and client in the design, produ ction, delivery, consum ption, and
other phases o f service activity. Service products are often produced and consum ed
in the course o f su p p lier-clien t interaction at a particular tim e and place ( coter-
minality” ). In novations m ay focus on this interaction as m uch as on conventional
product and process characteristics, and m ay rely less on technical know ledge and
more on social and cultural nous. M an y services are highly inform ation-intensive,
with a preponderance o f office-based w ork or com m unicative and transactional
43б IAN MILES

o peratio n s, such as telem arketing. Som e service p ro d u cts are d eliverab le electronic­
ally, such as text-based reports, T V p ro g ram s, m u sic reco rd in gs, c o m p u te r software,
an d w ebsites. T h e in fo rm atio n al co m p o n en ts o f m a n y o th er services are subject to
IT -based in n o vation . B u t in oth er respects, the con trasts w ith in the service sector are
as sign ifican t as those d ifferen tiatin g the sector fro m m a n u fa c tu rin g . A fter all, the
sector in cludes the m o st concentrated, k n o w led g e-in ten sive, an d IT-intensive
sectors in m o d e m in d u strial econ om ies (b an kin g, p ro fe ssio n al services, etc.), as
w ell as the least (retail, cleaning, etc.).

16.2*2 Research on Innovation in Services

D espite the econ o m ic im p o rtan ce o f the sector, in n o v a tio n in services received little
system atic attention until the 1980s, w h en a trickle o f studies ap p eared that focused
p rim a rily o n services an d new IT. In the 1990s, a n u m b e r o f m a jo r research projects
on services in n o vatio n w ere lau nch ed , and som e services w ere in clu d ed in R & D and
in n o vatio n surveys b y the end o f the tw entieth century.
A lth o u gh m an y recent surveys reveal grow th in R & D in vestm en t w ith in services,
R & D surveys m ay u nd errepresent the in n o vative activities o f service firm s. The
increases in service-sector R & D revealed b y m a n y o f these su rveys m a y reflect
(a) increased coverage o f services firm s b y su rveys (services w ere excluded from
m an y n ation al R & D surveys until the 1990s), and (b) reclassificatio n o f som e large
firm s' activities into the services sector. T h e U S N atio n al Scien ce Foundation's
su m m a ry o f recent trends in service-sector R & D investm ent, trends that incorporate
both o f the effects noted above, nevertheless suggests so m e system atic change in
service-sector R & D activities, as w ell as im p o rta n t d ifferen ces a m o n g the Triad
econ om ies in service-sector R & D ;

R&D performance by the U.S. service-sector industries underwent explosive growth between
1987 and 1991, driven primarily by computer software firms and firms performing R&D on a
contract basis. In 1987, service-sector industries performed less than 9 per cent o f all R&D
performed by industry in the United States. During the next several years, R& D performed in
the service sector raced ahead o f that performed by U.S. manufacturing industries, and by
1989, the service sector performed nearly 19 per cent o f total U.S. industrial R&D, more than
double the share held just two years earlier. By 1991, service-sector R&D had grown to
represent nearly one-fourth o f all U.S. industrial R&D. Since then, R&D performance in
U.S. manufacturing industries increased and began growing faster than in the burgeoning
service sector... Unlike the United States, Japan has yet to see a dramatic growth in service-
sector R & D . . . R&D in Japan s service-sector industries reached 4,2 per cent o f the total R&D
performed by Japanese industry in 1996 and 4.5 per cent in 1997
R&D within the EU s service sector has doubled since the mid-1980s, accounting for about
11 per cent o f total industrial R&D by 1997* Large increases in service-sector R&D are apparent
in many EU countries, but especially in the United Kingdom (19.6 per cent of its industrial
 I N N O V A T I O N IN S E R V I C E S 437

R&D in 1997)3 l^ly (15.3 percent), and France (10.0 percent). (National Science Foundation
2002: ch. 6)

This picture o f service-sector R & D investm ent underm ines the view that innov­
ation in services results solely from the sector’s adoption o f m anufacturing
innovations.1 A bun dant case-stu d y evidence also highlights innovation by services
firms (see e.g. A ndersen et al. 2000; T idd and H ull 2003). Indeed, som e service
industries invest h eavily in R8cD and pursue R & D program s that are at least as
sustained as are those o f m an u factu rin g. Som e useful inform ation on corporate
R&D can be gleaned from co m p an y annual reports and accounts, and C om pan y
Reporting Ltd. produces an annual analysis o f these data for the U K 's D epartm ent o f
Trade and In dustry (D T I). The 2003 analysis (D T I 2003) reports that ITservices rank
fifth among all industries in w o rld w id e R & D spending in 2003, and sixth am ong U K
industries in 2003 R & D spending. The analysis in this report o f R & D spending by IT
firms indicates that M icro so ft ranks eleventh in w orldw ide tabulations o f corporate
R&D spending, British Telecom is the fifth largest R & D spender in the U K , and
Reuters the eleventh. Fu lly 6 per cent o f the total R & D spending b y the world $
largest R & D investors— the top 700 com panies, w hose R & D spending exceeded
£35 million in 2002— is accounted for b y firm s classified as m ainly involved in
software and IT services. (T h e equivalent figure for the U K alone is 5 per cent.)
Data from the C o m m u n ity In n o vatio n Survey (C IS) also indicate that services
firms are m ajor innovators, and that it is not just “ high-tech” services (like software
and telecom m unications) w hich p lay significant roles in technology developm ent.
Although there are in n ovative activities in all branches o f services, the innovation
surveys depict low er levels o f Innovation and R & D investm ent w ithin services, on
average, than w ithin m an u factu rin g. The adoption by services firm s o f technologies
produced in other sectors is indeed a m ajor form o f innovation in the sector, m uch
of which thus displays “ su p p lier-d riven ” characteristics.

16*2*3 Services Attributes and Innovation Trajectories:

Industrialization and Modularization
Such patterns o f in n o vatio n m ay be related to the ways in w hich (m ost) services
differ from (m ost) m an u factu rin g firm s. H ow m ight the distinctive characteristics
o f service-sector firm s p rodu ce distinctive patterns o f innovation? Services inter­
activity m eans that their produ cts are often custom ized to particular client needs.
Historically this characteristic has entailed the provision o f services on a sm all scale
and local basis. Over thirty years ago Theodore Levitt (i 9 7 2) argued that service
firms needed to adopt a “production line approach,” emulating industrial practices
and moving toward mass production of standardized products, a m ore refined
43$ IAN M ILE S

d iv isio n o f lab o r and h igh er levels o f technology. In fact m a n y services have been
h ig h ly standardized and tech n o logy-in ten sive fo r a lo n g tim e c o n sid e r railways
an d con ven tion al teleco m m u n icatio n and b ro ad cast services.2 B u t it is app aren t that
the grow th o f large-scale firm s in o th er service sectors is asso ciated w ith a form o f
increased stan d ard ization . M c D o n a ld ’s an d o th er fast fo o d restau ran ts are one
fam iliar exam p le o f such firm s in the service sector. Fast fo o d ch ain s also display a
m easure o f cu sto m izatio n , in that their p ro d u cts are co m p o se d o f v a rio u s co m p o n ­
ents, o r m odules, w h ich can be co m b in ed in n u m ero u s w ays a cco rd in g to custom er
d em an d ; n ew m odu les m a y be added to increase v ariety and su p p o rt o th er form s o f
in n o vatio n .
T h e in d u strializatio n o f services has been criticized as gen eratin g low -quality,
lo w -skill jo b s (M cD o n ald izatio n , o r M cjo b s, to so m e co m m en tato rs). O ther types
o f in n o vatio n and reorgan ization in services m ay p ro d u ce n ew fo rm s o f social
exclusion. F o r exam ple, the use o f call centers an d o th er elem ents o f banking
a u to m atio n often occu rs in tandem w ith the closu re o f tra d itio n al outlets, such as
h igh street branches. O u t-o f-to w n h yperm arkets have h ad an ad verse effect on high
street sh o p p in g areas an d their environs. Fears have been expressed that consum er
e-com m erce w ill have sim ilar effects, since these n ew ch an n els o f cu sto m er contact
m a y not be available to all consum ers. Such co n cern s m a y trig g er consum er,
regu lato ry o r legislative responses that w ill affect service m ark ets an d innovation
strategies.
B u t exactly w hat is con n oted b y the “ in d u strializatio n ” o f services cited b y Levitt?
A fter all, m an y m an u factu rin g firm s n o w em phasize flexib le sp ecialization , mass
cu sto m izatio n , reintegration o f h igh ly atom ized d iv isio n o f lab or, a n d the like. Som e
o f these trends are m ak in g m an u factu rin g m o re like services an d so m e are being
em ulated b y large services co m pan ies, even as oth er service o rg an izatio n s continue
to fo llo w m o re classical in d u strializatio n trajectories.
O ne w ay in w h ich services have em ulated m a n u fa ctu rin g is in the ad o p tio n and
d evelopm en t o f an organ ization al in n o vatio n — q u ality co n tro l p ro ced u res. In ser­
vice firm s, as has been tru e o f m an y m an u factu rin g firm s, atten tio n to q uality has
often served as a trigger to in n o vatio n , b y req u irin g firm s to v ie w th eir services as
consistin g o f a n u m b er o f co m p o n en t parts to w h ich q u a lity co n tro l prin ciples can
be applied. T h is is typ ically used to id en tify areas w h ere there is w eakness in service
p erfo rm an ce and w here change thus needs to be engineered. In fo rm a tio n Technol­
o gy is often in trodu ced in the context o f im p ro v in g cu sto m er service quality,
especially in speeding responsiveness v ia m eans such as call centers. U n derstan ding
the co m po n en ts o f a service also often generates in sigh ts in to w ays in w h ich these
co m pon en ts can be tran sfo rm ed o r recon figu red into n ew service b u n d les (Sundbo
1998). M o d u larizatio n underpin s m uch services in n o va tio n , since d eco m p o sitio n o f
service processes and /or produ cts m ay sp u r process in n o va tio n an d the identifica­
tion o f new p ro d u cts and pro d u ct co m b in ation s. T h is typ e o f in n o vative activity
does not necessarily rely on R & D investm ent,3 th o u gh in the sectors w h ere there has
 I N N O V A T I O N IN S E R V IC E S 439

been much talk o f u n b u n d lin g — softw are and telecom m unications— there is
much conventional R & D .
Another m ajor im petus to “ services industrialization” has been the application o f
IT. This has m ade it possible to autom ate elem ents o f the back-office w o rk o f m any
service firms and large firm s in o th er sectors— for instance through the use o f
document processing, em ail, E nterprise Resource Planning software and systems,
etc. IT is widely applied in back-office w ork, and in the m anagem ent and execution
of customer-facing services such as m ail and telephony. M any o f these IT applica­
tions make possible the pro vision o f custom ized services through recom binations o f
standard service m odules. B ack-office autom ation through IT also has changed the
spatial location o f service activities. Telephone call centers— dedicated offices where
the work revolves aro u n d the com puter-assisted answ ering o f telephone calls,
normally for the p ro visio n o f rou tin e custom er service inform ation— are a case in
point.4 Over the last decade we have seen the relocation o f call centers in the U K to
lower-wage areas o f the cou ntry, and m ore recently to the Indian subcontinent— a
specific m anifestation o f the process o f “ offshoring.” There is considerable debate as
to the extent to w hich higher-level office service w ork will also be subject to
offshoring— som e quite sophisticated softw are activities have follow ed m ore basic
programming w o rk in the m ove to overseas locations. The new international
division o f labor clearly involves a redistribution o f service as well as m anufacturing
and extractive activities.5

16.2.4 Services Diversity and IT-related

Innovation Trajectories
IT has been very w idely applied across service sectors— indeed there is d isp rop or­
tionately m ore investm ent in IT fro m services than from m anufacturing (which has
given rise to som e discussion o f the “ p ro d u ctivity p arad o x” as a services p h en om ­
enon).6 In som e cases, these applications involve little m ore than adopting m obile
telephones, personal com puters, and sim ilar devices and their supporting software
and services. In m an y cases, how ever, IT applications require significant innovative
effort, as substantially new app lications o f the IT are developed. Large service users in
sectors like finance and retail have invested huge sum s in developing sophisticated
networks and new system s fo r capturing, archiving and analyzing data. Large
numbers o f graduates in IT -related subjects have been absorbed by these sectors,
to the extent that som etim es there have been com plaints from m anufacturers that a
skill shortage has been created.7
In m any ways IT has p ro vid ed a tech nology that can be applied to the generic
inform ation-processing activities o f services, m uch as earlier revolu tion ary in n ov­
ations in energy tech n o lo gy (e.g. the steam engine, or electric pow er) could be
ap p lied to generic m aterials-p ro cessin g activities in m a n u fa c tu rin g , IT is by no
m eans the o n ly tech n o lo gy em ployed in services, n o r the o n ly tech n o logical field
in w h ich services firm s are active in n o vato rs. M ed ical services an d specialized
b io tech n o lo g y service firm s are m a jo r users o f g en o m ics an d p o st-g en o m ics know ­
ledge an d techniques, fo r exam ple. B u t IT is p ervasive acro ss services, an d the uptake
o f n ew IT stim ulated a reco gn itio n that services w ere o ften u sers o f in n o va tio n and
indeed in n o vato rs in their ow n right. R ich a rd B arras (19 8 6 ,19 9 $ ) n o ted th at in many
w ays the IT revo lu tio n w as an in d u strial revo lu tio n in the service sectors. H e argued
that IT -b ased services in n o vatio n has fo llo w ed a pattern that d iffers fro m that
u su ally depicted fo r m an u factu rin g. In co n trast to the classic p ro d u c t cycle in
m an u factu rin g, he suggested that services in n o v a tio n — o r m o re precisely, their
IT -b ased in n o vatio n — typ ically follow s a “ R everse P ro d u ct C y cle ” (R P C ).
B arras argu ed that the R P C involved three ph ases— Im p ro v e d Efficiency,
Im p ro v ed Q uality, an d N ew Services. IT was first in tro d u ced to im p ro ve existing
processes, and o n ly later becam e the basis fo r service p ro d u ct in n o va tio n , reversing
the “ p ro d u ct cycle” m o d el o f m an u factu rin g in n o va tio n p o p u larize d b y A bernathy
and U tterback (19 78).8 F o r “ van g u ard services” (the tech n o lo g ica lly sophisticated
sectors such as fin an cial services, fo r instance), B arras su ggested that these three
phases ro u g h ly characterized the 1970s, 1980s, an d 1990s respectively. T h u s, insur­
ance services m o ved fro m co m p u terization o f p o lic y record s, to p ro v id in g online
p o licy q u otatio n s, and then to su p p lyin g com plete o n lin e services d u rin g these
three decades. A sim ilar accoun t applies as w ell to service fu n ctio n s ou tsid e o f the
service sector. M an y new telem atics service firm s— su ch as E D S , w h ich em erged
fro m G en eral M o to rs— were organized aro u n d m a n u fa c tu rin g firm s5 in-house
in n o vatio n s in co m m u n icatio n s and data m an agem en t processes that enabled
these firm s to stream line their internal co m m u n ica tio n activities. Eventually, these
service activities w ere spun o ff as IT service com pan ies. S o m e o n lin e in form ation
services origin ated fro m in -h o u se data m an agem en t services, e.g. fo r publishing
firm s.
T h e analysis o f B arras has influenced a large b o d y o f sch o larsh ip that highlights
the contrasts betw een services in n o vatio n trajectories and th ose o f m an u factu rin g
(B arras 1986, 1990). M a n y services m ay rem ain n o n -in n o v a tiv e, but increasing
n u m bers o f IT adopters are exp erim en tin g w ith ap p licatio n s an d related in n o v­
ations, thereby p ro vid in g new arenas for testing the im p lica tio n s o f the Barras
fram ew o rk .9
C ritics o f the R P C m odel have high lighted a n u m b er o f issues;

• counterexamples (e.g. in IT-based services, which often begin with product

innovations);
• conceptual difficulties (the blurring o f production and consum ption makes it
difficult to establish a point in time at which innovation shifts from efficiency
enhancement to product innovation);
 I N N O V A T I O N IN SERVICES 441

* the historically specific nature o f the story (after assim ilating new IT and going
through an R P C process, w ill services then follow the classic product cycle?).
Uchupalanan (1998, 2000) m ounted a system atic critique o f the R P C approach,
tracing the history o f five IT in n o vatio n s10 through all firm s in the banking services
sector in Thailand. He uncovered a diverse range o f innovation strategies and
trajectories that were far richer than the R P C account. The banks were influenced
by the strategies o f co m petitors w ith respect to each given innovation, b y their
experiences w ith earlier in n ovation s (and their plans for others), and b y pressures
from regulators and the m arket. The interrelation o f m arket com petition, firm
circumstances and in n ovation dynam ics m eant that the R P C “ sto ry” o f innovation
processes was rarely applicable in this context. At best it was one o f a num ber o f
possible patterns o f developm ent.
The R PC approach, w hich has proven influential in m any recent analyses o f
services innovation, m ay neglect n o n -IT innovations and innovations undertaken
in earlier periods o f service innovation. But this fram ew ork's attention to the
increasing innovative activities o f service firm s, as well as the historic om ission o f
their innovative activities from m ost scholarship, m eans that the R P C is an im p o rt­
ant contribution, p ro vid in g a starting point for further research on services in n ov­
ation.
Other studies o f services in n ovation place m ore em phasis on the relation between
service firm s and their clients. O ne line o f w ork uses the term “ servuction” (services
production, intended to high light analogies with m anufacturing production activ­
ities) to describe the activities and procedures involved in producing and sustaining
supplier-client relations and delivering “ service” — rough ly what we call “ interac­
tivity” above. Belleflam e et ah (1986) classified a sam ple o f innovations from service
companies in term s o f w hether these centered on servuction, production, or a
combination o f both (w ithout assum ing that these w ould necessarily be IT-centered
innovations),11 and fo u n d n u m erou s exam ples o f all types o f innovation. Tordoir
(1996) studied profession al services, and noted very different patterns o f interaction
between suppliers and clients, in trodu cin g a distinction between “ jo b b in g” relations
(where the service su p plier provides a relatively standardized service), and “ spar­
ring” ones (where the su p plier and client negotiate the details o f what service is to be
provided, and h ow ). M o re recent studies have addressed the question o f how service
innovation differs according to the degree to w hich services are standardized or
specialized to specific clients (cf. H ip p et al. 2000, 2003). The m ore interactive
services w ould seem to require greater exchange o f know ledge between service
firm and client, and the learning processes involved can be a fertile basis for
innovation: how ever, m an y o f the innovations are liable to be o n e -o ff solutions for
particular problem s, o r elaborate custom izations o f m ore generic solutions.
The R P C approach argues that IT foased innovation begins w ith back-office
processes and then m oves fo rw ard into functions involvin g m ore custom er contact.
4 42 IAN MILES

T hese latter fu n ctio n s m a y involve n ew service p ro d u cts (su ch as n ew types

o f b an k accounts m o re closely tailo red to the circu m stan ces o f in d iv id u a l clients),
o r d elivery in n o vatio n s (e.g. onlin e b an k in g and cash m ach in es). T h ere is also
scope for in n o va tio n in oth er cu sto m er-facin g fu n ctio n s, fo r exam p le in targeted
m arketin g and in com pu ter-assisted helplin es, etc. In teractio n around
service relations certain ly does involve in fo rm a tio n flo w s— esp ecially in the case
o f in fo rm atio n al services like cou nselin g, con su ltan cy, an d edu cation . But
in teractio n m ay involve m any physical elem ents, not all of w hich are
p rim a rily carriers o f in fo rm atio n al con ten t (co n sid er services p e rfo rm in g physical
and b io lo gical fu n ctio n s like catering, tran sp o rt, clean in g, su rgery, an d hairdress­
ing).
IT -b ased in n o vatio n m ay ap p ly to the in fo rm a tio n a l elem en ts o f such service
p ro v isio n — m arketin g, o rd erin g, tran saction s, etc. IT m a y be an im p o rta n t adjunct
to oth er types o f services in n o vatio n — w ith o u t teleph ones, pizza d elivery services
w o u ld be m u ch less attractive! B u t it is n o t the o n ly vector. O th er typ es o f interaction
are also involved in services, in clu d in g physical and o th er elem ents, an d these may
benefit fro m n o n -IT -b ased in n o vatio n — m o torcycles an d in su lated boxes in the case
o f pizza delivery. T h e R P C m odel, w h ich em phasizes the catalytic effect on services
in n o vatio n o f the in tro d u ctio n o f IT equ ip m en t and so ftw are fro m m anu facturing
sectors, could be extended to ap p ly to these tech n ologies in the service sector as well.
E xam p les include superm arkets investing in the d evelo p m en t o f refrigeratio n tech­
n o lo g y (as w ell as general refrigeration design in pu ts, su p erm ark ets played im po rt­
ant roles in d eterm in in g w hat sort o f shift w as m ad e to altern atives to C F C s in
eq u ip m en t), railw ay com pan ies su p p o rtin g the d evelo p m en t o f better trains (before
p rivatization , B ritish R ail h ad a substantial R & D facility c o n d u ctin g w o rk on topics
such as faster engines, m ore en viro n m en tally so u n d carriage p a in tin g practices, and
safety issues), and so on.
A n o th er ap p ro ach to exam in in g the w ay in w h ich services d iffe r co n sid ers the key
tran sfo rm atio n s that they effect, and distinguish es a m o n g three b ro ad groups o f
services (M iles 1993). T h is classification h igh lights the a d o p tio n an d innovation
d yn am ics o f services w ith respect to a n u m b er o f m a jo r tech n o lo gies in addition
to IT
A first set o f services, Physical Services (tran sport, d o m estic services, catering etc.)
involve physical tran sfo rm ation s. These have been p a rtic u la rly suitable for the
ap p licatio n o f au tom o tive and electric po w er tech n o logies— e.g. in freigh t haulage,
lau nd ries, co o kin g equ ipm en t, and so on. T h e a d o p tio n o f n ew tech n ologies in this
sector has freq uen tly created co m p etition am o n g d ifferen t m o d es o f provision;
betw een road, rail, and air services; laundries and launderettes; tra d itio n al restaur­
ants and fast food outlets; etc, But the co m p etitio n betw een m o d es o f p ro visio n can
go m uch deeper, In particular, con su m er services such as la u n d ry o r fo o d prepar­
ation have been subject to co m p etition from w hat G e rsh u n y (1978) described as
self-services, p ro d u ction by consum ers in the h ou seh o ld . Su ch tech n ologies as the
automobile, hom e refrigeration and other household appliances have enabled
consumers to use m an u factu red goods to provide services in the household in
competition with service firm s. The adoption by consum ers o f these m anufactured
goods also has given rise to som e new services, such as autom otive repair establish­
ments.
New IT is n o w w id ely used in physical services, such as transport, logistics,
retailing, and w arehousing. C o m p u ters have long been applied to their back-office
accounting and tran sactional fu nctions, and electronic cash registers and scanners12
introduced into superm arkets and sm aller shops. These IT-based devices are linked
to the office system s being used for stocktaking, and in som e cases data are fed to
automated w arehouses and m uch w ider system s o f superm arket autom ation. N ew
“transport in form atics” system s do m ore than just docum ent timetables and the
locations o f vehicles, p ro vid in g advanced routing and tariffing procedures, backed
up by mobile co m m u n icatio n s, “ sm art cards” and other innovations.
The large-scale Human Services subsector has often been organized under the
auspices o f the w elfare state, and m an y o f the front-office tasks in this subsector have
had little scope for ap p lication o f generic I T M an y o f the Social Welfare activities o f
this subsector have utilized office and com m unications systems, while the medical
field has exploited successive generations o f m edical technology— surgical equip­
ment, pharm aceuticals, radiology, etc. Typically both sets o f services have had to
combine large-scale adm in istrative data processing applications (payroll, pensions,
passports, d rivin g licenses, and the like) and planning (for exam ple, m anaging
housing system s and waste disposal services, m onitorin g epidem iological and
environmental statistics) fu nctions, w ith tasks that are m uch m ore custom ized in
terms o f the characteristics o f specific citizens o r clients. The large-scale tasks were
early pioneers o f co m p u ter use. N ow , PC s and data networks are being w idely
adopted, allow ing d ecision su p p o rt in the context o f the details o f individual
clients— for exam ple, expert system s to aid m edical diagnosis and prescription, or
to speed up the task o f assessing in d ivid u al entitlem ent to welfare benefits. There are
some self-service-type applications, too, where the clients use public access term in­
als or hom e-based equ ip m en t to gather inform ation on service provision, or access
the services directly (e.g. interactive teaching aids, databases on available jobs or
benefits, etc.) Som e integration w ith in and across public services is possible as
inform ation on the sam e client held in different databases can be com bined— but
privacy rules have often im peded such integration, and there are m any other
instances w here perceived threats to privacy and civil liberties have restricted
innovations in volvin g capture and distribution o f inform ation.
Finally, Information Services are relatively less dependent on m otor pow er tech­
nologies, although electricity is im po rtan t to them . They have m ade considerable
use o f inform ation technologies, including those that predate new IT and Its
foundations in m icroelectronics. Som e o f these services, such as broadcasting and
cinema, are fo u n d ed on these technologies, while others, like consultancies and
444 IAN MILES

technical services, have used them as im p o rta n t tools fo r p ro d u c in g an d delivering

th eir o u tp u ts. In som e o f the co n su m er services here, n ew an d even o ld e r generations
o f IT d isp lay self-service trends sim ilar to those that m o to rs in d u ced in physical
services. T h u s, in the field o f entertainm ent, tra d itio n al c o n su m e r services like
theater an d cin em a have been subject to co m p e titio n fro m n e w ones such as T V
and oth er au d io visu al equ ip m en t and, m o re recently, v id eo g a m es, P C s, and online
entertainm ent. P rofession al and business services are a m o n g the m o st IT -intensive
sectors o f industry, as are finan cial services, w h ich are also la rg e ly co n cern ed with
in fo rm a tio n p rocessin g (in this case p rocessin g and m a n ip u la tin g d ata ab ou t p ro p ­
erty relationships).
N ew IT is thus e n o rm o u sly im p o rtan t— and often v e ry v isib le — in in fo rm atio n
services. Such in n o vation s as au tom ated teller m ach in es an d sm a rt cards, new
teleph one and telem atics services, an d shifts fro m an alo g b ro a d c a stin g to interactive
d igital m edia and “ n arro w castin g” are all the focu s o f co n sid erab le investm ent and
activity. M o st n ation al b road castin g services in the W est n o w o ffer o p p o rtu n ities to
access archived rad io an d som etim es T V services th ro u g h the In tern et, and often
p ro vid e a g o o d deal o f ad d ition al content, d iscu ssion fo ra, an d the lik e .13 Som e o f
the service firm s involved p lay a p ro m in en t role in g u id in g the in n o va tio n s, as in the
case o f b an ks w h o are h eavily involved in d efin in g the ch aracteristics o f new teller
m achines. T h e pace o f in n o vatio n in these in d u stries is q u ick en ed b y the shift in
regu latory p o licy in m an y countries, w h ich has led to n ew en tran ts an d increased
in tern ation al co m p etition co n fro n tin g m an y firm s an d sectors (Sau ve and M attoo
2003).
T h is sort o f analysis pro vid es a useful m a p p in g o f d ifferen t typ es o f services
and in n o vatio n trajectories, b u t reveals little ab ou t the d y n a m ics and processes o f
in n o vatio n in the vario u s services that have been d istin gu ish ed . T h e R P C m odel is
not the last w o rd on this subject. O ne po ten tially fru itfu l aven u e fo r fu rth er research
applies evo lu tio n ary and characteristics-based ap p ro ach es to services innovation
(see G allo u j 2002).

16.3 I n n o v a t i o n S y s t e m s a n d t h e
O r g a n i z a t i o n of I n n o v a t i o n

IT pro vid es a generic tech n ology that can be w id ely ap p lied across service sectors— a
substantial break w ith past experience, w here even tech n o logies like autom obiles
and typew riters, alm ost u n iversally adopted in certain services b ran ch es, w ere rarely
 I N N O V A T I O N IN S E R V I C E S 445

applied in others.14 T h o u gh services vary in the extent and speed with which they are
adopting these innovations, and som e o f them are m ore relevant to certain branches
of services than to others, m any technological opportunities are opening up for all
services. Even personal services such as hairdressing are subject to innovations—
there are shops w here video tech nology and PC s are used to give clients an im pres­
sion o f how they w ou ld look w ith different hairstyles, for instance. And IT is a very
configurational technology, w hich challenges users to develop new software and
interfaces, and to b rin g diverse elem ents o f technologies together in new ways.
Their increased tech n o logy-in ten sity has not yet transform ed all o f the features o f
innovation in services. M an y services are argu ably laboring under a heritage derived
from past periods w here few generic technologies found ready application in their
activities. The low tech n ology-in ten sity that previously characterized services meant
that many services firm s paid little attention to strategies for the m anagem ent or
adoption o f innovation. Lim ited use o f advanced technology w ould also mean that
most service firm s w o u ld have little incentive to be linked into innovation systems
that would connect them w ith those responsible for generating the new technolo­
gies. And in the w id er in n o vation system s them selves, few o f the innovation-related
facilities offered b y institution s such as university departm ents, research institutes,
and government laboratories are tailored to the requirem ents o f services. It is thus
not surprising that few service firm s m ake m uch use o f these resources currently
(Institute o f In n ovation Research 2003).
There are, as alw ays w hen discussing services, exceptions, For exam ple rail,
broadcasting, and telecom m u nications services, m any o f which were state-owned
and very large organizations for m uch o f their existence, were closely related to
manufacturers. These large services firm s frequently ran their own laboratories,
testing sites, technical train in g program s, etc. O ther exceptions include very large-
scale service firm s in sectors like financial and retailing services. Especially in the
latter case, superm arket chains have becom e adept m anagers o f their ow n supply
chains, and are often active in dictating production processes and products—
including innovations— in their agricultural and m anufacturing suppliers. For
example, their suppliers m ay be requested to adopt specific environm ental or anim al
husbandry practices, to use ecom m erce techniques, etc. As well as influencing their
suppliers, such firm s have been pioneers in som e sorts o f technology for their own
use— for exam ple, data w areh ou sin g and data-m ining m ethods.15 Another, and
particularly im portant, class o f exceptions, are the business services that are, as we
shall see, im portant actors in in n o vation system s, contributing to innovation across
the economy. C onsultancies, training organizations, and m any firm s helping to
service new technologies are ju st a few o f the agents involved here- not to m ention
specialized R & D and design services themselves!
But relatively few service firm s and sectors have strong links w ith national or
regional innovation system s. T h e organization o f innovation is a new developm ent
in m any service firm s— for instance, in a set o f recent interview s the author found
446 IAN MILES

that m an y large service firm s in the U K w ere only ju st b eg in n in g to exam in e the

application o f IPR arrangements to their innovations, as opposed to their trade­
m arks and co p yrigh t content.16 In part, this new orientation reflects a general
em ph asis o n intangible assets, an d in part it is in resp on se to the em ergence o f
“ business m eth o d s” patentin g in the U S A , w h ich are seen as im p a c tin g u p o n m any
service o peration s.
O ne result o f this heritage is that m a n y service firm s are u n ab le to a p p ly relevant
know ledge, and, m o re particularly, are u nable to learn effectively. T h e y can rapidly
ad o p t “ o ff the s h e lf” technologies like P C s, b u t fin d it h ard er to d evelop m ore
cu stom ized o r in n o vative solu tions. In part this is becau se services in n o vatio n is
rarely organ ized in term s o f the “ stan d ard ” m o d els o f R & D m an ag em en t structures,
and is typ ically con du cted on a m o re ad -h o c, p ro ject m an ag em en t basis. Service
firm s tend to stress h u m an resources and tech n o lo g y a cq u isitio n rath er than form al
R & D (Tether et al. 2001; Su n d b o i99B)> resulting in lim ited c o o rd in a tio n o f learning
experiences— the in n o vation s that are m ade are often n o t rep ro d u ced in subsequent
projects, and flow s o f know ledge w ith in and betw een services firm s as to techno­
logical o p p o rtu n ities, g o o d practice in in n o vatio n , an d th e like, m a y be lim ited .17
C IS data discussed b elo w sh ow that m an y services sectors d isp la y so m ew h at lower
levels o f in n o vatio n than their parallels in m an u factu rin g . T h is m a y resu lt fro m weak
integration into in n o vatio n system s, the o rien tation o f existin g tech n o lo g y support
in stitu tion s tow ard m an u factu rin g activities, and w eak in tern al o rg an izatio n for
in n o vatio n in m an y services— even i f the firm s them selves d o n o t rep o rt exceptional
p ro b lem s in in n o vative efforts.
Such features o f services in n o vatio n p artly reflect the h isto rical legacy discussed
above. But they also p artly reflect the fact that the featu res o f d ifferen t types o f
econ o m ic activity can shape in n o vation processes an d trajecto ries in distinctive
w ays. A rguably, the nature o f service in n o vation s d em an d s d ifferen t app roach es to
m an agem en t and organ izatio n o f the in n o va tio n process— fo r exam p le, requiring
m o re em phasis on the service w orkers (especially i f p ro fessio n als) an d clients, and
on the interaction process betw een them . T h e trad itio n al R & D lab m ay not be well
suited to such in n ovation s, th ou gh we find large tech n o lo g y -in ten sive services firms
have been ru n n in g such laboratories (in som e cases, such as telecom m u n ication s
an d railw ays, fo r m any decades).
T h e differences between services and m a n u fa ctu rin g in n o v a tio n m ay reflect
h istorical legacies stem m ing fro m the low tech n o lo g y -in ten sity o f m an y services
activities (w hich m ay be overcom e in tim e), alon gside o f the sign ifican ce o f charac­
teristics such as in teractivity to m any services (w hich m ay d em an d sp ecific processes
and practices). But these generalizations need to be q u alified to take in to accoun t the
en o rm o u s h eterogeneity w ith in services— they do not a p p ly to large segm ents o f
the services sector. D evelop in g a m ore nuanced fra m e w o rk that is sensitive to the
en o rm o u s h eterogeneity am o n g the vario u s co m p o n en ts o f the services sector is an
im p o rtan t task for future research.
 I N N O V A T I O N IN S E R V I C E S 4 47

16.3л Innovation Surveys and Services

Despite substantial efforts to im prove the situation in recent years, data on m any
aspects o f services and services inno vation are less detailed and less com prehensi ve
than those for m anu factu ring. Even the C o m m u n ity Innovation Survey excludes
public and personal services. M an y features o f services innovation are quantified in
such innovation surveys and sim ilar data, but qualitative features are harder to
detect. Few available in n o vation indicators were designed with services in m ind, and
thus may fail to adequately capture the dynam ics o f services innovation.
With these caveats in m ind, as well as the discussion above on the differences
between innovation in services and m anufacturing, it is notew orthy that the evi­
dence from these surveys indicates that services innovation does not seem to follow
dramatically different paths from those displayed in m anufacturing. The differences
appear to be m ore o f degree than o f kind.
Tether et al. (2001) present the m ost extensive analysis to date o f services in n ov­
ation, based on the C IS2 (second C o m m u n ity Innovation Survey) data.18 Just under
half o f the service enterprises reported undertaking innovative activities between
1994 and 1996. This share is slightly sm aller than that reported for m anufacturers,
and this difference betw een m an u factu rin g and service firm s' innovation propen­
sities remains even when size is controlled for. M ost service branches include a
higher share o f sm aller businesses than do m anufacturing branches (though finan­
cial services are d om in ated b y very large firm s). A n d larger enterprises are m ore
likely to engage in in n ovative activities in (m ost) service branches, as elsewhere. But
this is not enough to account for the difference between services and m anufacturing
firms in reported levels o f innovation. (Looking at the size distributions, we find
36 per cent o f sm all services and 48 per cent o f sm all m anufacturers; 48 per cent and
55 per cent o f m ediu m -sized services and m anufacturers, respectively; and 73 per
cent and 79 per cent o f large services and m anufacturers, respectively, are classified as
innovative enterprises.)
The proportion o f in n ovators is high am ongst the technology-oriented services,
many o f w hom are co m parab le in term s o f reported innovation levels to high-tech
manufacturing— 68 per cent o f com pu ter services, 64 per cent o f telecom m unica­
tion, and 55 per cent o f technical services are classified as “ innovative enterprises in
CIS2. M ore trad ition al services appear to be particularly low innovators (only 24 per
cent o f transport services, for exa m p le )19.
Such results should be interpreted cautiously, since as this chapter has pointed
out, the m easures o f in n o vation em ployed in the C IS are probably less than ideal for
studying services. Services firm s are less likely to see what they are doing as techno­
logical innovation, rather than custom ization or o n e -o ff service production.
Organizational in n o vatio n m ay be im p o rtan t in services, but is not exam ined in
the survey. The C IS 2 also fails to specify the sorts o f technology involved in
innovations, m eaning that the role o f IT in services innovation cannot be traced.
44$ IAN M ILES

B ut the role o f IT in services in n o vatio n is d ifficu lt to o verstate Lich t and Moch

(1999) found that all innovating services firms in Germany undertook IT innov­
ations, even if they also applied other technologies. Unfortunately, the survey they
worked w ith only covered services, so we cannot say whether this conclusion would
also ap p ly to o th er sectors.
C o n tra ry to som e expectations, ju st u n d er h a lf the in n o v a tin g service enterprises
covered b y the su rvey analyzed in Lich t an d M o ch rep o rted that th ey h ad engaged in
R & D betw een 1994 and 1996. Indeed, a q u arter o f these firm s rep o rted having
engaged in R & D on a co n tin u o u s basis. H ow ever, R & D is less c o m m o n in innovating
services firm s than am on gst m an u factu rers o f sim ilar ch aracteristics, again control­
lin g fo r size.20 R & D is m o re co m m o n in large service en terp rises, an d in technology-
orien ted services.
O f course, R & D is n o t strictly sp eakin g the m o st c o m m o n o r m o st im portant
source o f in n o vatio n fo r m an y services. A cq u isitio n o f m a c h in e ry an d equipm ent,
acq u isition o f oth er external technologies (in clu d in g so ftw are) an d train in g directly
linked to in n o vatio n w ere the m ost w id ely u n d ertak en in n o va tio n -rela ted activities
in services, accord in g to C IS2 data (Tether et al. 20 0 1). T h ese so u rces illustrate the
im p o rtan ce o f tech n o logy ad o p tio n in services in n o va tio n . T h e im p o rtan ce o f the
h u m an elem ent for services is also u nd erlined b y the tra in in g exp en d itu res o f these
firm s. O n average, acquired technologies accou n ted fo r the largest share o f expend­
itures o n in n o vatio n ; w hile in -h o u se R & D accou n ted fo r a n o th er q u arter o f total
exp en ditures on in n o vation , w ith this share h ig h er a m o n g st tech nology-orien ted
services. T ech n ology-orien ted services tend to spend m o st o n in n o va tio n , but all
sectors surveyed contain som e very h igh sp en d in g (an d so m e v ery lo w spending)
enterprises.
A n o th er set o f C IS questions concerned sources o f in fo rm a tio n fo r innovation.
Tether an d Sw ann (2003) contrast services an d m a n u fa c tu rin g firm s' use o f infor­
m atio n sources, using C IS3 data for the U K . T h e b ro ad p attern o f results is fairly
sim ilar across the sectors, w ith rather m ore m a n u fa ctu rin g firm s ty p ica lly reporting
use o f each in fo rm atio n source. Sources w ith in the en terp rises w ere the most
co m m o n ly used, being cited by 85 per cent o f m an u fa ctu re rs an d 81 per cent o f
services. A substantial p ro p o rtio n o f n o n -R & D p e rfo rm in g in n o va to rs reported
u sing such sources. Su pp liers are also w id ely used b y both sectors (83 p er cent and
77 P er cent o f m anu factu rers and services firm s resp ectively cited these sources),
follow ed by custom ers (80 per cent and 73 per cent). R ath er less im p o rta n t sources o f
in fo rm atio n are the technical press (65 p er cent fo r b oth sectors), com petitors
(66 per cent and 62 per cent), and trade fairs (72 p er cent an d 58 p e r cent). Two
in fo rm atio n sources w here services firm s rep ort m ak in g m o re use than m anufac­
turers are m eetings and conferences (m an u factu rin g 52 p er cent, services 62 per
cent), and consultants (48 and 56 per cent). O th er sou rces, such as standards,
govern m en t offices, other parts o f the enterprise, and u n iversities (36 and 24 per
c e n t) are m ore often cited b y m anufacturers,
 The sectoral differences are less acute than m ight have been expected (and
doser inspection o f variation s across different services indicates considerable
diversity here). But the greater use of consultants and lower use o f sources such
as Universities tends to confirm the notion that m any services are poorly linked
into wider innovation system s, and the form al institutions that support them .21
Many specialized services firm s are carrying out functions that are also under­
taken within firm s in other sectors o f the econom y. Office w ork, transport,
commercial transactions, security, and catering, and sim ilar activities— ser­
vices— are undertaken in all sectors, though to differing degrees. Just as in n ov­
ation surveys were not designed w ith services in m ind at the outset, so there m ay
be some overlooking o f service fu nctions, which m ay well slip between the
“product innovation/process in n o v a tio n ' categories o f C IS-type instrum ents,
and may even be m atters that the respondents to the surveys are unaware of. A
number o f recent studies consider services supplied in support o f the core
products o f m an u factu rin g firm s (e.g. K uusisto 2000; Lay 2002; M athe and
Shapiro 1993); com m en tators such as D avies (2003) and Howells (2001) have
argued that the service co m po n en t o f such products is grow ing dram atically in
many sectors. The issues that con fron t adoption and innovation o f service-related
technologies in services firm s m ay well be liable to affect innovation in service
functions in m an u factu rin g firm s. Service functions can in principle be significant
loci o f organizational learning and innovation in all sectors. It rem ains to be
established whether the o rgan ization o f innovation in such service functions m ore
resembles conventional m an u factu rin g innovation, or m ore inform al systems
used for services innovation.

16,3.2 Innovation-Supporting Services in Innovation Systems

Business services have risen d ram atically in econom ic significance— for instance,
from little over 3 per cent o f U S valu e-ad ded and em ploym ent in 1970 to 9.9 and 13.8
per cent respectively in 2000, w ith equally striking increases in other countries
(E C O R Y S-N E I2003). T h ey have also becom e highly evident contributors to in n ov­
ation across the econom y.22 T h ey pro vid e interm ediate inputs to industry and other
organizations— they are business services as m uch because they are servicing b u si­
ness processes as because their clients are often private firm s themselves. Know ledge-
intensive business services (K IB S ), in particular, play im portant roles in innovation
systems (Leiponen 2001; M iles 1999b; and several chapters o f G adrey and Gallouj
2002). Some are tran snational firm s, dealing w ith know ledge from the frontier o f
practice. But m any are sm aller; locally based K IB S m ay be im portant agents o f
transfer o f locally-specific know ledge, em bedded in local networks, between actors
in regional innovation systems (cf. Kautonen 2001). Several studies suggest that the
450 IAN M ILES

presence and use of KIBS enhances the performance of economic sectors and
regio n s.23
K IB S in clu d e all kinds o f business services that are fo u n d ed u p o n h ig h ly special­
ized kn o w led ge— social and in stitu tion al k n o w led ge in m a n y o f the traditional
p ro fessio n al services* o r m o re tech n o logical an d tech n ical k n o w led ge. These
firm s typ ically have v ery h igh levels o f q u alified staff* su ch as u n iv e rsity graduates.24
S o m e K IB S are based on adm inistrative* legal* m arketing* o r sim ila r knowledge.
O thers are directly based on scientific an d tech n o lo gical k n o w led g e— testing* proto­
typing* en viron m en tal services* engineerin g con su ltan cy, etc. T h e knowledge
requirem en ts fo r tech n o lo gy users are b o u n d to be m o re ch allen g in g w here new
tech n o lo gy is involved. Firm s are less lik ely to have acq u ired the k n o w led ge neces­
sa ry to und erstand , m aster, and utilize n ew p ro d u ct an d p ro cess o p p o rtu n ities.25
T h u s, m an y technical K IB S focus on new tech n o logical o p p o rtu n itie s— exam ples
include W eb and Internet* softw are and co m p u ter services; o th ers on the production
and transfer o f know ledge ab ou t new tech n o logy; in fo rm a tio n an d tra in in g services*
fo r exam ple.
T ech nology-related K IB S sectors are a m o n g the m o st active in n o vato rs in the
econom y* as indicated b y C IS an d other data (e.g. D T I 2003, w h ich includes data
in d icatin g the h igh levels o f R8cD expenditure* patenting* etc. o f softw are and IT
services).26 We saw above that Tether and Sw ann (2003)* u sin g C IS 3 d ata fo r the UK*
fo u n d that m o st services are p o o rly linked w ith the p u b lic elem ents o f innovation
system s (e.g. in term s o f so u rcin g in fo rm a tio n fro m an d c o llab o ra tin g w ith univer­
sities). B u t they also rep ort that som e technical K IB S w ere o u tsta n d in g ly w ell linked
to these co m p o n en ts o f in n o vatio n system s. S o m e tech n ical services (e.g. contract
R & D ) had u n iq u ely h igh levels o f in teraction w ith the p u b lic scien ce base* higher
than any m an u factu rin g sector; however* IT services an d m o re p ro fe ssio n al services
tended to have lo w levels o f contact* relying m o re o n p ro fe ssio n al association s and
the like to refresh their know ledge.
M a n y K IB S p lay im p o rtan t roles in in n o vatio n processes in their client firm s and
sectors. These roles m ay not be altogether n ew — som e h ave existed fo r m any
decades. B u t there has been considerable grow th o f e m p lo y m e n t an d output in
K IB S sectors* w h ich can o n ly m ean that their use has also expanded* an d suggests
that their significance for in n o vation across the e c o n o m y w ill also have m ush­
room ed,
W hat are these roles? K IB S m ay p ro vid e the firm w ith general in fo rm a tio n about
its internal o p eratio n s and external environ m en ts. T ech n o lo g y and innovation-
related in fo rm atio n often form s p art o f this. K IB S m ay sim p ly p lay a role in
id en tifyin g the nature o f a p articu lar p ro b lem o r class o f p ro b lem s co n fro n tin g the
firm (fo r instance* that com petitors are lau n ch in g p ro d u cts w ith new fu n ction al­
ities* that regulations m ay m ean that processes w ill have to generate less o f a
p articu lar pollutant* etc.). K IB S m ay p ro p o se w ays o f so lv in g a technological
problem (fo r exam p le 1 ecom m en d in g that p articu lar strategies are undertaken for
 I N N O V A T I O N IN S ER V IC E S 451

product or process innovation). They may provide advice (for example, recom­
mending a specific technological solution), or actually implement such a solution
on a “turnkey” or long-term basis (as in the case of systems integrators and facilities
managers). The KIBS relate technological knowledge to the specific problems
encountered by the client. German survey data indicate that technology-related
KIBS are more likely to produce specialized products that are tailored to client needs
than were the other services studied (Hipp et al. 2003).
In addition to expertise in specific industrial, technological, or functional
domains, KIBS professionals require skills in interpersonal communication, presen-
tation o f materials, “ im pression m anagem ent ” and the like. These are fairly rare
capabilities, and their co m b in ation is rarer still; for this reason labor costs and wages
are high in m ost K IB S. Toivonen (2001) finds that in Finland effectively all o f the
KIBS studied require co m b in ation s o f generic and sector-specific skills. C om m on
requirements were m arketing and sales skills; social skills such as sensitivity to
others, willingness to share know ledge and m otivate others, capacity for self­
renewal, IT-related skills, and sector-specific know ledge related to o n e s particular
expertise, to the processes and business m echanism s characteristic o f the K IB S, and
knowledge about the industries and organizations o f clients.

16.4 C o n c l u s i o n s

The growth o f services to their d om in an t position in industrial econom ies means

that we can no longer ignore services innovation, o r sim ply assum e that it follows the
patterns and processes depicted in m anu factu ring production processes. U nder­
standing “ service in n o vatio n ” m ay w id en o u r approaches to explaining, m easuring,
and managing inn ovation . In n o vatio n studies w ill have to take on board the issues o f
organizational and m arket in n o vation , interorganizational and client-facing in n ov­
ation, and even aesthetic and cultural innovation. M any o f the m ostly w idely
remarked features o f the evo lu tio n o f technological innovation in the late twentieth
century depended centrally on com bin ed m aterial and non-m aterial innovation.
This suggests that o u r m odels o f in n ovation should put less em phasis on artifacts
and technological in n o vation , and m ore on seeing innovation as involving changes
in market relationships that can be effected at least partly through artifact and
service innovations, w ith organizational and technological dim ensions.
Service innovation n o w occupies a m ore prom inen t position in innovation
studies over the last few years, b u t this has yet to be reflected in an accum ulation
o f knowledge about services in in n ovation policy. There are few obviou s efforts to
452 IAN MILES

give services a prominent role in such policies, and their specific requirements may
be overlooked in current programs. (There is a dearth of research concerning how
innovation policies affect services, let alone analysis of service- related policies.) One
exception is Finland, where KIBS are being treated as important actors in the
innovation process. We have seen that KIBS are significant elements of innovation
systems, and as policy adopts a more systemic viewpoint, we can expect their role to
b ecom e m o re o f a focus in policy.
T h e greater p o licy salience o f services in n o va tio n is, in tu rn , lik ely to stim ulate
m o re detailed research into the area. It w ill be im p o rta n t to m o ve b ey o n d d ocu ­
m en tin g the grow th o f K IB S , to exam in e precisely h o w th ey fu n c tio n as know ledge
sources and interm ediaries: w h at sorts o f exch ange o f k n o w led g e take place in what
w ays an d at w hat steps o f the in n o vatio n process; h o w these are m an ag ed o n service
and client sides o f the equ ation ; w h at skills and cap ab ilities are req u ire d fo r effective
and in n o vative so lu tio n s to be im plem ented . T h ere are still v e ry fe w studies that
address these fu n d am en tal q u estio n s.27

N otes

1. This view has been widely expressed, for instance, with services being classified as
supplier-driven in Pavitt’s original taxonomy o f types o f innovation (1984)— though in
a subsequent paper (1994) he put software services into the “ specialized supplier” group,
and added a category o f “ information intensive” firms which included finance, retailing,
travel and publishing.
2. Hipp et al. (2000) show that a surprisingly large proportion o f German service firms
consider their outputs to be largely standardized. Least standardization was reported by
business services such as technical and computer services.
3. For just one example, consider Jones (1995) on in-flight catering services.
4. By 2000, about 1 per cent o f the UK workforce was employed at telephone call centers.
5. For a series of studies on service internationalization and innovation— still a largely
unexplored topic— see Miozzo and Miles (2003).
6. CL Miles and Matthews (1992), Roach (1988) for early analyses. The data, derived from
input—output tables, allow us to examine investment patterns across different industries.
IT investment from services sectors constitutes a greater share of this investment than
services output constitutes of all output. And for the UK in the 1980s, Miles and Matthews
found that the proportion of sectoral investment devoted to computers and telecommu­
nications equipment was 5.7 and 4.6 per cent respectively for services, 4.1 and 4.0 per cent
for manufacturing. (The expenditure on telecommunications equipment among services
at this time was heavily dominated by telecommunication services firms; and in general
the high levels of 11 investment are driven by specific sectors, such as finance.)
7. Not that the problem is not reported by manufacturers— 1 found U K public sector
organizations in the 1980s to be vociferously complaining that their projects were plagued
by departui e o f IT staff to financial services. Beyond the affluent West similar concerns are
 I N N O V A T I O N IN S E R V I C E S
453

also raised, for instance Computerworld Singapore (7(3); 20) documented in October
2000 that financial services computer professionals there were consistently paid more
than comparable staff in other sectors.
8. This model, while proving extremely influential and a useful starting point for analysis, is
widely viewed as o f limited applicability even to manufacturing.
9. US researchers have also suggested such developments: see e.g. Faulhaber, Noam, and
Tasley (1986). But see Bronwyn H alls Ch. 17 in this volume, which indicates that
“followers” may spend a long time in that state before becoming innovative in their
own right.
10. Interbranch on-line services; autom ated teller m achines; credit cards and associated
services; remote banking, and electronic funds transfer at point o f sale.
11. Gallouj and Weinstein (1997) provide a useful review, comparing “ servuction” to a
number of other formulations. Though there have been few “ servuction” studies focus­
ing on Innovation, the conceptual approach continues to be developed, e.g. by Gadrey
and de Bandt (1994).
12. Note that bar-code scanners require the cooperation of manufacturers in bar-coding
their products. Similarly, financial service innovations like credit and debit cards require
the cooperation o f retailers in accepting these cards and using validation systems.
13. The BBC website at http://www.bbc.co.uk is a case in point. New BBC radio and TV
channels are only available digitally, and remarkable volumes of archived content and
text are also online.
14. Probably the only universal innovations were ones involving the construction and
maintenance o f buildings and technologies o f heating and lighting— and telephones.
Such innovations typically involved relatively little user learning. Even where new skills
were required (e.g. automobile driving skills— which in any case were typically acquired
for everyday purposes) these technologies required relatively little configuration to meet
the needs o f specific users. Indeed, limited scope for configuration was presented, and
the relevant engineering services (e.g. garages) were a matter o f high-street crafts rather
than industrial laboratories.
15. NIST in the USA (TASC 1998) examined how technology-intensive services in several
sectors deal with technology barriers. Some barriers were technology-specific— thus the
high risks o f complex technology led to needs for technical expertise not available to
most individual firms. Collaborative R8cD was undertaken with other firms in their
industry (and often with manufacturing sector partners, too), typically to gain access to
complementary research or technical skills. IT development barriers were overcome by
codevelopment projects with manufacturing suppliers. Such collaborations were not in
general oriented to knowledge resulting from basic research. IT implementation barriers
were associated with high costs o f configuring and employing systems (often IT imple-
mentation cost four to five times more than acquiring the hardware and software). Other
barriers were “market-related” — thus high transaction costs associated with the sys-
temic nature o f new IT, which led to emphasis on standards-related activities to reduce
barriers to IT development and implementation. Standards and protocols were often
central to innovation strategies.
16. See the report by FhG-ISI (2003), which deals especially with services patenting activ­
ities. The general argument is that the patent mechanism is not very appropriate to many
service innovations, and thus the firms are not particularly oriented to these types of
454 ia n m il e s

intellectual property. Nevertheless, the study found numerous service firms active in
patenting activity and the development of new IP strategies.
17* Again, there are striking exceptions to this rule some service firms in areas like IT and
consultancy are pioneers in innovation and knowledge management techniques.
18. The sectors studied excluded public and consumer services such as retail and HORECA;
and microbusinesses and very small firms were also excluded.
19. Tether and Swann (2003) present detailed information on UK firms in CIS3 data, where
different types of manufacturing and service are contrasted in some detail.
20. Nearly 70 per cent o f the innovating manufacturers conducted R 5cD.
21. Other evidence supporting this view is presented in Miles 1999a; Tether and Swann 2003.
22. One— but only one— o f the sources o f growth in business services, in particular, is the
outsourcing o f activities previously undertaken in-house by firms and organizations in
other sectors.
23. Some authors even describe them as forming a “ second knowledge infrastructure.” The
traditional primary knowledge infrastructure is mainly a matter o f Higher Education
Institutions (HEIs) and government laboratories/public research and technology
organizations (RTOs).
24. It is difficult to attribute causality to the correlational data involved here, but some very
different types o f study yield broadly similar results. Researchers like Antonelli (2000)
and Tomlinson (2000)— using slightly different input-output datasets, and statistical
methods— show an association between the use o f KIBS as intermediate inputs, and the
performance o f the user sectors. Doubt is cast on the methodology by ECORYS-NEI
(2003), however. Peneder (in European Commission 2000: ch. 4) found that clusters of
industries characterized by high KIBS use performed particularly well. Hansen (1994)
reported that the growth performance o f the economies o f US cities was related to the
size o f the KIBS sectors in these economies. An interesting study by Muller (2001)
examines relations between KIBS and SMEs at regional level, suggesting benefits in
terms o f innovation on both sides.
25. Tether and Swann (2003) show that in the UK, at least, these are the sectors with the
highest graduate-intensity. There are clear differences between those KIBS with high
levels o f science and engineering graduates (e.g. Technical and IT services) and those with
other classes o f University graduate (e.g, consultancy and marketing services).
26. CIS2 data show that acquisition o f external technology through use o f consultancy
services is the second most frequent mechanism used by manufacturing firms. The
most important mechanism is direct equipment purchases.
27. Tomlinson (1999) analyzed U K survey data, finding that KIBS staff are more likely than
are others to learn new things, to receive training, to work with computers, and to move
between different types o f work. Labor mobility is often emphasized as a means for
diffusing knowledge around the economy. Tomlinson argued that KIBS provide an
alternative perhaps a superior— means. (The survey suggests that people moving
between jobs fare poorly on these indicators of “ life long learning.” Whether this result
is peculiar to the UK and/or to the recession underway at the time o f the study and the
downward mobility it induced, requires further study, since the argument is very
provocative.)
28. Cf. ECORYS-NEI (2003) for a review o f literature on business services and their clients.
Several relevant studies are presented in the recent collections by Dankbaar (2003) and
1 1 dd and Hull (2003).
 I N N O V A T I O N IN SERVIC ES
455

Re f e r e n c e s

A W., and U t t e r b a c k , J. (1978), "Patterns o f Innovation in Technology” Tech­

b e r n a t h y ,

nology R eview 80: 40-7,

♦ A ndersen , B.> H o w e l l s , f., H u l l , R., M i l e s , I., and R o b e r t s , J. (eds.) (2000), Knowledge
and Innovation in the New Service Economy, Cheltenham: Edward Elgar.
A C. (2000), "New Information Technology and Localized Technological Change
n t o n e l l i,

in the Knowledge-Based Economy,” in Boden and Miles 2000:170-91.

B a r r a s , R. (1986), “ Towards a Theory o f Innovation in Services,” Research Policy 15(4):
161-73*
-----(1990), “ Interactive Innovation in Financial and Business Services: The Vanguard of the
Service Revolution,” Research Policy 19: 215-37.
B ell , D. (1973)»The Coming of Post-Industrial Society, London: Heinemann.
B , C , H o u a r d , J., and M i c h a u x , B. (1986), Innovation and Research and Devel­
e l l e f l a m m e

opment Process Analysis in Service Activities, Brussels, EC, FAST. Occasional papers no 116.
Bessant, J., and R u s h , H. (2000), “ Innovation Agents and Technology Transfer” in Boden
and Miles 2000:156-69.
Boden , M., and M i l e s , I. (eds.) (2000), Services and the Knowledge Based Economy London:
Continuum.
Bo lisa n i , E., S c a r s o , E., M i l e s , I., and B o d e n , M. (1999), “ Electronic Commerce Imple­
mentation: A Knowledge-Based Analysis,” International Journal of Electronic Commerce
3(3); 53-6 9.
*B ryso n , J. R., and D a n i e l s , R W. (eds.) (1998), Service Industries in the Global Economy
2 vols., Cheltenham: Edward Elgar.
Cox, D., G um m ett , P., and B a r k e r , K. (2001), Government Laboratories— Transition and
Transformation, Amsterdam: IOS Press.
*D a n k b a a r , В. (ed.) (2003), Innovation Management in the Knowledge Economy London:
Imperial College Press.
Davies, A. (2003), “Are Firms Moving ‘Downstream" into High-Value Services,” in Tidd and
Hull 2003:321-41.
D en H e r t o g , P. (2000), “ Knowledge Intensive Business Services as Co-Producers o f Innov­
ation,"" International Journal ofJnnovation Management 4(4): 491-528.
-----and B ild er beek , R. (2000), “ The New Knowledge Infrastructure: The Role o f Tech­
nology-Based Knowledge-Intensive Business Services in National Innovation Systems,"" in
Boden and Miles 2000: 222-46.
D e p a r t m e n t o f T r a d e a n d I n d u s t r y (2003), The 2003 R&D Scoreboard (data provided by
Company Reporting Ltd), London: DTI available at: http://www.innovation.gov,uk/
projects/rd_scoreboard/introfr.html
ECORYS-NEI and CRIC (2003), Business Services: Contribution to Growth and Productivity in
the European Union, Report to European Commission DG Enterprise, Rotterdam:
ec o r ys .
E u r o p e a n C o m m i s s i o n (2000), European competitiveness report 2000, Brussels: Commis­
sion Staff Working Paper EN TR DT 2000/045/A1 Competitiveness, available at: http:// eur-
opa.eu.int/comm/enterprise/enterprise_policy/competitiveness/doc/compet_rep_ 2000/
cr-2000_en.pdf

* Asterisked items are suggestions for further reading.

45& IAN MILES

E v a n g e l is t a , R., and S a v o n a , M. (1998), “ Patterns o f Innovation in Services: The Results of

the Italian Innovation Survey,” paper presented to the 7th Annual RESER Conference,
Berlin, $—10 October; revised version forthcoming in Research Policy,:
F a u l h a b e r G., N oam , E., and T a s l e y , R. (eds.) (1986), Services in Transition, The Impact of
Information Technology on the Service Sector, Cambridge Mass.. Ballinger.
*F h G-ISI (Fraunhofer Institute for Systems and Innovation Research) (2003), Patents in the
Service Industries, Karlsruhe: FhG-ISI; report to the EC; available at: ftp://ftp.cordis.lu/
pub/indicators/docs/ ind_report_fraunhoferLpdf
F u c h s , V. (1968), The Service Economy>New York: NBER.
------ (1969), Production and Productivity in the Service Industries, New York: NBER.
G a d r e y , J., and de B a n d t , J. (1994), Relations de service, Marches de service, Paris: CNRS,
* ---- and G a l lo u j , E (eds) (2002), Productivity Innovation and Knowledge in Services,
Cheltenham: Edward Elgar.
G a l lo u j , C , and G a l lo u j , F. (2000) “ Neo-Schumpeterian Perspectives on Innovation in
Services,” in Boden and Miles 2000: 21-37.
G a l lo u j , R (2000), “ Beyond Technological Innovation: Trajectories and Varieties o f Services
Innovation,” in Boden and Miles 2000:129-45.
* ---- (2002), Innovation in the Service Economy: The New Wealth o f Nations, Cheltenham:
Edward Elgar.
------ and W e in s t e in , O. (1997), “ Innovation in Services,” Research Policy 26: 537-56.
G e r sh u n y , J. I. (1978), After Industrial Society? The Emerging Self Service Economy London:
Macmillan.
------ and M il e s , I. D. (1983), The New Service Economy: The Transformation of Employment in
Industrial Societies, London: Pinter.
G r a h a m , S. J, FL, and M o w ery , D. C. (2003), “ Intellectual Property Protection in the U.S.
Software Industry,” in W. Cohen and S. Merrill (eds.), Patents in the Knowledge-Based
Economy, Washington, DC: National Academy Press, 217-58.
G r e e n f ie l d , H. C. (1966), Manpower and the Growth of Producer Services, New York:
Columbia University Press.
H a l l , В. H. (2003), “ Business Method Patents and Innovation,” presented at the Atlanta
Federal Research Bank conference on Business Method Patents, Sea Island, Georgia, 3~5
April.
H ip p , C , T e t h e r , B., and M il e s , I. (2000), “ The Incidence and Effects o f Innovation in
Services: Evidence from Germany,” International Journal of Innovation Management 4(4):
417 - 5 4 *
-------------------- (2003), “ The Effects ot Innovation in Standardized, Customized and Be­
spoke Services: Evidence from Germany,” in Tidd and Hull 2003:175-210.
H o w ells , J, (1999), “ Research and Technology Outsourcing and Innovation Systems: An
Exploratory Analysis,” Industry and Innovation 6 :111—29.
(2001), The Nature o f Innovation in Services,” in D. Pilat (ed.), Innovation and
Productivity in Services, Paris: OECD, 55-79.
I n st it u t e of I n novation R esea rc h (2003), “ Knowing How, Knowing Whom: A Study of the
Links between the Knowledge Intensive Services Sector and the Science Basef Mimeo, loIR
(Manchester); Report to the Council for Science and Technology, available at: http://
wwwxst.gov.uk/cst/reports/files/knowledge-intensive-services-study.pdf
Jo n e s , P. (1995) Developing New Products and Services in Flight Catering,” International
Journal of Contemporary Hospitality Management 7(2/3): 24-8.
 I N N O V A T I O N IN S E R V IC E S

Kautonen, M. (2001), "Knowledge-Intensive Business Services as Constituents of Regional

Systems. Case Tampere Central Region, m M. Toivonen (ed.), Crowth and Stgnifcance of
Knowledge Based Services, Helsinki: Uusimaa TE Centre Publications 3.
K u u s is t o , J. (2000), "The Determinants of Service Capability in Small Manufacturing
Firms,” Ph.D. thesis, Kingston University Small Business Research Centre, Kingston.
------ and M e y e r , M. (2002), Insights into Services and Innovation in the Knowledge-intensive
Economy.i Helsinki: Finnish Institute for Enterprise Management, National Technology
Agency, Technology Review 134/2003.
Lay , G. (2002), Serviceprovider Industrie: Industrial Migration from Manufacturing to Selling
Products and Services, Karlsruhe: Fraunhofer Institute for Systems and Innovation
Research (FhG-ISI) 8 S (ISI-A-13-02).
Leiponen, A. (2001), Knowledge Services in the Innovation System, Helsinki, Etla: Working
Paper B185; SITRA 244 (publisher: Taloustieto Oy).
Levitt , T (1972), "Production Line Approach to Service,” Harvard Business Review 50(5):
41-52.
L ig h t , G., and M och , D. (1999), "Innovation and Information Technology in Services,” The
Canadian Journal of Economics 32(2): 363-83.
*М атнё , H., and S h a pir o , R. D. (1993), Integrating Service Strategy into the Manufacturing
Company, London: Chapman & Hall.
*Miozzo M., and M il e s , I. (eds.) (2003), Internationalization, Technology and Services,
Aldershot: Edward Elgar.
M i l e s , I. (1993), "Services in the New Industrial Economy,” Futures 25(6): 653-72.
----- (1999a), "Services and Foresight,” Service Industries Journal 19(2): 1-27.
— — (1999b), "Services in National Innovation Systems: from Traditional Services to Know­
ledge Intensive Business Services,” in G. Schienstock and O. Kuusi (eds.), Transformation
towards a Learning Economy: the Challenge to the Finnish Innovation System* Helsinki:
SITRA (Finnish National Fund for R8cD).
----- K astrin o s , N. (with K. F la n a g a n ), B ild e r b e ek , R., and den H ertog , P. (with
W. H u it in k and M, B o u m an ) (1995), Knowledge-Intensive Business Services: Users* Car­
riers and Sources of Innovation* Luxembourg: European Innovation Monitoring Service,
EIMS Publication no. 15 (ed./d-ooSoi mas).
------and M a tth ew s , M. (1992), "Information Technology and the Information Economy,”
in K. Robins (ed.), Understanding Information* London: Pinter.
*M uller , E., (2000), Innovation Interactions between Knowledge-Intensive Business Services
and Small and Medium-Sized Enterprises* Heidelberg and New York: Physica Verlag.
National S cien ce F oundation (2002), Science and Engineering Indicators 2002* Washing­
ton, DC: National Science Foundation (online version).
Pavitt , K. (1984), "Sectoral Patterns o f Technical Change: Towards a Taxonomy and a
Theory,” Research Policy 13(6): 343-73.
------(1994), "Key Characteristics o f Large Innovation Firms,” in M, Dodgson and
R. Rothwell (eds.), The Handbook of Industrial Innovation, Aldershot: Edward Elgar.
PREST (2002), A Comparative Analysis of Public* Semi-Public and Recently Privatised Research
Centers* Manchester: PREST available at: http:// les.man.ac.uk/PREST/Research/ Final%
20 Summary%2oReport.pdf
R o a ch , S. S. (1988), “ Technology and the Services Sector: America’s Hidden Competitive
Challenge in B. R. Guile and J. B. Quinn (eds.), Technology in Services, Washington, DC:
National Academy o f Engineering.
458 IAN MILES

S au ve , P., and M atto o , A, (eds.) (2003), Domestic Regulation and Service Trade Liberaliza­
tion, Oxford: Oxford University Press/World Bank.
Soete, L., and MiozzO) Nt. (1989)> Trade and Development in Services. A Technological
Perspective, Working Paper No. 89-031. Maastricht: MERIT.
S undbo , J. (1998), The Organization of Innovation in Services, Roskilde: Roskilde University
Press.
------ (2000), “ Organization and Innovation Strategy in Services” in Boden and Miles 2000:
109-28.
------ and G a l lo u j , R (2000), '"Innovation as a Loosely Coupled System in Services,” in
S. Metcalfe and I. Miles (eds.), Innovation Systems in the Service Economy, Dordrecht:
Kluwer.
TASC (1998), The Economics o f a Technology-Based Service Sector, National Institute of
Standards & Technology Program Office, Strategic Planning and Economic Analysis
Group, January 1998 NIST Planning Report: 98-102.
T e t h e r , B. S. et al. (2001), “Analysis o f CIS Data on Innovation in the Service Sector: Final
Report,” report to European Commission DG12, CRIC, University o f Manchester (112 pp.)
available at: http:// www.kiet.re.kr/ffies/econo/20021230-inno.pdf
—— H ip p , C , and M il e s , I. (2001), “ Standardization and Particularization in Services:
Evidence from Germany,” Research Policy 30 :115-38.
------ and S w a n n , G. M. P. (2003), “ Services, Innovation and the Science Base: An Investi­
gation into the UK's 'System o f Innovation' Using Evidence from the UK's Third Com­
munity Innovation Survey,” presented at the International Workshop: Innovation in
Europe: Empirical Studies on Innovation Surveys and Economic Performance, Institute of
Socio-Economic Studies on Innovation and Research Policy, National Research Council,
and University o f Urbino, Faculty o f Economics; Rome, 28 January 2003.
*T id d , J., and H u ll , F. M. (eds.) (2003), Service Innovation: Organizational Responses to
Technological Imperatives and Market Opportunities, London: Imperial College Press.
T o iv o n en , M. (2001), “ Megatrends and Qualification Requirements in the Finnish Know­
ledge Intensive Business Service Sector,” in M Toivonen (ed.), Growth and Significance of
Knowledge Based Services, Helsinki: Uusimaa TE Centre Publications 3.
T o m lin so n , M. (2000), “ Information and Technology Flows from the Service Sector:
a U K -Japan Comparison,” in Boden and Miles 2000: 209-221.
T o rdo ir , P, P. (1986), “ The Significance o f Services and Classifications o f Services,” in
P. Coppetiers, J.-C. Delaunay, J. Dyckman, J. Gadrey, F. Moulaert, and P. Tordoir, The
Functions of Services and the Theoretical Approach to National and International Classifica­
tions, Lille: Johns Hopkins University Center.
(1996), The Professional Knowledge Economy: The Management and Integration of
Professional Services in Business Organizations, Dordrecht: Kluwer.
U c h u p a la n a n , K. (1998), Dynamics of Competitive Strategy and IT-based Product-Process
Innovation in Financial Services: The Development of Electronic Banking Services in Thai­
land, D.Phil. thesis, University o f Sussex, Falmer, Brighton.
(2000), Competition and ІГ -based Innovation in Banking Services,” International
Journal of Innovation Management 4(4): 455-90.
 C H A P T E R 17

INNOVATION AND
DIFFUSION

B R O N W Y N H. H A L L

1 7. 1 I n t r o d u c t i o n 1

In 1953, a young female M acaq u e m o n key in the south o f Japan washed a muddy
sweet potato in a stream before eating it. T his obvious im provem ent in food
preparation was im itated q u ickly by other m onkeys and in less than 10 years it
became the n o rm in h er im m ediate group; by 1983, the m ethod had diffused
completely. In 1956, the sam e m onkey innovated again, inventing a technique in
which handfuls o f m ixed sand and w heat grains were cast u pon the sea, so that the
floating cereal could be skim m ed from the surface. A gain, by 1983, this m ethod o f
gleaning wheat had diffused alm ost com pletely throughout the local populations o f
Macaques.2 Besides the o b vio u s fact that hum ankind does not have a m o n o p o ly on
innovation, these exam ples illustrate a couple o f things about the diffusion o f
innovations: first, w hen they are clearly better than what went before, new ideas o f
how to do things w ill u sually spread via a “ learning by observing” process, and
second, the process can take som e tim e; in these cases it took thirty years, and the life
cycle o f the M acaque m o n k ey is som ew hat shorter than ours (Kawai, W atanabe, and
Mori 1992).
Turning to the w o rld o f h um ans, it is safe to say that w ithout diffusion, innovation
would have little social o r econom ic im pact. In the study o f innovation, the w ord
diffusion is co m m on ly used to describe the process b y w hich individuals and firm s in
4бО B R O N W Y N Н. H A L L

а so ciety/eco n o m y ad o p t a n ew tech nology, o r rep lace an o ld e r tech n o lo gy with a

newer. B u t d iffu sio n is n ot o n ly the m eans b y w h ich in n o v a tio n s b eco m e useful by
b ein g spread th ro u gh o u t a p o p u latio n , it is also an in trin sic p a rt o f the innovation
process, as learn in g, im itatio n , an d feedback effects w h ich arise d u rin g the spread of
a n ew tech n o lo gy enhance the o rig in al in n o v a tio n .3 U n d e rsta n d in g the diffusion
process is the key to u n d erstan d in g h o w co n scio u s in n o v a tiv e activities conducted
b y firm s and govern m en tal in stitu tion s (activities su ch as fu n d in g research and
d evelopm en t, tran sferrin g tech nology, lau n ch in g n ew p ro d u c ts o r creating new
processes) p ro d u ce the im p ro vem en ts in e co n o m ic an d so cial w elfare that are
u su ally the end goal o f these activities. F o r entities w h ich are 'c a tc h in g up,” such
as d evelo p in g econom ies, b ack w ard regions, o r tech n o lo g ica lly laggard firm s, diffu­
sion can be the m ost im p o rta n t p art o f the in n o vative p ro cess.4
T h irty years ago, an eco n o m ic h isto rian (R o sen b erg 1972) m ad e the following
o b servatio n ab ou t the d iffu sio n o f in n o vation s:

in the history o f diffusion o f many innovations, one cannot help being struck by two
characteristics o f the diffusion process: its apparent overall slowness on the one hand,
and the wide variations in the rates o f acceptance o f different inventions, on the other.
(Rosenberg 1972:191)

E m p irical m easu rem en t an d stu d y since then has c o n firm e d this view. This
chapter and the references in clu d ed in it review the d iffu sio n o f a num ber o f
in ven tion s and in n o vative processes, fro m the b o ilin g o f w ater to preven t diarrheal
diseases to m o b ile teleph ony in E u rop e. B o th these stu d ies an d the figu res showing
d iffu sio n rates in v ario u s countries d em o n strate the tru th o f R o sen b erg 's statement.
T h e studies go fu rth er than sim p ly n o tin g the speed an d v a ria tio n o f diffusion, in
that they correlate the rates o f ad o p tio n w ith ch aracteristics o f the technologies and
their poten tial adopters in an attem pt to exp lain the sp eed o f d iffu sio n and the
ultim ate acceptance o f the new p ro d u ct. B esides the w id e v a ria tio n in acceptance o f
in n o vation s, a second im p o rtan t ch aracteristic o f the d iffu sio n p ro cess is the way in
w h ich it interacts w ith the in n o vative process. T h is has p erh ap s been a som ew hat less
studied aspect o f d iffu sio n , o w in g to the d ifficu lty o f co llectin g system atic data, but
case studies aboun d. R osen b erg (1982), a m o n g oth ers, has em p h asized the fact that
the d iffu sio n o f in n o vation s is often accom p an ied b y le arn in g ab o u t their use in
d ifferent en viron m en ts, and that this in tu rn feeds b ack to im p ro vem en ts in the
origin al in n o vation .
W h y is d iffu sio n som etim es slow? W h y is it faster in so m e co u n tries or regions
than others, and fo r som e in n o vation s than fo r others? W h at facto rs exp lain the wide
variatio n in the rate at w h ich it occurs? T h is ch ap ter p ro v id e s a historical and
co m p arative perspective on d iffu sio n that lo ok s at the b ro a d determ in an ts, eco­
n o m ic, social, and institu tion al. The w ays in w h ich the d ifferen t so cial scientific
disciplines think about d iffu sio n is discussed and a fra m e w o rk is presented for
stu d yin g its determ inants. Som e o f the em p irical evid en ce o n these determ inants
 IN N O V ATIO N AND DIFFUSION 461

is reviewed, and a range o f exam ples given. The chapter concludes with a discussion
of gaps in our understanding and future research questions.

17. 2 C o n c e p t u a l F r a m e w o r k s

The diffusion o f innovations has been studied from a num ber o f different perspec­
tives: historical, sociological, econom ic (including business strategy and
marketing), and netw ork theoretical. The choice o f approach is often dictated by
the use to which the results w ill be put, but there is no doubt that insights from one
perspective can in fo rm the research in another discipline. Perhaps a key exam ple o f
this is the w ay in which historical study o f the developm ent and spread o f certain
major inventions has affected h ow econom ists understand the role o f the diffusion
process in determ ining the dynam ics o f p ro d u ctivity change, a topic I return to later
in this chapter. First, I lay out som e o f the fram ew orks that have been used by
different disciplines fo r the analysis o f diffusion.
The sociological and organizational literature is exem plified b y Rogers" well-
known book, Diffusion o f Innovations, now in its fourth edition. In this book, he
reviews the subject p rim arily from a sociological perspective, but one that is
informed by research on organizations, the role o f econom ic factors, and the
strategies o f firm s and developm ent agencies. Rogers provides a useful set o f five
analytic categories that classify the attributes that influence the potential adopters o f
an innovation:
(1) The relative advantage o f the innovation.
(2) Its com patibility, w ith the potential adopter’s current w ay o f doing things and
w ith social norm s.
(3) The co m plexity o f the innovation,
(4) Trialability, the ease w ith w hich the innovation can be tested b y a potential
adopter.
(5) O bservability, the ease w ith w hich the innovation can be evaluated after trial.
M ost o f these attributes are recognizable in one form or another in the m any
analyses o f specific in n ovation s that have been undertaken by researchers in the past,
albeit under different nam es. F o r exam ple, both trialability and observability are
characteristics that speak directly to the level o f uncertainty faced b y a potential
adopter. The latter characteristic is a key feature o f the real options m odel o f
technology choice w hich is discussed later in this chapter and which underlies
some o f the w o rk on tech nology adoption by business firm s. C om plexity as a
determ inant is clearly related to the econ om ist’s notions o f cost and com plem entary
462 B R O N W Y N Н. H A LL

investm ent, as is relative advantage, w h ich an eco n o m ist m ig h t co n sid er to be

d eterm in ed p rim a rily b y the benefit/cost ratio o f ad o p tin g the n ew technology.
B u t u n d erstan d in g the w a y in w h ich the d iffu sio n p ro cess u n fo ld s, in add ition to
sim p ly id en tifyin g features that d eterm ine its u ltim ate success o r failu re, requires a
larger fram ew o rk , one also p ro vid ed b y R ogers later in the sam e vo lu m e . In addition
to the attributes listed above, w h ich influ ence the a d o p tio n d ecisio n at the individual
level, he p o in ts to a v ariety o f extern al o r so cial co n d itio n s that m a y accelerate or
slo w the process:
(1) W h eth er the d ecision is m ade collectively, b y in d iv id u a ls, o r b y a central
authority.
(2) T h e co m m u n icatio n channels used to acq u ire in fo rm a tio n ab o u t an innov­
ation, w h eth er m ass m edia o r interperson al.
(3) T h e n atu re o f the social system in w h ich the p o ten tial ad o p ters are em bedded,
its n o rm s, and the degree o f interconnectedness.
(4) T h e extent o f change agen ts5 (advertisers, d evelo p m en t agencies, etc.) pro­
m o tio n efforts.
Like so m an y students o f the d iffu sio n process, R o g ers im p lic itly assum es that
neither the n ew in n o va tio n n o r the tech n o lo gy it replaces ch an ges d u rin g the
d iffu sio n process an d that the n ew is better than the old. T h ese assu m p tio n s have
been challenged stron gly b y R o sen b erg (19 7 2 ,19 8 2 ), w h o a rg u ed that not o n ly was
the n ew tech n o lo g y im p ro ved as user exp erien ce an d feed b ack accu m u lated , but
also that freq u en tly the replaced tech n o lo g y exp erien ced a 'la s t gasp 55 im provem ent
due to co m p etitive pressure and that this fact co u ld slo w the d iffu sio n o f the new.
A freq uen tly given exam p le is the rap id p ro d u c tiv ity increase in sailin g ships during
the nineteenth ce n tu ry d ocu m en ted b y G ilfillan (1935 a, 1935 b).
In con trast to the focu s on the extern al en viro n m en t favo red b y sociologists and
students o f o rgan izatio n al beh avior, m an y eco n o m ists have ten ded to view the
process as the cu m u lative o r aggregate result o f a series o f (ratio n al) individual
calculation s that w eigh the increm ental benefits o f a d o p tin g a n ew technology
against the costs o f change, often in an en viro n m en t ch aracterized b y uncertainty
(as to the fu tu re path o f the tech n o logy and its ben efits) and b y lim ite d inform ation
(about b oth the benefits and costs an d even ab ou t the v e ry existence o f the technol-
o gy). A lth o u g h the ultim ate d ecision is m ade b y d em an d ers o f the technology, the
benefits an d costs are often in flu en ced b y d ecision s m ad e by su p p liers o f the new
technology. T h e resulting d iffu sio n rate is then d eterm in ed b y su m m in g o ver these
in d iv id u a l decisions.
T h e virtu e o f this ap p ro ach to th in k in g ab ou t the a d o p tio n o f in n o va tio n s is that
it is gro u n d ed in the d ecision m ak in g o f the m ic ro e c o n o m ic u n it, b u t this virtue
com es w ith a cost, in that it ignores the social feedback effects (o r extern alities, to use
the econ o m ists term ) that m igh t result fro m on e in d iv id u a l a d o p tin g and therefore
en co u ragin g another. N aturally, in the recent past, e co n o m ists h ave risen to this
 IN N O VATIO N AND DIFFUSION 463

challenge and included such concepts as network effects in their m odels (see the
discussion in Box 17.1). N evertheless, the factors and m echanism s considered in
most o f their studies typ ically fall short o f m any that other disciplines might
consider im portant, such as social connectedness. A n interesting early debate on
this topic that reflected different view s o f the determ inants o f hybrid-corn adoption
in the United States was conducted by a pioneering econom ist in the study o f
diffusion and a n u m ber o f sociologists including Rogers in the pages o f Rural
Sociology (Babcock 1962; G riliches 19 60a and b, 1962; H avens and Rogers 1961;
Rogers and H avens 19 6 2).5 L o o kin g back at this debate from today, a reasonable
conclusion is that both econ om ic and non-econom ic factors probably mattered for
the diffusion o f hybrid corn, although econom ic factors by themselves did a pretty
good job explaining variatio n across states.
As an exam ple o f m icro eco n o m ic analysis o f the adoption decision in a m odern
technological setting, consider the decision to replace a w ired physical connection to
the Internet with a w ireless one, either at hom e or in an office. Benefits m ight include
the ability to w o rk on the n etw ork th rou gh ou t o n e s house or w orkplace rather than

Box 17.1 The QWERTY controversy—diffusion with network externalities

In an influential article published in 1985, Paul David proposed an answer to the

question o f why most keyboards have the QWERTYUIOP layout today, even though
studies done in the first half o f the twentieth century show that those trained on a
keyboard with the Dvorak layout are able to type more quickly. He attributed this
outcome to the importance o f lock-in where there are network externalities. The
argument is that the invention o f touch typing in the late 1880s made typewriters a
network good because o f the interrelatedness between the keyboard layout and the
typist's skills, the economies o f scale in the user costs o f typewriting due to training,
and the quasi-irreversibility o f investment in learning how to type. By the 1890s, these
factors led to a significant lock-in to QWERTY layout, because it was easier to recon­
figure the keyboard than to retrain the typist. The conclusion from this story of the
diffusion o f a new technology with network characteristics is that it is possible that the
version o f technology adopted (the “ standard") was not the necessarily the “ best"
available, because o f path dependence in the diffusion process induced by network
externalities. That is, small accidents early in the choice o f technologies can lead to the
adoption o f an inferior standard because the existence o f an installed base makes that
technology more attractive to new adopters. This point was also made by Brian Arthur
( 1919) using probability models o f stochastic diffusion processes developed by Arthur,
irmoliev, and Kaniovski 1983).
David's view has been challenged forcefully by Liebowitz and Margolis ( 1990), on at
least two grounds: First, they show that the historical evidence that the Dvorak
keyboard was preferable may be weak. Second, they argue that if society faces large
enough costs from adopting the wrong standard, it will pay individuals to change the
standard via some form o f collective action, One version o f David s response to this
critique was published by the Economist magazine in 1999
464 B R O N W Y N H. H A L L

at a fixed lo catio n such as a desk, and the absence o f w ires. T h e y m ig h t also include
the fact that several m em b ers o f the h o u seh o ld can be o n lin e at the sam e tim e using a
single telephone connection. T h e costs in clu de the pu rch ase o f a base statio n and the
services o f a tech nician to install it, but th ey m a y also in clu d e the tim e o f the user
(adopter) spent reco n fig u rin g his o r her co m p u ter an d en su rin g th at all the com ­
m u n icatio n tools needed are w o rk in g . C o sts m igh t also in clu d e the acq u isition o f
n ew softw are, o r the tim e spent train in g oth er m em b ers o f the h o u seh o ld o r office in
its use. W ere we to enrich this sto ry to in clu d e the a d o p tio n e n viro n m en t, w e m ight
focu s o n such factors as w h eth er n eigh b o rs o r colleagues alread y h ad undertaken
such an installation, the extent to w h ich it has been ad vertised b y the su p p lier o f the
tech n o lo gy (o r the extent to w h ich it has been “ sp o n so re d ” b y a g o vern m en t agency
o r lead in g firm ), and even the state o f d evelo p m en t o f the n ew tech n o lo g y and the
o p eratin g system necessary to use it (a co m p le m e n tary in p u t). N o te also that m ost o f
these factors have been ch an gin g rap id ly over tim e.
A s alluded to earlier, the first em p irical stu d y o f the d iffu sio n o f tech n o lo g y b y an
econ o m ist w as G rilich es’ (1957) stu d y o f the d iffu sio n o f h yb rid c o m seed in the
M id w estern U n ited States. T h is stu d y em phasized the role o f e c o n o m ic factors such
as expected pro fits and scale in d eterm in in g the v a ry in g rates o f d iffu sio n across the
M id w estern states. A t the sam e tim e, it fo u n d that the v a ria tio n in in itial start dates
fo r the process depended on the speed w ith w h ich the seed w as cu sto m ized fo r use in
p articu lar geograph ic areas. T h at is, d iffu sio n dep en d ed to a certain extent on the
activities o f the su p p liers o f the tech n o lo gy in ad ap tin g it to lo cal co n d itio n s, again
h igh ligh tin g the ten d en cy fo r the fu n d am en tal ch aracteristics o f the tech n ology to
change som ew h at d u rin g the a d o p tio n process. T h is th em e is repeated throughout
the h isto ry o f in n o vatio n . B ru lan d (1998, 2002) fin d s that the n in eteen th -cen tury
d evelo p m en t o f the N o rw eg ian textile in d u stry w as g reatly facilitated b y the
tech n o lo g y tran sfer activities u nd ertaken b y the m o stly B ritish m a c h in e ry suppliers
in the fo rm o f train in g, in creasin g the su p p ly o f skilled w o rk ers in N orw ay.
T h e m arketin g literature on d iffu sio n is p rim a rily fo cu sed o n tw o questions: how
to encou rage co n su m ers and cu stom ers to pu rch ase n ew p ro d u c ts o r technologies,
and h o w to detect o r forecast success in the m arketplace. T h at is, it o ften looks for
factors that can be in flu en ced in o rd er to increase the n u m b e r o f agents that will
ch oo se a p articu lar p ro d u ct. For this reason, the literatu re tends to em ph asize factors
such as m ed ia in fo rm a tio n o r the role o f social n etw orks and ch an ge agents, as well
as the characteristics o f the p ro d u ct itself, rather than in d iv id u a l ad o p ter factors
such as ed u catio n and incom e levels that are less subject to m a n ip u la tio n b y the
m arketin g organ ization . The w o rk h o rse m odel in m ark etin g fo r m a n y years has
been the Bass (1969) m o del, w hich assum es that m ass m ed ia are im p o rta n t early on
in the d iffu sio n process but that as tim e passes, in terp erso n al co m m u n icatio n
becom es far m o re im p o rtan t. E stim atio n o f this m o d el on a n u m b e r o f consum er
d urables has revealed that interperson al c o m m u n ica tio n p lays a m u ch bigger role
than the m edia in d iffu sio n (R ogers 1995). For an in terestin g d iscu ssio n o f the
 contrast between the econ om ic and m arketing view s and a com parison o f models
from the two literatures, see Zettelm eyer and Stonem an (1993). Recent work on
identifying and forecasting success in the m arketing literature is illustrated by
Golder and Tellis (1997)- I defer discussion o f their m odel to later in this chapter
when I discuss som e o f the findings obtained b y Tellis, Strem ersch, and Yin (2002)
using this m ethodology.
The activist view o f d iffu sion taken by the m arketing literature is also that pursued
by specialists in tech nology policy, w ho are generally interested in encouraging the
adoption o f particular new technologies for w elfare-enhancing reasons, either
because it serves particu lar pu blic po licy goals (such as encouraging the boiling o f
water to reduce disease in less-developed countries) or because certain technologies
are viewed as conferring externalities on society as a w hole (such as the adoption o f
Internet use or vaccin ation against a com m unicable disease). In understanding the
variation across countries in d iffu sion , variables describing their institutions and
culture have proved essential in som e cases (but not all, see the discussion o f Tellis,
Stremersch, and Y in 2002 in Section 17.5).

17.3 M o d e l i n g D i f f u s i o n

The most im portant th in g to observe abou t the decision to adopt a new invention is
that at any point in tim e the choice being m ade is not a choice between adopting and
not adopting but a choice betw een ad optin g now or deferring the decision until later.
It is im portant to lo o k at the decision in this w ay because o f the nature o f the benefits
and costs. By and large, the benefits fro m adopting a new technology, as in the
wireless com m unications exam ple given above, are flow benefits that are received
throughout the life o f the acquired innovation. H owever, the costs, especially those
of the n on-pecuniary ‘T eam in g” type, are typically incurred at the tim e o f adoption
and cannot be recovered. There m a y be an on goin g fee for using som e types o f new
technology, but it is u sually m uch less than the initial cost. Econom ists call costs o f
this type “ sunk.” T h at is, ex ante> a potential adopter weighs the fixed costs o f
adoption against the benefits he expects, but ex post, these fixed costs are irrelevant
because a great p art o f them have been su n k and cannot be recovered.
The argum ent that ad o p tion is characterized b y sunk costs im plies two stylized
facts about the adoption o f new technologies: first, adoption is usually an absorbing
state, in the sense that w e rarely observe a new technology being abandoned in favor
of an old one.6 T h is is because the decision to adopt faces a large benefit m inus cost
hurdle; once this hu rd le is passed, the costs are sunk and the decision to abandon
466 B R O N W Y N Н. H A LL

requires givin g up the b en efit w ith o u t regain in g the cost, so even i f the gross benefit
is reduced relative to w hat w as expected, the net ben efit is still lik ely to be positive.
S eco n d , u n d er u n certain ty ab ou t the benefits o f the n ew tech n o logy, there is an
o p tio n value to w aitin g b efore sin k in g the costs o f a d o p tio n , w h ich m a y tend to delay
a d o p tio n .7
A n im p o rtan t exception to the rule that a d o p tio n is n o rm a lly an ab so rb in g state is
the p o ssib ility o f fads o r fashion s, w h ich m igh t be d efin ed as th in gs su ch as the “ hula
h o o p ” craze o r vario u s types o f w eigh t-loss diets, w h ich d iffu se ra p id ly and then
d isap p ear after a tim e. The experien ce o f a w ave o f a d o p tio n fo llo w ed b y a wave o f
disuse seem s to be som ew h at m o re likely in the case o f in n o v a tio n s in “ practice,”
such as m edical practice or business practice, than in the case o f p h ysical products,
p o ssib ly because in the latter case the costs that are su n k are o u t o f p o ck et costs paid
to others, w hereas in the fo rm er m u ch o f the cost (alth o u g h b y n o m ean s all) comes
in the fo rm o f the ad o pter's tim e an d effort. T h at is, the p o ssib ility o f sun k costs
m a y lo o m larger to the ad o p ter w hen d en o m in ated in d o lla r o r eu ro sym bols.
N elson et al. (2002) discuss this p h en o m en o n m o re fu lly an d g ive so m e exam ples
(such as the q u ality circle m o vem en t). T hese au th o rs place co n sid erab le em phasis
o n the d ifficu lty in these cases o f getting feedback that the in n o va tio n truly is
an im p ro vem en t. R elatively lo w su n k costs co m b in ed w ith u n certain benefits
w ill m ean that the d ecision to ad o p t is m o re easily reversible in the case o f practices.
Strang and Sou le (1998) also discuss the cyclicality o f fash io n s in business
practices.
It is a w ell-k n o w n fact that w h en the n u m b er o f users o f a n ew prod u ct or
in ven tion is plotted versus tim e, the resu lting cu rve is ty p ic a lly an S-shaped or
ogive d istrib u tio n . T h e n o t v e ry su rp risin g im p lica tio n is that a d o p tio n proceeds
slo w ly at first, accelerates as it spreads th ro u g h o u t the p o ten tial ad o p tin g p o p u la­
tio n , and then slow s d o w n as the relevant p o p u latio n b eco m es satu rated. In fact, the
S-sh ap e is a natu ral im p licatio n o f the o b servatio n that a d o p tio n is u su ally an
ab so rb in g state. Figu re 17 .1, w h ich represents the d iffu sio n o f electric m otors in
U S m an u factu rin g betw een 1898 and 1955, show s su ch a cu rve. In 1898, the share o f
m an u factu rin g h o rsep o w er p ro d u ced b y electric m o to rs w as ab o u t 4 per cent. It
increased stead ily and sm o o th ly betw een 1900 and ab ou t 1940, at w h ich p o in t nearly
all h o rsep o w er is p ro d u ced b y electricity. S atu ratio n ap p ears to be reach ed at around
90 per cent, p resu m ab ly because fo r som e specialized uses, o th er typ es o f m otors are
preferred.
L o o ked at in term s o f the benefits and costs o f tech n o lo g y a d o p tio n , a range o f
sim p le assu m p tio n s w ill generate this curve. T h e tw o lead in g m o d els explain the
d isp ersion in a d o p tio n tim es using tw o d ifferent m ech an ism s: c o n su m e r heterogen­
eity, o r co n su m er learning. The heterogen eity m o d el assu m es that d ifferent con­
su m ers expect to receive d ifferent benefits fro m the in n o va tio n . I f the d istrib u tion o f
benefits over co n su m ers is n o rm al (or a p p ro x im a te ly n o rm a l, that is unim odal
with a central tendency), the cost o f the new p ro d u ct is co n stan t o r declines
 INNOVATION a n d d iffu sio n 467

Year

Figure 17.1 Diffusion of electric motors in US m anufacturing

monotonically over tim e, and it is assum ed that consum ers adopt when the benefit
they receive for the prod u ct is greater than its cost, the diffusion curve for the
product will have the fam iliar S-shape.
An im portant alternative m odel is a learning o r epidem ic m odel, which is m ore
popular in the sociological and m arketing literatures (the Bass m odel is an exem ­
plar), but has also been used b y econom ists. In this m odel, consum ers can have
identical tastes and the cost o f the new technology can be constant over tim e, but not
all consumers are in form ed ab ou t the new technology at the sam e time. Because each
consumer learns about the tech n o logy fro m his or her neighbor, as tim e passes, m ore
and more people adopt the tech n ology du ring any period, leading to an increasing
rate o f adoption. H ow ever, eventually the m arket becom es saturated, and the rate
decreases again. This too w ill generate an S-shaped curve for the diffusion rate.8 In
general, com bining this m odel w ith the previous m odel sim ply reinforces the
S~shape o f the curve. C o ld e r and Tellis (1997) define a concept they call “ take-off,”
which is their attem pt to id en tify the p oin t at which the em pirical diffusion curve
appears to have its greatest inflection relative to the initial growth rate.9 For the data
in Figure 17.1, this poin t w o u ld be in about 1910. Because for m any consum er
products the existence o f such a p o in t is a good predictor o f eventual success, the
focus o f their w ork is to id en tify predictors o f this point.
Regardless o f the details o f the m echanism generating the probability distribution
of adoption tim es, the question w hich concerns both social scientists and those
interested in encouraging the spread o f new technologies is the question o f what
factors affect the rates at w h ich these events occur. A second and no less interesting
question is what are the determ inants o f the ceiling at which the S-curve asym ptotes.
That is, when w ou ld we expect this ceiling to be less than 100 per cent o f the potential
# 8 BRONWYN H. HALL

user base? T h e next section o f this ch ap ter review s these facto rs an d som e o f the
em p irical evidence co n cern in g their im p o rtan ce.

17.4 D e t e r m i n a n t s of t h e
D iffusion Rate

Figure 17.2 sh ow s the n u m b er o f U S h ou seh o ld s that have a d o p te d p articu lar new

in ven tion s as a fu n ctio n o f tim e. A lth o u g h n o t sm o o th , these cu rves clearly follow
the S-sh ap ed pattern n oted b y m a n y observers. T h e y also exh ib it the characteristic
w ide v a ria tio n in the elapsed tim e fo r d iffu sio n . F o r exam p le, it to o k o ver fo rty years
for the clothes w ash er to go fro m o n e-q u a rter o f all h o u seh o ld s to three-quarters,
whereas it to o k less than ten years fo r the vid eo cassette reco rd er o r c o lo r television
(not sh ow n ) to m ake the sam e leap. Table 17 .1 sh ow s the d iffu sio n o f com m on
household electronic applian ces in Jap an betw een 1989 an d 1995. It is notew orthy
that there is con siderab le variatio n in the d iffu sio n rates fo r d ifferen t p ro d u cts even
during the sam e six-year p erio d , and this v a ria tio n is n o t exp la in ed b y the level o f
diffusion that w as alread y achieved in 1989 (co m p are the refrig e ra to r to the air
:o n d itio n er, o r the C D /cassette/radio player to the v id e o cam era).

Year

Figure 1 7 .2 Diffusion of major innovations in the United States

Source; Dallas Federal Reserve Bank.
 INNO VAT ION AND DIFFUSION 469

Щ е І І І Ш Штіт in Japanese Households (%}

NewPnxhict 1989 1995 Change

Confess phones NA 43.7 NA

'Щ а А Ііз я е іІе player 31.5 68.2 36.7
Qjovection heater/cooler 34.7 57.3 22.6
Washing machine 34.7 55.4 20.7
Wort processor 25.1 43.7 18$
Microwave own 72.9 89$ 16$
Vidro camera 17.5 34$ 16$
Mr conditioner 64.8 79*3 14$
МйояюЬІІе 78$ 82.1 5$
12.4 ms 4*2
тл 9 9 .3 0J
62.9 ш 0*7

Bulletin No. 18.

From the considerations reviewed earlier in the chapter, one can derive a list o f
factors that m ight be expected to influence the d iffu sion o f innovations. These can be
classified into four main groups, those that affect the benefits received, those that
affect the costs o f adoption, those related to the industry or social environment, and
those due to uncertainty an d information problems. Alternatively, using the clas­
sification system o f Rogers, one can identify the first and second as combining to
yield relative advantage an d com plexity, the third as com patibility, and the fourth as
being determined by trialability and observability.

174*1 Benefit Received from the New Technology

Qeariy the m ost im portant determinant o f the benefit derived from adopting a new
technology is the am ount o f improvement which the new technology offers over any
previous technology* This is to a great extent determined by the extent to which there
® st substitute older technologies that are fairly close. For example, in Figure 17*2*
we see that rad io an d the autom atic clothes w asher were both introduced in the
United States in the early 1920s, but that diffusion o f the former was m uch m ore
than the latter. This may be partly because a fairly good substitute for the
470 B R O N W Y N H , HALL

au to m atic clothes w ash er in the fo rm o f m an u al clothes w a sh in g m ach in es existed

w h ereas there w as no v e ry g o o d substitute fo r rad io. It is also co n sisten t w ith the
Tellis et al. (2002) fin d in g that across E u ro p ean co u n tries d u rin g the latter h a lf o f the
tw entieth century, the single m o st im p o rta n t facto r that exp lain s speed o f d iffu sio n
is w h eth er the go o d in q u estio n is “ w h ite ” (h o u seh o ld ap p lian ce) o r “ b row n ”
(en tertain m en t o r in fo rm a tio n co n su m er d u ra b le ). T h ese au th o rs hypothesize
that the general exp lan atio n fo r this fin d in g is that “ b ro w n ” g o o d s are m ore
status-enhancing, in that they are more readily observable to non-m em bers o f the
h ou seh old . U n fo rtu n ately th ey d id not co n tro l fo r the p rices o f the g o o d s because o f
lack o f consistent data across cou ntries, so it is difficult to k n o w w h eth er this finding
might also be related to differences in the full costs o f adoption across goods and
countries.
A n im p o rta n t facto r in e xp lain in g the slow ness o f tech n o lo g y a d o p tio n is the fact
that the relative advantage o f n ew tech n ologies is freq u en tly rath er sm all w hen they
are first introduced. As many authors have emphasized, as diffusion proceeds
learning about the tech n o lo gy takes place, the innovation is im proved and adapted
to d ifferent en viron m en ts, thus m ak in g it m o re attractive to a w id e r set o f adopters
(R o sen b erg 1972; N elson et al. 2002). T h e im p licatio n is that the b en efits to adoption
gen erally increase o ver tim e; i f th ey increase faster than costs, d iffu sio n w ill appear to
be delayed (because the number o f potential adopters will increase over time,
expanding the size o f the adopting population). In the Rosenberg (19S2) study, the
leading example was the airframe, specifically the stretching o f the Boeing 747, but in
fact one could argue that any technology in which learning by doing or using is an
im portant aspect o f its development will display feedback between diffusion and
innovation. A good example might be applications software, most o f whose devel­
opment after initial launch is dictated by the experience and dem ands o f users, or the
worldwide web, where enhancements after the first web browser was created were
dramatic.

17,4*2 Network Effects

Increasingly, the valu e o f som e new tech n o lo g y to the co n su m er d ep en d s p artly on
the extent to w h ich it is ad o pted b y other co n su m ers, either becau se the technology
is used to co m m u n icate w ith others (such as the In ternet, o r in stan t m essaging) or
because the p ro v isio n o f softw are and services fo r the tech n o lo g y dep en d s on the
existence o f a large cu sto m er base. G o o d s o f this type are u su ally term ed netw ork
go od s b y econom ists: their c h ie f ch aracteristic is that th ey rely o n standards to
ensure that they can communicate either directly or indirectly. For these goods, an
important determinant o f the benefit o f adoption is therefore the current or
expected n etw ork size.
 INNOVATION AND DIFFUSION

For exam ple, Saloner and Shepard (1995) exam ine the adoption o f ATM machines
by banks, under the assum ptions that consum ers prefer a larger network o f ATM
machines to a sm aller and that banks respond to consum er preference. These
authors do indeed find that banks with m ore branches adopt an ATM network
sooner, even after con trollin g for overall bank size, and argue that this confirm s that
a higher netw ork valu e leads to earlier adoption o f a new technology, other things
equal.10 This exam ple illustrates both the im portance o f networks and also the role
o f large firm s as interm ediates between technology and consum ers in sponsoring
particular standards for networks.
A fam ous exam ple o f the role o f "n etw o rk externalities” in consum er adoption o f
new technologies is the V H S/B eta com petition, which resulted eventually in a single
standard for video recorder/players in a large part o f the w orld. M ost observers
attribute this outcom e to the consum er desire for a large range o f software in the
form o f pre-recorded tapes to go w ith this hardware, and to the fact that V H S had an
initial early advantage in the length o f program that could be recorded. See Park
(2002) for details on the diffusion o f this technology to consum ers.
Although netw ork effects (particularly those from networks that diffuse know ­
ledge about o r experience w ith an innovation) have always been view ed as im portant
for the diffusion o f innovations, especially in the sociological literature, recent w ork
in econom ics has focused on the role played by standards in accelerating or slowing
the diffusion process, as in the V H S/B eta exam ple (D avid 1985; Katz and Shapiro
1985; A rthur 1989; E con om ides and H im m elberg 1995). The central message o f the
modern econom ic literature on standards and netw ork externalities is that con­
sumers and firm s receive benefits from the fact that other consum ers and/or
firms have chosen the sam e tech nology that they have. These benefits are viewed
as being o f tw o kinds, direct and indirect. D irect netw ork benefits are those that
arise because they allow the adopter to com m unicate with others using the same
technology. Exam ples are the choice o f fax com m unication technology or the choice
o f w ord processor docum ent form at. Indirect benefits arise from the fact that
adoption o f a p ro d u ct that uses a particular technological standard by a greater
num ber o f people increases the p rob ab ility that the standard will survive and that
goods com patible w ith that standard w ill continue to be produced. The V H S/Beta
example alluded to earlier can be view ed as an exam ple where indirect network
benefits w ere very im po rtan t, although direct benefits presum ably also play a part
(the benefits from being able to loan a video m ade on one's ow n m achine to a friend
or neighbor).
The close connection between technological standards and network externalities
comes from the fact that standards create a number o f effects all o f which go in the
direction o f making it more likely that a good will exhibit network externalities.
First, a technological standard increases the probability that communication
between two products such as telephones, instant messaging services, or a CD player
and a CD, will be successful Second, standards ease consumer learning and
472 B R O N W Y N H, H A L L

encourage adoption when the same or sim ilar standards are used in a range o f
products. The use o f a particular standard, such as a W indows operating system, by
others in a consumer's network, also helps learning and will encourage adoption,
because o f the relative ease with which a new adopter can obtain advice from those
nearby. Third, a successful standard increases the size o f the potential market for a
good, which can be im portant in lowering the cost o f its production and in
increasing the variety and availability o f com plem entary goods. Besides the VH S/
Beta example referred to earlier, an example o f this latter effect m ight be the wider
availability o f software for the W indows operating system, in com parison to M acin­
tosh OS or Linux.
A lth o u g h standards have alw ays m attered fo r d iffu sio n , the in creasin g im p o rt­
ance o f d igital an d in fo rm a tio n tech nologies have in creased th eir salien ce and led to
a v ariety o f “ standards b attles" and to strategic b e h a v io r o n the p art o f firm s that
h op e to in fluen ce their ad o p tion . E arlier exam p les o f stan d ard s battles are the
co m p etitio n betw een A C and D C m ethods o f d istrib u tin g electricity (D avid
19 9 0 a), an d the failu re o f gas-p o w ered refrigerators to succeed in the m arket despite
their app aren t efficiency, because o f the sp o n so rsh ip o f electric p o w e r b y G E and
W estinghouse (R ogers 1995). N evertheless, it is clear that the im p o rta n ce o f this
p h en o m en o n has increased recently, w ith increase in in fo rm a tio n an d co m m u n i­
cation tech nologies. C o n sid er fo r exam p le, the battle b etw een N etscap e an d M icro ­
soft Internet E x p lo rer fo r d o m in an ce in the w eb b ro w ser m arket.
T h e increase in the im p o rtan ce o f standards that has acco m p a n ied the grow th in
im p o rta n ce o f the in fo rm atio n and teleco m m u n icatio n s in d u stries has led to a wave
o f eco n o m ic m o d elin g. T hese m odels in co rp o rate the in creasin g retu rn s p h en o m ­
enon that results fro m the p o sitive feedback fro m installed base to a d o p tio n b y other
consum ers. A n early effort is that b y A rth u r, E rm o liev, an d K a n io v sk i (1983), which
em phasizes D a v id 's insigh t that w h ere there are m u ltip le p o ssib le stand ard s, small
events early in the process that fav o r one o f the stan d ard s can lead to an adoption
process that settles on an in ferio r standard. B y ad d in g h etero g en eity in consum ers'
tastes o r lo calization in in fo rm a tio n spillovers, later researchers h ave p ro d u ced m ore
co m p lex m o d els o f d iffu sio n in the presence o f n etw o rk extern alities that results in
m o re than on e stand ard su rv iv in g in the m arket even in the presen ce o f increasing
returns in ad o p tio n (B assan in i and D osi 1998; W endt an d v a n W estarp 2000).
In d u strial o rgan izatio n and strategy theorists have centered th eir m o d e lin g efforts
on the im p licatio n s o f increasing returns in a d o p tio n fo r co m p etitive strategy and
m arket structure. E xam ples o f this literature include K atz an d S h ap iro (19 8 5,198 6 ,
l 994 )y Farrell and Salo n er (1992), and Sh ap iro an d V arian (1999). In a series o f papers,
Katz and S h ap iro have explored the im p licatio n s o f co n su m er a d o p tio n b eh avio r in
the presence o f n etw ork externalities for the strategic in teractio n s a m o n g firms
o fferin g co m p etin g produ cts. In general, the theoretical literatu re o f w h ich these
pap ers are an exam p le identifies m u ltip le possible eq u ilib ria a m o n g firm s com peting
in such en viron m en ts, so that it is d ifficu lt to draw firm co n clu sion s.
 I N N O V A T I O N AND D IFFU S IO N

Farrell and Saloner study the speed o f diffusion (relative to the socially optimal rate)
when the good in question is subject to network externalities, so that early adopters
ignore the consequence o f their adoption on future adopters and on the users of the
previous technology. They show that in this setting, diffusion can be either too fast
(excess momentum) or too slow (excess inertia). Finally, the book by Shapiro and
Varian draws out the implications o f these various theoretical models for the produc­
tion and marketing o f information goods (broadly defined), many of which exhibit
the properties that give rise to network externalities. They describe strategies for
competing in markets where network externalities are important and where it is
important to win standards battles because losing them means business failure.

17.4,3 Costs o f Adopting the New Technology

The second main class o f factors affecting the decision to adopt new technology are
those related to its cost. This includes not only the price o f acquisition, but more
importantly the cost o f the complementary investment and learning required to
make use o f the technology. Such investm ent m ay include training o f workers and
the purchase o f necessary capital equipm ent (whose diffusion is therefore affected
by the sam e factors). It is difficult to overem phasize the im portance o f this point
about the need fo r co m p lem en tary investm ent, especially for com plex m odern
technology that requires the reorganization o f the process that will use it (see Ch. 5
by Lam in this volu m e, for m ore on this topic).
For exam ple, in a series o f recent papers Eric Brynjolfsson has argued that the full
cost o f adopting n ew com pu ter inform ation systems based on networked personal
com puters is about ten tim es the cost o f the hardw are.11 G reenan and Guellec (1998)
use data on French firm s and w orkers to m ake a sim ilar point, that the effective
adoption o f IC T requires organizational change as well, and that this raises the cost
o f adoption, w hich slow s diffusion. Caselli and C olem an (2001) com pare the rates o f
com puter investm ent across O E C D countries between 1970 and 1990 and highlight
the importance both o f w ork er skill level and o f com plem entary capital investments
in determ ining the rate o f purchase o f new com puting systems. The im plication o f
this w ork is that the use o f new com pu tin g technology requires both the training
o f workers and the installation o f related equipm ent (for exam ple, rem odeling
expenses for space to install servers, along w ith the necessary cooling equipm ent).
The need fo r co m p lem en tary investm ent therefore has tw o effects: it slows diffusion
because it raises the cost, and because this type o f investm ent usually takes tim e, it
slows d ow n the rate at w hich the benefits o f the new technology are seen by the firm
and the econ om y in the fo rm o f increased productivity.
D avid (1990b) has argu ed that a sim ilar adjustm ent took place in m anufacturing
industry use o f electric pow er, w hich took 40 years to diffuse com pletely in the
474 BRONW YN H . HALL

U n ited States (also see Figu re 17 .1 and M o w e ry an d R o sen b erg 1998)* T h e installation
o f electric p o w er in a facto ry requ ired a com plete red esign o f its la y o u t an d a change
in task allo catio n , w h ich m eant that ad o p tin g this n ew te c h n o lo g y w as a rath er costly
process, an d tended to o ccu r slow ly, o r w hen green field in vestm en t w as being
undertaken* D a v id argu es that a sim ilar reo rgan izatio n o f w o rk flo w takes place
w h en co m p u ter tech n o lo g y is in tro d u ced into the w o rk p lace o r w h en Internet-
b ased pro cessin g replaces telephone o r m ail o rd er p ro cessin g. R ecent p ro d u ctivity
grow th evidence in the U n ited States app ears to co n firm the v ie w that m ajor
tech n o lo gical-o rg an izatio n al change takes tim e fo r its effects to be felt (G o rd o n
2003; Econom ist 2003).
Shaw (2002) has d o cu m en ted this k in d o f p h e n o m en o n in the replacem ent o f
m an u al m o n ito rin g o f p ro d u ctio n lines in co n tin u o u s h ot steel p ro d u c tio n lines by
co m p u terized p u lp it o p eratio n . N o t o n ly does this in vo lve a su b stan tial investm ent
in h igh tech n o lo g y equ ip m en t, b u t it also requ ires few er w o rk ers w ith substantially
h igh er cogn itive skills. W here th ey used to be on the p ro d u c tio n lin e w orkin g
ph ysically w ith the m ach in ery, they are n o w in sm all ro o m s (“ p u lp its” ) above the
line, m o n ito rin g and ad ju stin g the process u sin g co m p u te r tech nology.
T ech n o logy p ro d u cers o ften try to subsidize the a d o p tio n o f n ew tech n ologies by
p ro v id in g free train in g and o th er help to (poten tial) users an d b y ch arg in g reduced
in tro d u c to ry rates fo r a certain p erio d . A n o th er sy m p to m o f the desire o f innovating
firm s to reim b u rse new cu stom ers for their su n k costs in p re v io u s tech n o logies is the
w id esp read practice am o n g softw are firm s o f o fferin g c o m p e titive u pgrades to
ow n ers o f rival p ro d u cts as well as to the ow n ers o f th eir o w n p ro d u cts. For a
m o re com plete d iscu ssion o f strategies used b y tech n o lo g y p ro d u ce rs to encourage
d iffu sio n and increase the installed base o f their p ro d u ct, see S h a p iro an d Varian
(1999).
Because m ost o f the costs o f a d o p tio n are fixed, firm s' ch oices to change or
in tro d u ce tech nologies m a y be in flu en ced by their o w n scale an d b y the m arket
structu re o f the in d u stry w ith in w h ich they operate. A n in terestin g exam p le o f this
p h en o m en o n is given b y Paul D av id in a series o f pap ers on the in tro d u c tio n o f the
m ech an ical reaper in U S and B ritish agricu ltu re in the n in eteen th cen tu ry (D avid
1975a and b). H e argues p ersu asively that a d o p tio n w as delayed in B rita in relative to
the U nited States fo r tw o reasons: first, because the reaper w as a fixed cost invest­
m ent, profi tab ility required a farm and fields o f a certain size; secon d, because it was
i n co m p atib le w ith the typ ically B ritish pattern o f sm all fields d iv id ed b y hedgerow s.
In ad d itio n to the difference betw een cou ntries, he also fin d s that d iffu sio n was
delayed in the U S itself until the price o f lab o r rose to a level that m ad e the invest­
m ent in the reaper (a lab o r-savin g device) profitab le.
In the p resen t-d ay context, a sim ilar em p irical fin d in g can be fo u n d in m any
studies o f d iffu sio n , M a ju m d a r and V enkataram an (1998) lo o k ed at the replacem ent
o f m echan ical sw itch in g by electronic sw itch in g in the U S telecom m u n icatio n s
industry and fo u n d that larger firm s ad o p ted first, p re su m a b ly b ecau se the costs
 I N N O V A T I O N AND D I F F US I ON 475

per customer were somewhat lower. Note that even when technology adoption
involves an investment in equipment that is proportional to the existing size o f the
firm, the requirement that the firm have sufficient absorptive capacity, and the need
for worker training or other complementary changes may create a fixed cost that is
not proportional to firm size.
As in the case o f investment in innovation, firm investment in new technologies is
also sensitive to financial factors. As was suggested earlier, the decision to adopt new
technology is fundamentally an investment decision made in an uncertain environ­
ment, and therefore we should not be surprised to find that all the arguments for a
relationship between sources o f finance and choice o f investment strategy that have
been advanced in the investment literature have a role to play here. Chapter 9 by
O'Sullivan in this volume reviews these financial factors in some detail. For example,
Mansfield (1968) reports that the adoption o f diesel locomotives by railways depends
somewhat on their liquidity, implying these firms faced a higher cost o f external than
internal finance.

1 7 4 4 Information and Uncertainty

The choice to adopt a new technology requires knowledge that it exists and some
information about its suitability to the potential adopter s situation. Therefore an
important determinant o f diffusion is information about the new technology, which
may be influenced by the actions o f the supplier o f the new technology. Obviously in
many cases this takes the form o f advertising, which influences the cost o f the new
technology directly. The choice to adopt may also depend on the information
available about experience with the technology in the decision maker's immediate
environment, either from those in geographic proximity or from those with whom
he or she interacts.
Because benefits for adoption are spread over time while costs are usually incurred
at the beginning, expectations about the length o f life o f either the technology or the
adopter will matter. Uncertainty about benefits, costs, or length o f life will slow the
rate o f adoption, and may often turn the decision problem into an options-like
computation. As discussed earlier, the latter is a consequence o f the fact that in most
cases, once a new technology has been chosen, the costs are sunk and cannot be
recovered. That is, the potential adopter has an option on new technology; if he sees
the uncertain payoff reach a certain value (the strike price), he will exercise the option
by adopting the technology (see Stoneman 2001& for a theoretical development).
Empirical work on diffusion that incorporates real options is rather scarce,
although descriptive work that confirms the role o f trialability and observability is
widespread (for some recent examples, see Nelson et al. 2002). One notable example o f
an investigation o f technology adoption as the exercise o f an option is that of Tuque
4 7б B R O N W Y N Н. H A LL

(2002). She looks at the decision by US plants to adopt three advanced manufacturing
technologies, and finds that plants operating in industries with lower degrees o f
demand and technological uncertainty and a thicker resale market (higher resale
prices for used machinery) are more likely to adopt these technologies. She argues
that this confirms the importance o f uncertainty in the decision; if adopting a new
technology corresponds to the exercise o f an option, we expect adoption to happen
more often in industries with lower uncertainty and lower sunk costs.

17.4.5 Market Size, Industry Environment,

and Market Structure
The relationship between firm size or industry concentration and the adoption of
new technology by a firm is subject to m any o f the same considerations as the
relationship o f these factors to innovation. As discussed above, large dom inant firms
can spread the costs o f adoption over m ore units, but also m ay not feel the pressure
to reduce costs that leads to investment in new technologies. Empirically, in the case
o f technology adoption, most studies have found that large firms adopt any given
technology sooner, but there are some exceptions. Oster (1982) found that small
firms in the steel industry replaced the open hearth furnace with the basic oxygen
furnace during the post-World War II period sooner than large firms. In a study of
twelve m ajor innovations in the coal, rail, iron and steel, and brewing industries,
Mansfield (1961) found weak evidence that firms in competitive, less concentrated
industries adopted new technologies sooner, as did Rom eo (1977) in a study o f the
diffusion o f numerically controlled machine tools.
In some cases the adoption o f new technology is determined by firm s, acting for
the benefit o f consumers and for their own benefit. As an example, consider airline
adoption o f computerized reservation systems. Consum ers have little say in this
decision although they ultimately benefit in the form o f lower prices for air travel or
better service, such as seat reservations. In other cases, the decision fundamentally
rests with the consumer, for example the choice o f video recording technology such
as V H S, Beta, and now DVD. Although the same considerations o f cost versus
benefit apply broadly in both cases, the role o f market structure m ay be more
im portant in the form er case than in the latter, because the adopting firm s are likely
to be few in number and therefore able to interact strategically with respect to the
adoption decision itself. In the latter case, the strategic interaction occurs in choos­
ing the technologies that are offered; in principle, firms can produce the same set o f
strategic outcomes as in the form er case (via penetration pricing, etc.), but lack of
perfect information about consum ers’ tastes and limits on their ability to segment
the market sufficiently may prevent the firms from fully internalizing consumers’
preferences.
 IN N O V A T IO N AND DIFFUSION

Market structure can affect tlie decision to adopt in two distinct way si via seller
behavior and via buyer behavior. Highly concentrated providers of new technology
will tend to have higher prices, slowing adoption, but they also have the ability to
determine a standard more easily, increasing the benefit o f adoption. If two or more
oligopolistic firms are competing to offer different standards, we may in fact get too
rapid adoption o f a new technology, because o f the incentives they face to price below
cost in order to build market share (Farrell and Saloner 1992). In the case o f potential
adopting firms, market concentration affects both their ability to pass through any
costs to consumers and also the incentives they face in incurring the costs o f adoption.
Many of the issues raised by the tension between the fear of displacement and the
exercise of market power here are familiar from the literature on monopolists'
incentives to innovate (for example, see Gilbert and Newberry 1982).
Along with market size and structure, the general regulatory environment will have
an influence, tending to slow the rate o f adoption in some areas due to the relative
sluggishness o f regulatory change and increasing it in others due to the role of the
regulator in mandating a particular technological standard. As an example o f
the former situation consider the use o f plastic pipe for plumbing, which lowers
construction cost, but has been slow to diffuse in many localities due to existing
building codes. As an example o f the latter, Mowery and Rosenberg (1982) have
written about the extent to which airline regulation by the Civil Aeronautics Board
in the United States was responsible for promoting the adoption o f new innovation in
airframes and jet engines, in its role as standard setter and coordinator for the industry.
An important example o f the unintended consequences o f regulation for diffu­
sion is the difference between the United States and Europe (and Japan to some
extent) in the diffusion o f household Internet use. Historically, pricing in the US
telecommunications industry has permitted unlimited local calling at a single
monthly rate, whereas pricing for local calls in other countries has usually been
proportional to usage. These policies are largely determined by regulatory bodies,
but once in place, are difficult to change because consumers and firms adapt to them.
In the absence o f direct connection to the network such as is available in large
institutions, household Internet use requires the ability to connect over local phone
lines for extended periods o f time. The marginal cost o f the Internet for households
is therefore to a great extent determined by the cost o f local calling, so diffusion o f the
Internet along with email and instant messaging use has been far more rapid in the
United States than in other countries that are just as developed. Only with the recent
advent o f ISDN service charged by the month in some European countries has
household Internet use begun to spread there. In contrast, the diffusion o f various
“ text-messaging” services on wireless phones, which are a form o f communication
popular with teenagers and similar to the Internet instant messaging widely used in
the US, has been more rapid in Japan and Europe. Relative costs o f the two forms o f
instant com m un ication , w hich in turn are due to regulatory reasons, are probably
the m ain explanation fo r the differences.
478 B R O N W Y N H. H A LL

17.5 C u l t u r a l a n d S o c i a l D e t e r m i n a n t s

Econom ic factors like these can go a long way toward explaining differences in rates
o f diffusion (Griliches 1957 and subsequent authors) but other factors m ay also be
important; For example, m any have stressed differences in cultural attitudes towards
risk and simple ‘"newness.” 12 These characteristics can vary within cultures as well as
between them, leading to dispersion in adoption rates that are not accounted for by
the economic variables. Am ong others, Strang and Soule (1998) provide a useful
discussion o f the cultural basis o f diffusion.
Rogers (1995) cites a number o f situations where compatibility with existing social
norms has strongly influenced the adoption o f health-related innovations such as
the boiling o f water for consumption or various types o f contraceptives in under­
developed countries, whose relative popularity depends greatly on local religious
and cultural mores. He cites as example an instance where a strong traditional
distinction made between the qualities of cold and hot water discouraged the use
of the very simple preventive measure of boiling water destined for human con­
sumption in order to prevent diarrheal diseases.
On the other hand, for consumer household durables, Tellis et al. (2002) find that
variables such as gender, cultural attitudes, religion, etc. have little predictive power
for "‘take-off’ on average (across European countries) in the presence o f lagged
market penetration. When these variables are considered separately as predictors,
""industriousness” (which is measured by a climate variable) and ""need for achieve­
m ent” ( which is measured by the ratio o f Protestants to Catholics in the country)
speeds diffusion, and a measure o f ""uncertainty avoidance” slows diffusion. This
study is noteworthy in that it includes economic, cultural, and communication
variables jointly in the same predictive equation.

17.6 C o n c l u s i o n s

Traditionally, diffusion is one o f the three pillars on which the successful introduc­
tion o f new products, processes, and practices into society rests, along with inven­
tion (a new idea) and commercialization/innovation (reducing the invention to
practice). In some ways it is the easiest part o f the process to study, because it is
m ore predictable from observable factors than the other two. Certainly countless
studies o f the diffusion o f individual innovations exist, and even exhibit some
commonalities (see the references in this chapter and in Rogers 1995), such as
 INN OV ATION AND DIFFUSION 479

the familiar s-shaped curve, and the importance o f both economic factors and social
networks,
Although many have criticized the linear model that lies behind the division of
innovative activity into three parts as oversimplified, it remains true that without
invention it would be difficult to have any thing to diffuse, so that the model still
serves us as an organizing principle, even if we need to be aware of its limitations.
Nevertheless, an important insight from the many historical case studies of individ­
ual inventions has been the extent to which the diffusion process enhances an
innovation via the feedback of information about its operation or utility under
varying conditions and across different users, information that can be used to
improve it. A second major finding from this literature has been the possible
feedback from differences in the rate or scale of adoption across geographic areas
to the rates of improvement in the innovation.
In the introduction to this chapter, Rosenberg’s observations on the slowness and
variability o f the diffusion o f different innovations were cited. The studies reviewed
in this chapter have identified some explanations for these observations, such as the
size of sunk costs (trialability), the adaptations and improvement necessary to make
the invention useful after its initial conception, and the inherent slowness o f
interpersonal communication networks in spreading information. In the case of
major innovations such as electricity or the computer, some studies have empha­
sized that the necessi ty o f reorganizing the workplace to take advantage of the new
innovation means that diffusion will be greatly delayed, and also that the expected
gains from innovation may take time to be realized.
Several areas stand out as potentially fruitful for future research. First, most of the
studies conducted to date have been methodologically rather simple; the most
ambitious have used a hazard model to correlate the time until adoption with
various characteristics o f the innovation and the adopter (depending on the par­
ticular dataset). There is room for an approach that is more structural and grounded
in the choice problem actually faced by the adopter. One promising avenue for
modeling is the real options approach suggested by Stoneman (2001b); such a model
would yield a hazard or waiting time model rather naturally, while explicitly
incorporating the effects o f uncertainty on the decision.13 The cumulative distribu­
tion for adoption derived from a hazard model has the familiar S-shape.
Second, although many studies have described the process o f innovation
enhancement during its diffusion qualitatively, there has been relatively little
systematic collection o f data or modeling o f the process. Investigations of this type
would be very helpful in quantifying the importance o f this effect, which is similar to
but not the same as the well-known learning curve. One technological area where
this process has been very important and might be worth study is the area o f user-
driven software developm ent. '
Finally, an area o f research that is receiving increasing attention in a globalizing
economy is that concerned with international technology transfer.14 This literature
4$0 B R O N W Y N H . HALL

is generally positive (as opposed to normative) in approach and empirically based,

focused on identifying the mechanisms through which technology diffuses from
m o re d evelo p ed to less d eveloped co u n tries rather than o n the a d o p tio n ch oice itself.
T h at is, this analysis is co n d u cted at the aggregate level rath er th an at the level o f an
in d iv id u a l d ecision m aker. It is p ro b ab ly safe to say that there is ro o m fo r further
research in this area, as the diffusion o f tech n o lo g y is an im p o rta n t source o f
e co n o m ic an d so cial d evelopm ent. Indeed, fro m a w elfare persp ective, o n e o f the
m ost im p o rtan t areas for fu rth er stu d y is the co m p a rative d iffu sio n o f v a rio u s health
and m ed ical practices across d evelo p in g cou n tries, esp ecially becau se it is apparent
that there are w id e variatio n s even am o n g sim ilar lo w in co m e co u n tries in rates o f
ad o p tio n .

N o tes

1. University of California at Berkeley, Scuola Sant’anna Superiore Pisa, NBER, and the
Institute of Fiscal Studies, London. I am grateful to Beethika Khan for contributing some
of the literature review that lies behind the issues discussed in this paper, and other
contributors to this volume, especially my discussants, Kristine Bruland, John Cantwell,
and Ove Granstrand, for their very helpful comments. Finally I owe an immense debt to
the editors for their careful reading of multiple drafts of this chapter,
2. I am grateful to Chris E. Hall for calling this example to my attention. It is described in
McGrew (1998), where a more complete set of references to the anthropological literature
is given. A third feature of this example, perhaps not directly relevant to this chapter, may
be noted: the fact that once having innovated, innovators tend to innovate again.
3. As discussed in the introduction to this volume, the view that every adopter develops and
adapts an invention to his own use has led so me of the literature to refer to adoption itself
as “ innovation.” I will follow the more conventional practice of reserving the term
innovation for the first “public” use of a new product, process, or practice.
4. See Godinho and Fagerberg (Ch. 19 in this volume) on the role of adoption of new
technology in the catch-up process and in long run economic growth.
5. I am grateful to Paul David for calling some of these references to my attention.
6. Although see Rogers (1995) for some examples of innovations that failed to diffuse because
they were rejected after trial.
7. An option is a choice between doing nothing and paying a fixed amount to purchase an
uncertain return. It is real (as opposed to financial) if it involves investment in real assets.
In this setting, the investment is the adoption of a new technology, which has uncertain
benefits and costs that may change over time. The option value arises from the fact that
waiting may reduce the chance that the wrong decision is made.
8. For a good presentation of this class of models and their extensions, see Geroski (2000).
David (2003) provides an evolutionary interpretation of this mechanism.
9. For any particular parametric distribution function, this point might be defined at the
point where the curvature of the cumulative distribution (the second derivative) is
 INNOVATION AND DIFFUSION 4&l

maximized. Such з point is well defined if it exists. It occurs when shout 20 per cent o f the
population has adopted in the case o f a logit and when about 15 per cent have adopted in
the case o f a norm al. C o ld er and Tellis (1997) define a non~parametric discrete version o f
this measure by looking at the current rate o f adoption as a share o f adoption to date.
10. On the adoption o f A TM systems, see also Hannan and M cDowell (1984a and b), who
emphasize the role o f bank size and industry concentration, which are chiefly cost side
and market structure considerations.
11. See Brynjolfsson (2000) for a su m m ary o f this work and further references.
12. For a discussion o f various cultural explanations, see M okyr (1990).
13. In the labor econom ics literature, Lancaster and Nickell (1980) developed a sim ilar
model for the probability o f obtaining a job when unem ployed (see also Lancaster 1990).
14. See Keller 2001 for a review o f this literature.

R e f e r e n c e s

Arthur , W. B. (1989), “ C om petin g Technologies, Increasing Returns, and Lock-in by

Historical Events,” T h e E co n o m ic Jo u r n a l 99 (M arch): 116 -3 1.
------ Erm oliev , Y., and K a n io v sk i , Y. (1983), “ Generalized Urn Problem and Its A pplica­
tions,” C ybern etics 19: 6 1-7 1.
Babcock , J. M , (1962), “A doption o f H ybrid C orn — A Com m ent,” R u ra l Sociology 27:332-8.
Bass , F, M. (1969), “A N ew Product G row th M odel for Consum er Durables,” M a n a g e m en t
Science 13(5): 215-27.
B a s s a n i n i , A., and D o si, G. (1998), “ Heterogeneous Agents, Com plem entarities, and
Diffusion. Do Increasing Returns Im ply Convergence to International Technological
M onopolies?” in D. D, G atti, M . Gallegati, and A. Kirm an (eds.), M a rk e t Structure,
A ggregation , a n d H etero g en eity , Berlin: Springer, 163-85.
B r u l a n d , K. (1998), “ Skills, Learning and the International D iffusion o f Technology,” in
M. Berg and K. Bruland (eds.), Techn ological R evo lu tio n s in E u ro p e , Cheltenham : Edward
Elgar, 4 5 -6 9 .
------ (2002), B ritish T echn ology a n d E u ro p ea n In d u stria liz a tio n : T h e N o rw eg ia n T extile In d u s ­
try in the M id -n in e te e n th C en tu ry, O xford: O xford University Press.
B r y n j o l f s s o n , E. (2000), “ Beyond C om putation: Inform ation Technology, Organizational
Transform ation and Business Perform ance,” Jo u r n a l o f E co n o m ic Perspectives 14: 23-48,
C aselli , F., and C o lem an , W, II (2001), “ Cross-country Technology D iffusion: The Case o f
Com puters,” A m e ric a n E c o n o m ic R e v ie w 9 1(2): 328-35.
David , P. A. (1975a), “ The M echanization o f Reaping in the Ante-bellum M idwest,” in
P. A. D avid, T echn ical C h oice, In n o v a tio n , a n d E co n o m ic G ro w th , Cam bridge: Cam bridge
University Press, 195-232.
------ (1975b), “ The Landscape and the M achine: Technical Interrelatedness, Land Tenure,
and the M echanization o f the C orn H arvest in Victorian Britain, in P. A. D avid, Technical
Choice, In n o v a tio n , a n d E co n o m ic G ro w th . Cam bridge: Cam bridge University Press,
233- 90 .
D a v i d , P. A. (1985), “ C lio and the Econom ics o f QW ERTY,” A m eric a n E co n o m ic R e v ie w 75:

ЗЗ2-7*
482 B R O N W Y N H. H A L L

------- (1990), “A t last, a R em edy for C h ro n ic Q W E R T Y -S k ep ticism !” O xfo rd : A ll Souls

College, Oxford University. Manuscript.
-------(19 9 0a), “ H eroes, H erds, and H ysteresis in T echnological H isto ry : T h o m as Ed ison and
the Battle o f the System s R econsidered,” I n d u s t r ia l a n d C o r p o r a te C h a n g e 1 ( 1 ) : 129-80.
*(i9 9 o b ), “ T h e D yn am o and the C o m p u ter: A n H istorical P erspective o n the M o d em
P ro d u ctivity Paradox,” A m e r ic a n E c o n o m ic R e v ie w 80: 355-6 1.
-------(19 9 9 ),“ 'M yth ' - in fo rm in g the public about the p u b lic good s an d Q W E R T Y ” (a reply
to “ Lock and K ey” ), E co n o m ic Focus, The E c o n o m is t , 18 Septem ber.
-------(2003), “ Z v i on D iffu sion , Lags an d P ro d u ctivity G r o w t h . . . C o n n ectin g the D o ts”
Paper presented at the C on feren ce o n R8cD, Ed ucation and P ro d u c tivity held in m em ory
o f Z v i G riliches in Paris, A ugust 2003.
D a v ie s , S. (1979), T h e D iffu s io n o f Process In n o v a t io n , C am b rid ge: C am b rid ge U niversity
Press.
D ix it , A ., and P in d y c k , R. (1994), In v e s tm e n t u n d e r U n c e rta in ty > Princeton : Princeton
U n iversity Press.
E co n o m id es , N ., and H im m e l b e r g , C . (1995), “ C ritical M ass and N etw o rk Size with
A p p licatio n to the U S . Fax M arket,” N ew York U niversity, Sa lo m o n Brothers W orking
Paper S/95/26 (A ugust).
E c o n o m is t (2003), “ The N ew "New Econom y',” 11 Septem ber.
F a r r e l l , J., and S a lo n er , G . (1992), “ Installed Base and C o m p atib ility: In n o vatio n , Product
Preannouncem ents, and Predation,” A m e r ic a n E c o n o m ic R e v ie w 7 6 : 9 40-55.
*G e r o sk i , P. A. (2000), “ M odels o f Technology D iffu sion ,” R ese a rc h P o lic y 29(4/5): 603-25.
G il b e r t , R. J., and N e w b e r r y , D. M . G. (1982), “ P reem ptive Paten tin g and the Persistence o f
M o n op oly,” A m e r ic a n E c o n o m ic Review 72(3): 514-26.
G il f il l a n , S. C. (1935a), In v e n tin g the S h ip : A S t u d y o f th e In v e n tio n s m a d e in h e r H isto ry
b etw een F lo a t in g L o g a n d R o to rsh ip y C hicago: Follett.
-------(i935 b )y T h e S o c io lo g y o f In v e n tio n : A n E ssa y in the S o c ia l C a u s e s o f T ech n ic In v e n tio n a n d
S o m e o f its S o c ia l R esu lts; E s p e c ia lly as D e m o n s tra te d in the H is to r y o f th e S h ip y Chicago:
Follett.
C o ld er , P. N ., and T e l l is , G . J. (1997), “ W ill It Ever Fly? M o d elin g the T ak eo ff o f Really New
C o n su m er D urables,” M a r k e t in g S c ie n c e 16(3): 256-70.
G ordon , R, J. (2003), “ Five Puzzles in the B eh avio r o f P ro d u ctivity, Investm ent, and
Inn ovation , in G lo b a l C o m p etitiv e n ess R e p o rt , 2003—2004, W orld E co n o m ic Forum .
G r e e n a n , N ., and G u e l l e c , D. (1998), “ Firm O rganization, Technology, an d Perform ance:
A n E m p irical Study, E co n o m ic s o f In n o v a tio n a n d N e w T ech n o lo g y 6: 313—47.
^G r il ic h e s , Z , (1957), ""Hybrid C orn : A n E xp lo ratio n in the E co n o m ics o f Technological
Change,” E c o n o m e tric a 25: 50 1-22.
(1960a), H ybrid C o rn and the E conom ics o f In n o vatio n ,” S c ie n c e 132: 275—80.
(1960b), C on gruen ce Versus P rofitability: A False D ich otom y,” R u r a l S o cio lo g y 25:
354~ 6 .
“ (1962), Profitability Versus Interaction: A n oth er False D ich otom y,” R u r a l S o cio lo g y 27:
325-30.
H a n n a n , 1 ., and M c D o w ell , J. (1984a), I he D eterm inants o f T echn ology A doption : The
Case o f the Banking Firm , R a n d Jo u r n a l o f E co n o m ic s 15(3): 328—35.*

* Asterisked items are suggestions for further reading.

 -------------(1984b), "M ark et C oncentration and the D iffusion o f New Technology in the
Banking Industry,” The Review o f Economics and Statistics 66(4): 686-91.
Havens , E. A., and R o g ers , E. M. (1961), "P rofitability and the Interaction Effect,” Rural
Sociology 26: 409-14.
Jovanovic, R., and Sto lyaro v , D. (2000), O ptim al Adoption o f Com plem entary Tech­
nologies,” American Economic Review 90(1): 15-29,
Katz, M. L., and S h a pir o , C. (1985), "N etw o rk Externalities, Com petition, and C om pati­
bility,” American Economic Review 7 5(3); 424-40.
------------ (1986), "T echn ology A doption in the Presence o f Network Externalities,” Journal
of Political Economy 94; 822-41.
* ---------- (1994), "System s C om petition and N etw ork Effects,” Journal o f Economic Per­
spectives 77: 9 3-115.
Kawai, M ., W a t a n a r e , K., and M o r i , A. (1992), "Pre-Cultural Behaviors Observed in Free-
Ranging Japanese M onkeys on Koshima Islet over the Past 25 Years,” Prim , Rep. 32:143-53,
Keller , W. (2001), "In tern ation al Technology Transfer,” N B E R W orking Paper N um ber
W8573-
Lancaster , T. (1990), The Economic Analysis o f Transition Data, Cam bridge: Cam bridge
University Press.
------ and N ic k e l l , S. (1980), "T h e Analysis o f Reem ploym ent Probabilities for the
Unemployed,” Journal o f the Royal Statistical Society A 143(2): 141-6 5,
Liebowitz, S. J., and M argolis , S. E. (1990), "T h e Fable o f the Keys,” Journal of Law and
Economics 3 3 :1-2 6 .
Luq ue , A. (2002), "A n O ption-Value A pproach to Technology Adoption in U.S. M anufac­
turing: Evidence from M ic ro d a ta ” Economics o f Innovation and New Technology 11(6):
543-68.
McG rew , W. C. (1998), "C u ltu re in N onhum an Primates?” Annual Review o f Anthropology
27:301-28.
Majum dar , S., and V en k a ta r a m a n , S. (1998), "N etw ork Effects and the Adoption o f New
Technology: Evidence from the U.S. Telecom m unications Industry,” Strategic Manage­
ment Journal 19:1045-62.
Mansfield , E. (1961), "Technical Change and the Rate o f Im itation,” E co n om etrica 29(4):
741-66.
* ------ (1968), Industrial Research and Tech n ological In n o v a tio n , New York: Norton.
*M okyr, J. (1990), The Lever o f Riches. O xford: O xford University Press.
M owery , D., and R o sen berg , N . (1998), Paths o f Innovation, Technological Change in 20th-
Century America, Cam bridge: Cam bridge University Press.
--------------(1982), "G o vern m en t Policy and Innovation in the Com m ercial Aircraft Industry,
1925-75,” in R. R. N elson (ed.), Government and Technical Progress: A Cross-Industry
Analysis, O xford: Pergam on Press.
N elson , R. R „ P eterh ansl , A ., and Sa m pa t , B. N . (2002), "W hy and H ow Innovations Get
Adopted: A Tale o f Fou r M odels,” N ew York: C olum bia University (Photocopied).
Oster , S, M . (1982), "T h e D iffusion o f Innovation am ong Steel Firm s: The Basic Oxygen
Furnace,” Bell Journal o f Economics 13(1): 45-56.
Park, S. (2002), "Quantitative Analysis o f Network Externalities in Competing Technolo­
gies,” New York: SUNY at Stony Brook. Photocopied.
^R o g e r s , E, M . (1995), Diffusion o f Innovations, 4th edn., New York: The Free Press.
Ф 4 B R 0 N W Y N Н * HALE

----- and H a v e n s , А. Е. (1962), "Rejoinder to Griliches* 'Another False Dfcbotomy;*” Jfanl

Sociology 2 7 :3 3 2 -4 .
R o m e o * A . A . (1977)* "T h e Rate o f Im itatio n o f a C ap ital-em b o d ied P rocess In n o v a tio n s
Economica 44: 6 3-9 .
^Rosenberg , N. (1972), "Factors A ffectin g the Diffusion o f Technology" E x p lo ra tio n s in
Economic History 10(1): 3 -33 .
------ (1982)* "L e arn in g b y Using,” in N. R osen berg, Inside the Black Box, C am b rid g e: C am ­
b rid ge University Press, 120-40.
Salo ner , G., and S h epard , A. (1995), "Adoption of Technologies with Network Effects: An
Empirical Examination o f the Adoption o f Automated Teller Machines,” Rand Journal o f
Economics 26(3): 479-501.
S hapiro , C , and V a r ia n , H. (1999), Information Rules, Boston: Harvard Business School
Press.
S h aw , K. (2002), "B y What Means Does Information Technology Affect Employment and
W a g e s" in Greenan, N., Y. D ’Horty, and J. M airesse, (ed$.)> Productivity, Inequality, and the
Digital Economy. Cambridge, Mass.: The M IT Press, 229-68.
*S to n em an , P. (2001a), The Economics o f Technological Diffusion, Oxford: Blackwells.
------ (2001b), "Financial Factors and the Inter Firm Diffusion o f New Technology: A Real
Options Model,” University o f Warwick EIFC Working Paper No. 20 0 1-0 8 (December).
*S tr a n g , D., and Soule , S. A. (1998), "D iffu sio n in Organizations and Social Movements,”
Annual Review o f Sociology 24: 265-90.
T ellis , G. J., S trem ersch , $., and Y in , E. (2002), "The International Takeoff o f New
Products: The Role o f Economics, Culture, and Country Innovativeness,” Marketing
Science 22(2): 18 8 -208 .
W en d t , O., and v a n W estar p , F. (2000), "Determinants of Diffusion in Network Effect
Markets,” Paper presented at the 2000 IRM A International Conference, Anchorage,
A laska.
Z ett elm ever , F., and Sto n em an , P. L. (1993 k "Testing Alternative M o d els o f New Product
Diffusion,” Economics o f Innovation and New Technology 2: 283-308.
 PART IV

INNOVATION AND
PERFORMANCE
 In t r o d u c t i o n to P a r t IV

T h e literature on the relation ship betw een in n o v a tio n an d eco ­

n o m ic p erfo rm an ce has been d o m in ated b y eco n o m ists, an d the
first ch ap ter in this section focuses on an issue that h as received
m o re attention than any o th er in this area, the relatio n sh ip betw een
in n o vatio n and econ o m ic grow th (C h ap ter 18 b y V ersp agen ). A s
the au th o r show s, econ o m ists have ad o p ted d ifferen t fram ew o rk s
for the analysis o f this relation sh ip; the tw o m o st relevant
ap p ro ach es b ein g “ e v o lu tio n a ry ” an d “ n ew g ro w th ” theory.
These ap p ro ach es d iffer less in their view s o f the im p o rta n c e o f
in n o vatio n for grow th , w hich is ackn ow led ged in b o th , th an o n the
precise m ech an ism s th ro u g h w h ich in n o va tio n affects gro w th. O ne
im p o rtan t w ay th ro u gh w h ich in n o va tio n sp u rs gro w th is th ro u g h
the d iffu sio n o f tech n o lo g y fro m the d evelo p ed to the less-d e­
veloped w o rld (so called “ late-co m ers” ). F agerb erg a n d G o d in h o
(C h ap ter 19) p ro vid e a h istorical and in terp retative su rv e y o f the
literature on catch in g-u p b y “ late-co m ers”, w ith sp ecial fo cu s o n
the role played b y in n o va tio n fo r the o u tco m e o f such processes. A
related issue is that o f ch an gin g patterns o f co m p etitiven ess, and
the role o f in n o va tio n in this context. C an tw ell (C h a p te r 20)
exam in es the literatu re on in n o va tio n an d co m p etitiven ess. A l­
th o u gh the com petitiveness issue has sp aw n ed n u m e ro u s c o n tro ­
versies since the 1980s, a perhap s even m o re h o tly d eb ated issue,
p a rticu larly w ith in E u rop e, concerns the e m p lo y m e n t effects o f
in n o vatio n . Pianta (C h ap ter 21) p ro vid es an exten sive su rv e y o f the
large em p irical literature on this subject. T h e sectio n — an d the
entire vo lu m e— concludes w ith C h ap ter 22 b y L u n d v all an d B o rra s
on science, technology, and in n o vatio n policies.
 C H A P T E R l8

I N N OV A T I ON
AND ECONOMIC
GR OWT H

BART VERSPAGEN

18.1 I n t r o d u c t i o n

According to Maddison (2001), the world economy began to grow in roughly the
year ad 1000 following a long period o f stagnation (Figure 18.1).1 A marked increase
in growth rates occurred in the period 1600—1800, and, as the figure suggests, growth
has been increasing ever since.
Econom ic h isto ry addresses the issue o f the w ay in w hich the record o f econom ic
growth in Figure 18.1 is related to historical developm ents such as the Industrial
Revolution (see also B ru lan d and M ow ery, Ch. 13 in this volum e). A lthough the exact
impact that this has had on growth rates, particularly at the sectoral level, remains a
subject o f debate (e.g. Crafts 1985), it seems beyond dispute that a change ot
technology in the pure sense, coupled with organizational changes at various eve s
of aggregation, are the main driving factors behind the continuous increase о lving
standards entailed b y this process.
The historical perspective also nicely illustrates the fact that there is m uc m ore о
econom ic grow th than ju st the data on per capita incom e that are s o wi e У use У
economists, E co n o m ic grow th is an historical process o f structur c ange m
 time (AD)

...— Europe — — USA *----------China —— A frica - - - - W o rld

Fig. 18.1 Long-run growth in the world economy, according to the

data in Maddison (2001)

broadest sense, and only the most elementary aspects o f this process can be meas­
ured by the data on production and income. The form o f structural change most
visible in the statistics is the changing sectoral m ix o f the economy. Chenery,
Syrquin, and Robinson (1986) have illustrated the regularity between the changing
sectoral com position o f the econom y and the increasing level o f productivity, while a
“ deeper” manifestation o f structural change in the long-run process o f economic
growth remains largely the dom ain o f historical research.
Although the argument that technological and organizational innovation are
responsible for this lengthy period o f gradually accelerating growth is appealing,
in fact econom ic theories explaining any such relationship are far from straightfor­
ward. Growth theory, especially when focused on the issue o f technology, is a field
characterized by spirited scholarly debate. An im portant current debate is that
between the evolutionary approach and the more neoclassically inspired “ endogen­
ous growth theory”.
This chapter argues that the gap between these two approaches is rooted in
fundamental differences in their basic worldviews. While the neoclassical tradition
adheres to a worldview in which cause and effect are clearly separable, and growth
is an ordered, steady state phenomenon, the evolutionary worldview is one o f
 IN N O VATIO N AND ECONOMIC GROWTH 489

historical circum stances, co m p lex causal m echanism s that change over tim e, and,
above all, turbulent grow th patterns that appear to be far from a steady state.'
Before these tw o approaches are com pared (in Section 18.3), I discuss some o f the
perspectives found in the earlier literature (Section 18.2). This discussion includes
both highly applied m ethods from the m ainstream toolbox o f econom ics (growth
accounting and the literature on R & D and productivity) and applied w ork from a
post-Keynesian o r Schu m peterian perspective. Section 18.4 outlines a few lines for
future research.

1 8. 2
G rowth and T echnology:
Trad itio nal Economics A pproaches

While technological change and economic growth were at the core of the work of the
classical economists (think of Adam Smith or Karl Marx), these topics largely
vanished from the scene with the neoclassical revolution in economic thinking in
the late nineteenth and early twentieth centuries. The neoclassical growth models
that appeared h a lf a centu ry ago (So lo w 1956) treated technological change as an
exogenous phenomenon. Technology was an explanatory factor “of last resort” in
the sense that growth not explained by the variables included in the model was
assumed to be the result of exogenous technological change. However, when empir­
ical work—so called “growth-accounting” (Abramovitz 1956; Solow 1957)2— indi­
cated that the unexplained share of long run economic growth tended to be very
high, the interest in technological change and other possible explanatory factors not
taken into account by the modelers increased.
Following Solow (1957), growth accounting commonly starts from the assump­
tion of so-called “ neutral technological change ” implying that technological change
improves the productivity of both labor and capital equally. Moreover, all markets
are assumed to be “ perfectly” competitive and in equilibrium. Economies of scale
are assumed to be insignificant.
These assumptions support the following approach to calculating the contribu­
tion of “technological progress” to economic growth: subtract from the growth rate
of GDP the weighted growth rates of the capital stock and employment, using the
share of wages in GDP as a weight for employ ment, and subtract from one to get the
weight for the capital stock. What remains, the “ residual, is labeled total factor
productivity” (TFP) growth. This should, following Solow, be seen as the result of
technological progress. Although convenient, the strong assumptions underlying
these calculations are likely to be violated in practice, and the residual almost
certain ly includes m a n y m o re facto rs than ju st the c o n trib u tio n o f tech n ology. T h is
is w h y A b ra m o v itz (1956) called the resid u al “ a m easu re o f o u r ign o ran ce.
O ver the years, the grow th accou n tin g m eth od has been greatly refined. First, the
collection o f m ore refined statistical data allow ed m o re p ro d u c tio n factors to be
distinguish ed, e.g. h u m an capital, vario u s types o f lab or (different b y education al
level), different types o f capital, etc. In this way, the residual sh rin ks, attrib u tin g a
larger part o f it to the factors that are n o w better m easured (D en iso n 19 6 2 ).3 T h e
second line o f extension has been to refine the concept in a theoretical w ay, fo r exam ple
b y assu m in g that som e factors (capital) are q u asi-fixed , i.e. can n ot be reduced or
increased as a result o f sh o rt-ru n flu ctu ation s in o u tp u t grow th (e.g. M o rriso n 1986).
T h e T F P concept rem ains im p o rtan t in studies o f grow th b y eco n o m ists, as it
p ro vid es a “ p ro x im a te ” in d icato r o f the im p act o f tech n o lo gical ch an ge o n grow th.
N on eth eless, the p ro b lem s that rem ain in co n cep tu alizatio n an d m easu rem en t have
m ad e m an y sch olars in the field critical o f its use. P erh ap s the m o st fu n d am en tal
critiq u e is that m an y o f the factors go in g in to the grow th a cco u n tin g calcu latio n s are
interrelated b y causal links n ot accoun ted fo r b y the u n d erlyin g th eo ry.4
G ro w th th e o ry in the 1950s and 1960s w as b ased o n a sim p listic v ie w o f tech n o logy
as a “ p u b lic go o d ” Tech nological kn ow led ge o b v io u sly has so m e ch aracteristics o f a
p u b lic go o d , i.e. m o re than o n e firm can use the sam e piece o f k n o w led ge at the sam e
tim e (n o n -riv a lry ), an d once kn ow led ge is in the open, it is h ard to exclu d e specific
firm s fro m u sin g it (n o n -exclu d ab ility). In its extrem e fo rm , this v ie w leads to the
co n clu sio n that all kn ow ledge can be acquired extern ally as “ general kn ow ledge,”
an d firm s need not d evelop kn ow led ge them selves.
O th er im p o rtan t aspects o f tech nology, how ever, m ake it a p riva te rath er than a
p u b lic go o d (see also Fagerberg, C h . 1 in this vo lu m e ). P u re p u b lic g o o d s do not
require an y special effort o r special skills o n the side o f the co n su m er o r receiver o f
the services o f the go od . T h is is o b v io u sly n o t the case fo r tech n o lo g ical know ledge.
U sin g tech n o logical know ledge, even i f it stem s fro m the p u b lic d o m a in , requires
con siderab le skills and efforts o n the side o f the receiver o f this k n o w led ge. The
reason fo r this is that kn ow led ge has a stro n g ly cu m u lative an d o ften tacit character.
E v ery piece o f n ew kn ow led ge b u ild s to a large extent o n p re v io u s kn o w led ge, an d to
a p p ly know ledge requires that one have co m m a n d o ver the o ld er kn ow led ge on
w h ich the new kn ow led ge builds.
A n u m b er o f m odels developed d u rin g the 1950s and 1960s m ad e tech n o logy
en dogen ou s. In K ald o r (1957) this to o k the fo rm o f a so -called “ tech n ical progress
fu n ctio n ,” w h ich assu m ed a lin ear relation betw een grow th o f la b o r p ro d u ctivity
and the gro wth o f capital p er w orker. K a ld o r’s w o rk gave rise to a sp ecific trad ition ,
often labeled “ P o st-K eyn esian ism ” W ork in this trad itio n takes the role o f dem and
into account exp licitly.5
T h e P ost-K eyn esian trad itio n also em phasizes the role o f “ cu m u lati ve cau satio n ”
o r “ p ositive fe e d b a c k ” C o n tra ry to the neoclassical idea o f k n o w led ge as a public
go o d , these m odels assum e that know ledge is specific to the agents that d evelop it
 IN N O V A T IO N AND ECONOMIC GROWTH 49I

and does not spill over easily to other agents or nations. This idea was applied to
regional growth in Kaldor (1970), and goes hack to Verdoorn (1949), Fabricant (1942)
and Young (1928). In this view, generating knowledge is mainly a learning process
deeply rooted in gaining experience with specific production processes and prod­
ucts: learning-by-doing and learning-by-using are key concepts. Only those engaged
in the actual learning experiences will gain from it, and others, who do not profit
from experience, will be left behind.
The consequence o f this is a tendency for “ success to breed success” : those nations
(or regions, or agents) that are grow ing rapidly accum ulate experience and hence
learn faster than others. T his leads to a better com petitive position for those already
ahead and enables them to m ove fu rther ahead. Hence, the crucial tendency here is
one o f divergence, in w hich som e nations (regions) are able to grow rapidly while
others are left behind. A m odel o f regional grow th along these lines was presented in
Dixon and Thirlw all (1975):6
An important contribution in this post-Keynesian tradition is Cornwall (1977),
who argues that manufacturing is the leading sector in economic growth because of
the externalities it generates for other sectors. The motivation for this hypothesis is
consistent with the Schumpeterian idea that large innovations have a broad impact
across many sectors (see also Section 18.4.2 below). This is coupled with a view that,
for many countries, the inflow of knowledge from abroad is paramount (see also
Fagerberg and Godinho, Ch. 19 in this volume).
Attempts to generate m odels o f endogenous technological change were also
formulated in the neoclassical trad ition in the 1960s. A rrow (1962) introduced a
model o f learn in g-b y-d oin g as the source o f technological progress, and Uzawa
(1965) and Shell (1967) form ulated full-fledged grow th m odels with endogenous
technological change, w hich in m an y respects can be considered as the front-runners
of the wave o f “ endogenous grow th m odels” that em erged in the late 1980s and early
1990s (see Section 18.3.3).
The w ork on grow th accounting also contributed to the em ergence during the
1970s o f a purely em pirical approach to the issue o f growth and technology that
formulated and estim ated econom etric m odels o f the relationship between G D P and
R&D investment (e.g. G rilich es 1979, 1984). These studies em ploy a production
function that adds a “ know ledge stock” m easure (typically, cum ulative, depreciated
R&D investm ent) to the trad ition al factors o f labor and capital. Estim ates o f the
elasticities o f ou tp u t w ith regard to the variou s production factors suggest that
knowledge (R & D ) has a significant im pact on produ ctivity growth. This approach
has been used at variou s levels o f aggregation: firm s (e.g. G riliches and M airesse
1984)1 sectors (e.g. Verspagen 1995) or countries (e.g. G riliches 1986).
An important issue in this literature is the empirical identification of so-called
R&D spillovers. This goes back to the notion that knowledge is at least partly a public
good and can be used by others than the firm that developed it. In the context of
a production function, spillovers are incorporated by introducing two R&D
k n o w led ge stocks: o n e fo rm ed b y R & D u n d ertak en b y the firm (o r n a tio n , o r sector)
itself, and an o th er o n e fo rm ed b y R & D u n d ertak en b y o th er firm s (n atio n s, sectors;
see Lo s an d V erspagen 2000, fo r a m icro -level ap p lica tio n ). T h ese studies generally
conclude that the social rate o f return to R & D is larger th an the p riv a te rate o f return,
at a n y level o f aggregation. Firm s thus tend to ben efit fro m o th er firm s5 R & D , an d the
sam e h o ld s at the in tern atio n al level: o n e n a tio n s p ro d u c tiv ity g ro w th is to an
im p o rta n t extent d eterm in ed b y that o f others. D espite th eir e co n o m e tric so p h isti­
catio n , how ever, these studies reveal little ab ou t the exact ch an n els th ro u g h w hich
sp illovers operate. These channels m ay in clu d e trad ed g o o d s, em p lo yee m obility,
tech n o lo g y alliances, o r even kn ow led ge that is usim p ly in the a ir ” 7

18.3 C o m petin g Pa r a d ig m s for

Explaining the Relation betw een
G rowth and T echnology

Two m ajor approaches emerged during the 1980s and 1990s as the dominant
approaches to the analysis o f the relationship between technology and growth.
These are the neoclassical approach, which is also dom inant in other fields o f
economics, and the neo-Schumpeterian or evolutionary approach. While the
neoclassical approach consists o f a relatively hom ogenous set o f interrelated
sub-approaches (models), the field o f neo-Schumpeterian or evolutionary econom ­
ics consists o f a more loosely connected set o f contributions. The evolutionary
approach includes formal models as well as m ore “ appreciative” or historical
approaches, as will be explained in m ore detail below. Even the label used to describe
this approach is not yet com m on understanding. Here, we will use, m ainly for
convenience, the short description o f “ evolutionary economics,” but we include
under this heading a broad category o f work, including what some have called
neo-Schumpeterian economics.
Both o f these approaches agree on basic issues such as the im portance o f innov­
ation and technology for economic growth, as well as the positive role that can be
played by government policy for science and technology. Yet they disagree on the
behavioral foundations underlying these respective theories. These differences can
be characterized by saying that the neoclassical theory sacrifices a significant amount
o f realism in terms o f describing the actual innovation process in return for a
quantitative modeling approach that favors strong analytical consistency, while
the evolutionary approach embraces the micro com plications o f the innovative
 IN N O V A T IO N AND ECONOMI C G RО W T H 493

process and applies a m ore eclectic approach. G iven these differences, it is useful to
start with an overview o f their analytics in term s o f the m icroeconom ic aspects o f
endogenous technological change and innovation.

18.3.1 Microeconomic Aspects of Technology and Innovation

of Importance for the Analysis of Economic Growth
We focus in this section on tw o im portant aspects o f the m icro-foundations o f
innovation and technological change: uncertainty and differences in the significance
of innovations. Econom ists typically deal with uncertainty by postulating a prob­
ability distribution for a certain range o f events. Using these probability distribu­
tions, the econom ic consequences o f decisions can be weighted b y their probability.
Rational actors can m ake calculations that are m ore com plex than those in an
environment o f certainty, but the results do not differ appreciably. We refer to
such a situation as a case o f w eak uncertainty.
The situation changes, however, when the possible outcom es o f an uncertain
process are not k n ow n in advance, i.e. the events for which a probability distribution
is needed cannot be identified. Arguably, this is a better description o f the innovation
process, at least w here m ore radical innovations are concerned (Box 18.1 discusses
some exam ples o f this in the h isto ry o f com puting). We will refer to the situation in
which the possible outcom es o f an uncertain process are not know n in advance as
strong uncertainty. U nder strong uncertainty, the elegant calculations using p ro b ­
ability-weighted outcom es to calculate the expected value o f a stochastic process
no longer apply. As will be seen below, the treatm ent o f uncertainty as either weak
or strong is an im po rtan t distinction between neoclassical and evolutionary
approaches to econom ic grow th.
The second issue to be discussed in this section is the technological or economic
significance of innovations. The history of technology is filled with innovations that
have transformed the world— a non-exhaustive list includes the steam engine,
electricity, the automobile, the computer, and genetic engineering. Each of these
innovations had an almost immeasurable impact on the economy. But there are
many examples of less significant innovations that have had far less economic
significance.
One m ay argue that the above com parison is not a fair one, since “ the com puter
or “ the steam engine” never existed. All o f the above exam ples o f radical innovations
took decades to develop, and were the result o f a com bination o f radical techno­
logical breakthroughs as well as m any cum ulative increm ental innovations.
Although one can therefore not speak o f “ the com pu ter” or “ the steam engine, it
still rem ains true that som e innovations, no m atter at which level they are defined,
are m uch m ore valuable than others. In fact, a large share o f innovations eventually
 Box 18.1 Technological change and uncertainty . . „. a
II — 11 - - * fn -f I I - i‘ ‘ ..................................... .............................................................. ............ ' ' n * 1 - i. ■ П i nminif t f r ' Y i Гі -щ

Uncertainty with regard to technologies may be understood in different degrees.

Consider, for example, the difference between the first conception o f a computer in
the days o f Turing, Mauchly, and Eckert (i.e. the 1940s and early 1950s), and the
introduction o f the Pentium chip by Intel in 1992. In the first case, according to a
history o f the early days o f computers in the United States (Katz and Philips 1982), the
leading business men o f the day saw no commercial possibilities for the computer. }
They quote Thomas f. Watson Sr., CEO o f IBM , as having expressed the feeling that f
‘The one SSEC machine which was on exhibition in IBM ’s Mew York offices could solve ^
all the scientific problems in the world involving scientific calculations.” The same T. J.
Watson, by the way, quickly led IBM into leadership in the global computer industry in
the late 1950s.
These pessimistic views o f the commercial potential o f computers reflect the fact
that businessmen such as Watson had no familiarity with computer technologies in
their modern form. Under such circumstances, it was impossible to appreciate the
many new uses that were to be found, or the possibility that the functions and
capabilities o f the room-sized computers o f the early 1950s could be made to fit in
the space o f a desktop. One o f several problems in recognizing the commercial
opportunities o f a major technological breakthrough, and a factor contributing to
uncertainty, is the lack o f any frame o f reference for judging these impacts.
The situation was very different in respect o f the introduction o f the Pentium chip by
Intel in 1992. By that time, Intel and other firms, as well as users, had accumulated
knowledge about the applications for computers and the devices attached to them.
Intel also knew that its products were a major input in the small computers that were
being purchased by a large population o f consumers and firms. But Intel still faced
some degree o f technological uncertainty, because o f the complex nature o f the new
design. Indeed, Intel’s engineers had made a small mistake that could produce errors in
the Pentium’s calculations. The publicity over the so-called “ Pentium Bug” eventually
forced Intel to take back all faulty chips and offer free replacements. The example shows
that even for a relatively mature technology, some degree o f uncertainty remains.

turns out to be useless, simply because a prom ising technological idea never makes it
into a successful commercial application.
This has given rise to a distinction in the literature between incremental and
radical innovations. But this distinction obscures the fact that the size distribution of
innovations is not a dichotomy, but instead covers a continuous range o f innovation
sizes. Moreover, there is an im portant interaction and interdependence between
radical and incremental innovations. For example, the first workable steam engine
(the so-called Newcomen engine) was very large and had a lim ited applicability as
well as efficiency. It took more than fifty years for the next step to be taken, Le. James
Watts engine with a separate condenser. If we can characterize the impact o f some
innovations as major, basic, or ’‘radical,” it is only because o f a continuous
stream o f incremental innovations following the introduction o f a basic new design.
 Box 18.2 Evolution and the blind watchmaker

Let us use Richard D aw kins's m etaphor o f the blind watchm aker to illustrate the
general idea behind economic growth as an evolutionary process. Dawkinss story
starts from the idea o f W illiam Paley, an eighteenth-century theologian. Paley argued
that certain objects, like a watch, are by their nature obviously created by conscious
design, whereas for others, like a rock, it is easy to believe that they “ have always been
around” His argument then went on to stress that nature contains many such objects
that are obviously created by conscious design, The m ost fam ous o f such objects
discussed by Paley is the human eye. He then used this argument to offer the propos­
ition that the world must have been created by a conscious being (God).
Dawkins uses Paley s exam ples to argue that the watch m ay look as if it was carefully
designed (and in the case o f a watch it really was), but it might just as well have been
created by an evolutionary process that can be thought o f as a blind watchmaker. This
blind watchmaker is unable to design the watch by carefully planning it on a drawing
board and then implementing it using precision instruments. Instead, he operates
through the processes o f random mutation and natural selection. His approach is to
start with a simple device and add sm all and simple changes in a random way. These
changes are subjected to a real-world test, i.e., whether or not they lead to an
improvement in keeping the time. Only if they do so are they kept; otherwise they
are discarded. From a new design that incorporates such a successful small change, the
process may start again, and step-by-step a more complicated design emerges. In the
end, after a long and gradual process, a complicated artifact such as a watch may result,
Although this artifact looks as if it were carefully designed, it was instead the blind
watchmaker and his tools o f random mutation and natural selection that created it.
Carrying the metaphor over to economic growth and technology, our watchmaker is
blind because o f the strong uncertainty facing the individual economic decision maker.
No businessman can perfectly foresee the huge potential o f a new innovation when it
first emerges. But it is through a process o f incremental innovations, each one o f which
is implemented by an entrepreneur who sees some market for the newly resulting
artifact, that the full potential o f the technology unfolds. The incremental innovations
are the economic counterpart o f biological mutation. Natural selection has its coun­
terpart in economic selection, i.e. markets that decide whether or not certain innov­
ations become successful. Just as in biology, many of the “ m utations” (incremental
innovations) are not successful, and the selection process erases them from history.
The metaphor is thus concluded by arguing that, although the individual entrepre­
neur has to cope with strong uncertainty and therefore cannot design a process that we
may call a technological revolution, the capitalist system, working by means of a
combination o f the creation o f novelty (innovation) and economic selection
(markets), can create “objects” that seem as if they have been carefully designed.
With hindsight, technological revolutions, such as the diffusion o f steam power or
Information and Communication Technologies (ICTs) may look as if they were
planned from the very beginning to create a “ new economy, but in reality, so it is
argued by evolutionary theory, these technological systems were created by the trial-
and-error method o f the blind watchmaker,
49 6 BART VERSPAGEN

18.3.2 The Evolutionary Approach to Technology and Growth

18.3.2.1 The Evolutionary Philosophy

T h e e v o lu tio n a ry ap p ro ach to the analysis o f eco n o m ic g ro w th is b ased in p art on
the a x io m that in d iv id u a l h u m an s are u nable to cope in a fu lly m a x im iz in g w a y w ith
the co m p lexities o f tech n o lo g y that w ere d iscu ssed in S ectio n 18 .3.1. A single
e co n o m ic d ecision m aker, b e it an en trep ren eu r fro m the early d ays o f the In dustrial
R evo lu tio n o r a large m u ltin atio n al co rp o ra tio n fro m the tw en ty-first century,
sim p ly can n o t see all business o p p o rtu n ities that result fro m tech n o lo g ical p o ssib il­
ities an d /o r m an age th em in a w ay that m axim izes p ro fits. T h ese d ecisio n m akers
thus operate u n d er a schem e o f b o u n d e d ration ality, in w h ich relatively sim p le and
o ccasio n ally ad ap tive b eh avio ral rules (“ rules o f th u m b ” o r “ ro u tin e s” ) are used to
m ake decisions. These are n o t fixed, b u t can be ch an ged o ver tim e, especially so
u n d er the in flu en ce o f feedback fro m e co n o m ic p erfo rm an ce.
A lth o u g h these sim ple b eh avio ral rules help e co n o m ic d ecisio n m akers in a
tu rb u len t an d co m p le x w o rld cope w ith stro n g u n certain ty, th eir role sheds little
ligh t on the m ech an ism s th ro u gh w h ich co m p lex m o d e rn e co n o m ies rem ain on a
path o f constant tech n o logical im p ro vem en t that w e call e co n o m ic grow th . The
exp lan atio n o f aggregate eco n o m ic p erfo rm an ce in e v o lu tio n a ry eco n o m ics relies
on tw o forces: selection and the gen eratio n o f novelty. O ver tim e, the v a rie ty present
in the system is reduced b y selection— i.e. the grow th o f th o se entities that are better
ad ap ted to circu m stan ces, an d the decline o f those that are n ot. N o v e lty is constantly
add ed to the system , how ever, an d thus evo lu tio n is the o u tc o m e o f a constant
in teractio n betw een variety an d selection. In n o v a tio n is an im p o rta n t novelty­
gen eratin g process, and the m arket and oth er e co n o m ic in stitu tio n s are am ong
the m ost im p o rta n t selection m ech an ism s in m o d e rn econ o m ies.
In b io lo gy, the gen eratio n o f n o velty (m u tatio n ) is p u re ly ran d o m , an d there is no
w ay in w h ich the m ech an ism o f m u tatio n itse lf can learn to generate “ sm arter”
m u tatio n s. Each m u tatio n is tru ly “ b lin d ” in the sense that there is no ex ante w ay o f
telling w h eth er o r n o t it w ill im p ro ve the p erfo rm an ce o f the o rg an ism . In econom ic
evo lu tio n , how ever, d ecision m akers at the m icro level are n o t “ co m p letely b lin d ” —
th ey p lan their action s in o rd er to generate p o ten tially su ccessfu l in n o vatio n s in a
process that m o re clo sely resem bles the L a m arck ian v ie w o f evo lu tio n . Thus,
in n o vatio n s in tro d u ced b y p ro fit-seek in g, “ satisficin g” en trep ren eu rs w ill have at
least som e co m m ercial poten tial; in other w o rd s, they are m o st lik ely biased in a
“ p o sitiv e ” d irection . N evertheless, u n certain ty rem ain s im p o rta n t, since it is d iffi­
cult to foresee the cu m u lative effects o f n u m ero u s sm all, in crem en tal im provem en ts,
and because o f the system ic nature o f k n o w led ge that results fro m know ledge
sp illovers am o n g fields. A n actor o p eratin g in one field m a y invent so m eth in g for
w h ich he does not see the full poten tial in oth er fields.
The e vo lu tio n a ry ap p ro ach is p articu larly suited fo r an alyzin g h isto rical p ro ­
cesses. E vo lu tio n and h isto ry are both a co m p lex m ix tu re o f ran d o m factors, or
 IN N O V A T I O N AND ECONOMIC GROWTH 497

contingencies, sn d m ore system atic tendencies* It is a w ell-know n error to think, that

the biological evo lu tio n ary process is goal-oriented, i*e. that is strives to achieve a
predefined aim . O u r discussion o f the blind w atchm aker m etaphor m ay have misled
the reader into thinking that such a goal exists, i.e. that it would be the aim o f
evolution to create a co m p lex artifact such as a watch o r a hum an eye* Instead, it is
only the individual m u tation that has a sense. The accum ulation o f incremental
innovations m ay seem to have a purpose, but in fact there is no force in the system
that has form ulated or even tried to achieve such a goal. The same applies to
economic evolution.
Such a view o f the w o rld as a m ixtu re between chance and necessity is shared
between the historical view o f the w orld, the evolutionary view o f the world, and the
dialectic (Hegelian) view o f the w orld. It is opposed to the N ew tonian or Laplacean
view that portrays the w o rld as a clockw ork in which future states o f the system can
be predicted with full accuracy if o n ly enough inform ation about the present state is
known. We will argue below that the neoclassical econom ic growth theory is m uch
more similar to the latter view.

18*3*2.2 Non-Formal Evolutionary Theorizing about

Economic Growth and Technology
The evolutionary approach to econom ic grow th also draws heavily on econom ic
history and the h isto ry o f science and technology in its analysis o f econom ic
development. H istorical analysis often is used b y evolutionary scholars to develop
heuristic patterns that can be used to describe and categorize these developm ents in
a more general way. “ In the appreciative and applied evolutionary literature m uch
has been m ade o f the concepts o f technological paradigm (D osi 1982) and natural
trajectories (N elson and W inter 1982). This is indeed an attem pt to im pose add ­
itional structure on tech n ology and differentiate discrete interrelationships in tech­
nological space from one another, i f o n ly ex p o s t . . . This should be contrasted with
the smooth, substitutable, u n b o u n d ed p rodu ction possibility sets o f neoclassical
theory” (Silverberg 2 0 0 1 :12 7 7 ).8
Dosi (1982) defines a technological paradigm as a “ m odel and pattern o f solution
of selected technological problem s, based on selected principles from the natural
science and on selected m aterial technologies.” The term is borrow ed from Kuhn s
philosophy o f science (K uh n 1962), w hich posits that the norm al developm ent path
Of scientific knowledge relies h eavily on a dom inant framework join tly adhered to by
the leading scientists in the field. The paradigm thus lim its the possible directions
technological developm ent m ay take.
In the interpretation o f Freem an and Lou^a (2001), a sm all num ber o f basic
innovations set out a technological paradigm that m ay dom inate techno-econom ic
developments fo r a lo n g time* W ithin the paradigm , the basic design o f the in n ov­
ation is constantly altered b y increm ental innovations, but the direction o f
49$ BART V E R S P A G E N

technological development is limited by the paradigm. Still, there is some room for
choice within the paradigm, and these choices are governed by the specific circum­
stances (e.g., scarcity o f a particular resource) in which the technology develops. This
development is termed a “technological trajectory”
Thus, in the paradigm/trajectory heuristic, a basic innovation can be thought of as
setting out developments in the techno-economic domain for a number o f years to
come, but the success of the paradigm, and hence of the basic innovation, depends
crucially on how well incremental innovation is able to adapt the paradigm to local
(e.g. industry, geographical and temporal) circumstances. These circumstances
include the skills and capabilities o f the workforce that has to work with new
machinery, as well as factors such as cultural aspects o f the society in which the
paradigm develops.
Another set of heuristics developed in the historical part o f evolutionary econom­
ics relates to the temporal clustering o f innovations. This part o f the literature starts
from Schumpeter's observation that innovations “ are not evenly distributed in time,
but that on the contrary they tend to cluster, to come about in bunches, simply

M anagerial
C ap italism

the factory
system
The telegraph process
■.■r*

Networks'
Canal mania

A ssem bly
company belt
ham as A. Ed ison

The rise of
Information
lechanizatiori The rise of The rise of The A m e ri­ and Communi­
of textiles, steam as a electricity as can System cation Techno-
production of pervasive a pervasive of M an u fac­ logy, - ■
pig iron technology technology turing

1750 1800 1850 1900 1950 2000

Industrial Age of Age of Age o f The

Revolution Steam and Electricity M ass Information
Railways and Steel Production Age

Figure 1 8 . 2 Approximate chronology of technological revolutions,

based on Freeman and Soete (1 9 9 7 ) (dates are approximate)
 IN NO VATIO N AND ECONOMIC GROWTH
499

because first som e, and then m ost firm s follow in the wake o f successful innova-
tioiT(Schumpeter 1939: 75), A lth ou gh Schum peter was in fact referring to a ten­
dency for increm ental in n o vation s to cluster follow in g a large innovation (this is an
idea not incom patible w ith the paradigm view sum m arized above), his idea has been
interpreted in the literature as im p lyin g that large (or “ basic” ) innovations cluster in
time (e.g. Mensch 1979: K leinknecht 1987). In this view, som e historical periods are
characterized by an above average rate o f (basic) innovations, while other periods
show a relatively low rate o f such activity.
Together, these tw o sets o f heuristics have interesting im plications for growth.
They suggest that technological in n ovation can introduce an uneven tem poral
pattern into econom ic grow th. In the early, exp loratory stages o f a paradigm , the
technology progresses rapidly, but the pace o f change slows when the paradigm goes
into its phase o f “ n o rm a l” developm ent, and slows still further when technological
opportunities becom e less n u m erou s (and the paradigm m ay start to break dow n as
a result o f this). The clustering-heuristic suggests variations over time in the rhythm
of growth sim ply because the rate at w hich large, influential innovations occur
differs over time.
One extreme interpretation o f this tem poral pattern o f innovation is the idea o f a
“long wave” in econom ic grow th, in w hich periodicity is bounded in a short range o f
50-60 years (e.g. K leinknecht 1987; Freem an and Lou^a 2001). Another view claims
that growth patterns are inherently turbulent, but w ith little regularity in terms o f
strict cycles. In any case, the evo lu tio n ary view argues that the uneven tem poral rates
of technological change m ean that the econ om y is alm ost always away from anything
that could be characterized as a steady state.
Theories and historical analyses o f this type propose a view o f the interactions
among technology, the econom y, and the institutional context. The institutional
environment is im po rtan t because it is both a facilitator o f and an im pedim ent to
technological change. M oreover, the institutional context is itself an endogenous
factor that changes under the influence o f technological and econom ic develop­
ments. Although it is som etim es claim ed that theories o f this type suffer from
“technological d eterm in ism ” (i.e. a tendency for one-w ay causality from technology
to growth: see e.g. B ijk er et al. 1987), w o rk such as that o f Perez (1983) proposes an
interactive relationship am o n g institutions, the econom y, and technology that
emphasizes m utual causality.

^ ‘3*2.3 Form al E volu tion ary Grow th M odels

Evolutionary ideas have also been used to form ulate m odels o f econom ic growth and
technology. The starting p o in t o f this tradition is the m odel in N elson and W inter
(1982), in which heterogeneity is defined in term s of firms, using production
techniques that employ a fixed ratio of labor and capital (so-called Leontief-
technology). The generation of novelty (new fixed proportion techniques) occurs
;O0 BART VERSPAGEN

is a result of search activities by firms, but search is initiated only when the firm's rate
af return falls below a certain (arbitrarily set) value. Search may take two different
forms: local search or imitation. In the first case, firms search for new, yet undiscov­
ered techniques, each o f which has a probability o f being discovered which linearly
declines with technological distance from their current technology (hence the term
lo c a l search). In the second search process, imitation, a firm searches for techniques
currently employed by other firms but not yet used in its own production process.
Like most models in this tradition, the Nelson and Winter model has to be
simulated on a computer to obtain an impression of its implications. The model,
which is calibrated with the Solow (1957) data on total factor productivity for the
United States in the first half of the century, yields an aggregate time path for capital,
labor input, output (GDP), and wages (or labor share in output) that corresponds in
a qualitative sense to those observed by Solow. Based on these results, Nelson and
Winter argue that “ it is not reasonable to dismiss an evolutionary theory on the
grounds that it fails to provide a coherent explanation o f . . . macro phenomena" (p.
226). More specifically, they argue that although both the neoclassical explanation of
economic growth offered by Solow and the Nelson and Winter model seem to
explain the same empirical trends, the causal mechanisms underlying the two
perspectives differ greatly:
the neoclassical interpretation o f long-run productivity change. . . is based upon a clean
distinction between “ moving alon g5 an existing production function and shifting to a new
one. In the evolutionary theory. . , there was no production fu nction.. . . We argu e. . . that the
sharp “ growth accounting” split made within the neoclassical paradigm is bothersome
empirically and conceptually. (Nelson and Winter 1982: 227)

Evolutionary models following Nelson and Winter (1982), such as Chiaromonte

and Dosi (1993) and Silverberg and Verspagen (1994), extend these conclusions.
A more complete overview is in Silverberg and Verspagen (1998). The model by
Chiaromonte and Dosi shows how growth rates in a cross-section o f nations may
differ. The models by Silverberg and Verspagen show how “ routines" of R&D
investment may arise endogenously in a population of firms, and how growth
patterns vary along the history of an economy that learns in such a “ collective" way.
O ne o f the rare m o d els in this trad itio n that is solved an a ly tica lly rath er than by
n u m erical sim u latio n is that o f C o n lisk (1989). U n d er the a ssu m p tio n that tech n o l­
o g y advances are ran d o m , C o n lisk co n stru cts a m o d el in w h ich the g ro w th rate o f the
aggregate e co n o m y is a fu n ctio n o f three variables: the stan d ard e rro r o f the
p ro d u c tiv ity d istrib u tio n o f n ew plants (w h ich can be in terpreted as the average
in n o vatio n size), the savings rate (w h ich is d efin ed so m ew h at u n co n ven tio n ally),
an d the speed o f d iffu sio n o f n ew know ledge. M o reo ver, b y ch an g in g so m e o f the
assu m p tio n s ab o u t the sp ecificatio n o f technical change, the m o d el em u lates three
stand ard sp ecification s o f technical change fo u n d in grow th m o d els in the n eoclas­
sical trad ition . In this case, the first an d third factors no lo n g er have an im p act on
 IN N O V A T IO N AND ECONOMIC GROWTH 501

growth (they are specific to the “ evo lu tio n ary” technical change specification o f the
model). However, the im pact o f the savings rate can be com pared between the
various model setups. C o n lisk finds that using purely exogenous technical change
(as in the Solow m odel), or learning b y doing specifications (as in A rrow 1962), the
savings rate does not have an im pact u p on (long-run) econom ic growth. This result,
which is in fact also well k n ow n from standard neoclassical growth theory, marks an
important difference between these m odels and his m ore evolutionarily-inspired
specification.
The recent so-called “ h isto ry-frien d ly m odels” (M alerba et a l 1999) aim to bring
evolutionary m odels closer to em pirical reality by reproducing the historical evolu­
tion of a particular industry, e.g. the com pu ter industry. To this end, they start with a
descriptive analysis o f in d u stry variables such as grow th, concentration, and em ­
ployment, and incorporate the insights fro m this analysis into a m odel the behav­
ioral foundations o f w hich are consistent w ith the evolutionary view. This m odel is
calibrated and sim ulated to reproduce real-w orld trends as closely as possible. While
this approach generates em p irically relevant m odels, the sim ulations em ploy a
relatively narrow set o f param eter values. The w ork devotes little attention to a
more open-ended investigation o f w hich minimal set o f assum ptions is necessary to
generate certain aspects o f the structural evolution o f specific industries.
These more open-ended uses o f evo lu tio n ary m icro m odels could lead to a new
class o f models that em p lo y relatively sim ple, evolutionary m icroeconom ic founda­
tions to generate a b roader range o f phenom en a in the evolutionary interpretation
of technology and grow th, rather than increasing the sophistication o f the m icro-
foundations. A m uch clearer focus on the salient m acro features and what really
drives them at the m icro level m ay result from this approach, which is necessary to
close the gap between the h istorical, evo lu tio n ary view and m odel building.

18.3.3 Neoclassical Views o f Economic Growth

and Technology
183.3.1 Endogenous G row th M odels
How has m ainstream econ om ic th eo ry coped w ith the com plexity o f technological
change? The literature on neoclassical m odels o f endogenous technology grew
rapidly in the 1980s and 1990s fo llo w in g the publication o f R om er (1986). Rom er s
niodel and others in this trad ition w ere m otivated b y the apparent flaws associated
with the assum ption in the S o lo w m odel o f decreasing m arginal returns to capital,
holding all other p ro d u ctio n factors (labor, land, infrastructure, buildings) fixed, the
productivity o f an extra (m argin al) unit o f investm ent w ould fall with grow th in the
existing capital stock. D ecreasing m argin al returns to investm ent could cause
502 BART VERSPAGEN

grow th to slow down or even cease in the long run. As growth proceeds, capital
accumulates, i.e, the capital stock increases, and hence an extra unit of investment
generates less and less growth. Exogenous growth or productivity (knowledge) had
been the traditional answer, but Romer (1990) and Grossman and Helpman (1991)
proposed to make technology endogenous by modeling the R&D process. Abstract­
ing from technicalities (a survey is provided by Verspagen 1992), this can be
summarized as follows.
A ll the m odels assum e that R & D is essentially a lo tte ry in w h ich the prize is a
successful in n o vatio n . In the m o d el b y A g h io n an d H o w itt (19 9 2), this in n o vatio n -
prize buys the firm a te m p o ra ry m o n o p o ly o f su p p ly in g the b est-p ractice capital
g o o d used fo r p ro d u c tio n o f c o n su m p tio n goods. T h e te m p o ra ry m o n o p o ly v a n ­
ishes w h en the next firm m akes an in n o vatio n . H ence, the in n o v a tio n process is
m o d eled as a “ q u a lity ladder” o f in n o vatio n s, in w h ich each n e w in n o vation
supersedes the o ld one. In the in d u strial o rg an izatio n literatu re, this is called
“ vertical d ifferen tiatio n ” o f produ cts.
In the m o d el b y R o m e r (1990), the in n o va tio n prize bu ys the su ccessfu l firm a new
v a rie ty o f capital that w ill be d em an d ed b y p ro d u cers o f c o n su m p tio n g o o d s forever,
b u t has to com pete w ith all o th er varieties (in vented in the past, w ith the range
co n tin u in g to exp an d in the fu tu re as a result o f R & D ). In this m o d e l, varieties o f
g o o d s (in n o vatio n s) d o n o t go o u t o f the m arket. S u b stitu tio n b etw een variatio n s o f
g o o d s is govern ed b y a u tility fu n ctio n o r p ro d u c tio n fu n c tio n (d ep en d in g on
w h eth er in n o va tio n takes place in co n su m er g o o d s o r in term ed iate g o o d s) w ith a
“ constant elasticity o f su bstitu tion.” T h is is called “ h o rizo n ta l d ifferen tiatio n .”
M o re tickets fo r the R & D lo ttery can be b o u g h t b y d o in g m o re R & D , w h ich is o f
course a costly process. R elative to the e v o lu tio n a ry m o d els co n sid ered above, the
cru cial assu m p tio n Is that the o u tco m es o f the R & D pro cess can be characterized
realistically b y w eak uncertainty, i.e. the firm is able to estim ate the p ro b a b ility that it
w ill get the in n o vatio n -p rize given its level o f R & D sp en din g. W ith exp ected benefits
an d costs o f R & D k n o w n , the firm m a y m ake a c o st-b e n e fit an alysis an d derive an
o p tim al le vel o f R & D spen ding. T h is w ill, on average, co rresp o n d to a given am ount
o f in n o vatio n , and p ro d u ce a given grow th rate. A lth o u g h a d d itio n a l assum ption s
are necessary (e.g. w ith regard to the w o rk in g o f capital m arkets in w h ich R & D
exp en d itu res have to be fin an ced ), this m ech an ism is the k ey to gen eratin g en d o gen ­
ous grow th.
Before en d o gen o u s grow th is possib le in these m o d els, there is o n e essential
assu m p tio n ab ou t the nature o f tech n o lo g y that needs to be m ade. T h is is related
to the (partly) p u blic go od nature o f technology. In the n ew gro w th m o d els, this is
represented by the assu m p tio n that there are tech n o lo g y sp illo vers b etw een firm s in
the R&D process. T h e assu m p tio n takes tw o fo rm s, d ep en d in g o n w h ich flavor o f
m odel is used. In the horizontal d ifferen tiation type m o d els (also called “ love-
o f-v a rie ty m o d els), each in n o vatio n increases the level o f gen eral know ledge
available in the econom y, and this increases the p ro d u c tiv ity o f the R & D process
 in n o v a t io n and ec o n o m ic gro w th
503

itself (Romer 1990). T h is assu m ptio n is necessary because o f the ever m ore severe
competition between the varieties o f capital goods, and the falling profit rates that
this causes. A tendency fo r R & D to be m ore productive (i.e. the costs o f R & D to fall)
offsets this falling pro fit rate, and keeps R & D feasible in the long run (see Grossm an
and Helpman 1991).
In the quality ladder m odels (vertical differentiation), each new innovation
destroys the m o n o p o ly o f the old innovator. H owever, the new innovator also builds
on the previous in n ovation , because the quality o f the new capital good is a fixed
increase over the previou s one. In other w ords, each new innovator is “ standing on
the shoulders o f g ia n ts” and know ledge spills over intertem porally from one
innovator to the next one. W ithout this spillover, endogenous growth would not
be possible.
The technological spillovers in endogenous growth m odels lead to increasing
returns to scale at the aggregate level. Even though the production functions o f firms
at the micro level are characterized by constant returns to scale, the R & D spillovers
that flow from one firm to the rest o f the econom y im ply increasing returns at
the aggregate level. In term s o f the expression for the aggregate growth rate o f the
economy, this feature o f the endogenous grow th m odels im plies that growth at the
country level depends (ceteris paribus) on the size o f the country. Taken literally, this
means that (ceteris paribus) larger countries will grow m ore rapidly. Related to
this issue is the fact that the basic endogenous grow th m odels are quite sensitive to
small changes in the m odel specification w ith regard to technology spillovers.
A slightly different specification o f the im pact o f “ general knowledge” on R & D
productivity will lead to either zero grow th in the long run, or to increasing growth
rates in time (G rossm an and H elp m an 1991).
Technological spillovers m ake endogenous grow th possible, but pose a challenge
for policy makers. W hen tech n ology generates positive externalities, the social
benefits o f R & D are larger than the private benefits (a rational firm investing in
R&D does not consider the benefits o f its R & D for its com petitors). Hence the
amount o f R & D investm ent “ generated b y the m arket” w ill be too low from a social
point o f view. Technology p o licy in the form o f R & D subsidies m ay bring the
economy to a higher, so cially optim al grow th path. A sim ilar conclusion is reached
in a model o f h um an capital and grow th in Lucas (1988). In A ghion and H owitt
(1992), there is also a negative externality: each new innovator destroys the rents o f
the existing m o n o p o list (this is called “ business stealing,” or, in line with Schum p-
eter (1939), “ creative destru ction ” ). In this m odel, private R & D investm ent also can
be too high from a social w elfare perspective, depending on which o f the two form s
of externalities (creative destruction or standing on the shoulders o f giants) is
stronger.
The development of this new class of models raises promise and problems. On the
positive side, it can be argued that this new growth theory takes seriously a number
of arguments about technological change previously championed by evolutionary
5©4 B A S T VEBSPAGEM

theorists but' ignored by mainstream economists, These include the notion that
R&D and technology are essentially stochastic phenomena (although evolutionary
theory would argue that the type o f uncertainty, i.e. weak uncertainty in which the
probability distribution is known, is still not very adequate), and the importance o f
technology flows between agents (spillovers) for growth in the long run, The
implication in many o f these models that technology policy matters for growth
also is relatively consistent with evolutionary theory, but may be less easily accom­
modated by mainstream economic theories that emphasize the efficiency o f market
forces.
On the negative side, these new growth models still propose a view o f the
interaction between economic growth and technology that differs substantially
from that of evolutionary theory. The evolutionary view is one in which contingency
and more systematic factors blend together in the dialectical process of historical
time, but the new growth theory is still much closer to a Newtonian clockwork world
in which there is a certain degree of “weak” uncertainty. In other words, the new
growth theory still portrays the relationship between technology and growth as one
of a steady-state growth pattern, which can be “tweaked” relatively easily by turning
the knobs of the R&D process.
The evolutionary inclination, on the other hand, is that the nature of the growth
process is more complex and variable over time. While the importance attached to
the technology factor is shared with the new growth models, the belief that the
relation between technology and growth is easily tweaked is not. In the evolutionary
view, it is hard to predict exactly the impact of a policy measure, because it impacts
on a complex range of interrelated factors. Moreover, while relations between a
number o f factors may have been revealed by careful research for a specific instance
in time, it is to be expected that the nature of this relationship will change over time,
exactly because of the (co)evolutionary nature of the process.
A more recent branch of new growth theory is the group of models that comes
under the heading of “general purpose technologies” (GPT, Helpman 1998). A GPT
is defined in essentially the same way as a basic innovation or paradigm in the
evolutionary tradition. It consists of a basic technology (radical breakthrough), but
this needs to be developed in the form of a range of intermediate (capital) goods.
Within each GPT, the determinants o f productivity are essentially the same as in one
o f the variants of the new growth models discussed above: technological change
takes the form o f an ever-expanding range of capital goods, but this is time-specific
to the GPT. Thus, we see that at least two ideas from the evolutionary tradition are
captured: the idea o f differences in innovation size, and the idea that incremental
innovations are responsible for the diffusion of a basic technology.
The GPT model generates cyclical growth. In its simplest form, the cycle consists
o f two phases. In the “ low growth phase,” the new GPT has been discovered, but is
not yet in operation. New capital goods are being developed for it, and this activity
has been halted for the old GPT. Thus, economic growth is low, because the main
 IN NOVATION AND ECONOMIC GROWTH 505

technology in use is no lon ger being developed. O nce enough capital goods are
available for the new G PT, its p ro d u ctivity outperform s that o f the old GPT, the old
GPT vanishes, and the econ o m y shifts into a “ high growth phase.”
The G PT m odel resem bles the evolutionary, Schum peterian idea o f long waves in
economic growth. But scholars in the latter tradition have m oved away from the
fixed and determ inistic cycle that characterizes the G P T m odel. Its clockw ork view o f
economic growth has been dom in an t in the neo-classical tradition since the Solow
model. One illustration o f the lim itations resulting from this view is the fact that, in
the GPT view o f the w orld , there is o n ly roo m for substitution between subsequent
paradigms. But econom ic and technological histories are filled with exam ples o f the
adaptation and survival, often in m o dified form , o f old paradigm s. For exam ple,
although the autom obile is typ ical o f the m ass-production paradigm , it still plays a
crucial role in the m odern “ In fo rm ation Econom y,” although IC T has indeed been
applied in the pro d u ction o f cars.
In conclusion, the evo lu tio n ary tradition and the neo-classical tradition have
converged som ew hat in the ph enom en a deem ed central within each analytic ap ­
proach. But they disagree on the essential nature o f the growth process. The
neoclassical th eo ry conceptualizes grow th as a determ inistic process in which
causality is clear-cut, and p olicies can be built on an understanding o f tim e-invariant
determinants o f grow th patterns. In the evolutionary view, on the other hand,
contingencies and specific historical circum stances play a larger role, and causal
mechanisms that prevail in one period m ay be subject to endogenous change in the
next. In such a w orld , designing p o licy is harder, but not im possible.

18.3.3.2 E m pirical Work on G row th an d Technology Follow ing

the Endogenous Grow th M odels
The new growth m odels led to a tidal w ave o f em pirical w ork on growth. Temple
(1999) provides a detailed o verview o f this literature. The source o f data for nearly all
of this w ork was either the data b y M ad dison (1995) or the so-called Penn W orld
Tables (PW T, Sum m ers and H eston 1991). The PW T provides a broad cross-section
of data for over a h u n d red countries. A crucial topic in the em pirical debate
following the endogenous grow th m odels is the respective roles o f steady state
growth rates and convergence tow ard them . W hile the Solow m odel predicts that
countries will converge to identical steady states (dependent on the exogenous rate
of technological progress available to everyone), endogenous growth m odels predict
that steady states w ill generally differ between countries. Em pirical w o rk on this
issue has used a w ide range o f variables in regressions o f growth rate differentials
between countries, in o rd er to exam ine cross-national differences in steady state
growth rates.
Unfortunately, this app roach is d ata-driven rather than theory-driven: an overall
framework that governs and ju stifies the selection o f factors is lacking. Also, m any o f
506 BART VERSPAGEN

the estim ation results are sensitive to a sm all n u m b e r o f o b serva tio n s in the
large sam ple (L evin e and Renelt 1992). N on eth eless, this w o rk lead s to the co n clu ­
sio n that stead y state grow th rates d iffer betw een n ation s, G ro w th rates m ay
converge tow ard a co u n try-sp ecific stead y state gro w th p a th at best (so-called
co n d itio n al con vergen ce), lead in g to the d ivergen ce o f g ro w th path s a m o n g co u n ­
tries, G ro w th seem s to be h eterogen eou s am o n g co u n tries startin g fro m lo w levels o f
G D P p er capita, w ith so m e cou n tries falling b eh in d , and so m e co u n tries b ein g able
to catch up. T h is p h en o m en o n is discu ssed in m o re detail in F ag erb erg an d G o d in h o
(C h , 19 in this vo lu m e).
Jo n es (1995a an d b) has argu ed that the o b served e m p iric a l reco rd o n R & D and
grow th is incon sisten t w ith the theoretical p red ictio n s o f e n d o g en o u s grow th
m o d els (see B o x 18,3 on the wJon es critiq u e” and se m i-e n d o g e n o u s gro w th m odels).
H e observes that the p o stw ar em p irical eviden ce does n o t c o n firm the relationship
p ro p o sed b y R & D -b a se d en d o gen ou s grow th m o d els that an in crease in the n um ber
o f R & D w orkers leads to h igh er rates o f e co n o m ic grow th . Jo n e s notes that the
n u m b er o f R & D w o rk ers has increased since the 1960s, b u t gro w th rates ( o f total
facto r p ro d u ctivity) have either been constant o r d eclin in g d u rin g the sam e period.
T h e so-called “ Jo n es critiq u e” has led to still m o re w o rk in the e n d o g en o u s grow th
trad itio n since its pu b licatio n . Jo n es (1995a) suggests а so -called sem i-en d o gen ou s
grow th m o del, w h ich appears to be m ore consisten t w ith the e m p iric a l facts, but in
w h ich en d o gen o u s grow th o n ly takes place w h en the p o p u la tio n grow s.

Box 18.3 The "Jones critique" and semi-endogenous growth models

Figure 1 8 ,3 illustrates the Jones critique for the United States and the European Union
during the 1 9 8 0 s and 1 9 9 0 s. The R&D-based endogenous growth models predict that
the growth rate of an economy, which is here approximated by total productivity factor
growth, depends on the number of researchers in R&D. We see a steady increase in the
latter, both in the US and the EU, but total factor producti vity growth does not display a
clear trend—instead it fluctuates widely around a roughly constant level. Does this
constitute evidence against the relationship between innovation and economic growth?
Jones suggests an alternative model, which differs from the R&D-based endogenous
growth models by Romer, Grossman and Helpman, and Aghion and Howitt by a
different specification of the invention process. Whereas these original R&D-based
growth models assumed that the growth rate of knowledge depends on the number of
R&D workers in a linear way, Jones assumes that there are decreasing returns to R&D
labor. This assumption is based on the idea that "the most obvious ideas are discovered
first, so that the probability that a person engaged in R&D discovers a new idea is
decreasing in the level ofknowledge ... [and] the possibility that at a point in time the
duplication and overlap of research reduce the total number of innovations” (Jones
1 9 9 5 a; 765)* in this so-called semi-endogenous growth model, endogenous growth is
only possible when the population grows.
 INN О V A 1 ION AND E C O N O M IC GROWTH
507

№
C
О

1980 1985 1990 1995 2000

-ResearchersUS ResearchersEU ■—TFPgrowthEU *TFP growthUS

R§. 18.3 The Jones critique. Total factor productivity growth trends are
flat while the number of researchers in R&D increases (R&D researchers
on left scale, tfp on right scale).
Source: Source for tfp data: Groningen Growth and Development Centre Total Economy Growth Accounting Database.
■'Ш шш -Ші Ш Ш - i ste: OECO М а ш Science andTertinology Indicators Database.

From the point of view of evolutionary growth theory, the Jones critique appears to
be the result o f the misguided emphasis on steady-state growth states in the R
based endogenous growth models. The assumed relationship between R&D labor, the
number of innovations, and resulting economic growth is based on assumptions ot
va^Mbrium bdiavior and weak, uncertainty, in the less mechanistic wo:utionary
woiii innovation, R&D, and growth are linked in a less rigid relationship that may
f aenge over time as a result of new and radical technological deve pments. n is
S i i wHdie specific relationship between R&D labor and TFP grewt о serve у ones
j|iB8 yweB be specific to the historical circumstances of the period, and may be subject о
change in the future.

International en d o gen ou s grow th models have provided ot ei'n ew mp - .

empirical trad ition o f research o n R& D and p ro d u ctivity initiate у n іс t .
recent research focuses on the channels fo r Internationa^ transmissio
spillovers. The assum ption by Coe and Helpman (i 9 9 5 ) is that these R P
508 BART VERSPAGEN

are em b od ied in traded go od s, and, hence, that R & D w eigh ted b y trad e flo w s m a y be
used to m easure them . T h e em p irical analysis b y C o e an d H elp m an sh ow s th at the
co rrelatio n s betw een T F P grow th an d this m easu re are in d eed stro n g , suggesting
that trade is an im p o rta n t sou rce o f kn o w led ge sp illovers. H o w ever, subsequen t
co n trib u tio n s sh ow that oth er w eig h tin g schem es m a y p ro v id e d ifferen t in terpret-
ation s. Fo r exam p le, Lich ten b erg an d V an Pottelsbergh e (1996) sh o w th at Foreign
D irect In vestm ent (F D I) m ay be a carrier o f sp illo vers an d V erspagen (x997) show s
the im p o rta n ce o f in ter-sectoral sp illovers, w h ile K eller (1998) ls critical o f the
vario u s w eigh tin g schem es an d b en ch m ark s th em again st a ra n d o m w eig h tin g
schem e. These results also are sensitive to the m easu rem en t b y e m p irica l researchers
o f ab sorp tive cap acity in the sp illo ver-receivin g cou n tries.
A n interesting “ m erg er” betw een the em p irical trad itio n on p ro d u c tiv ity and
R & D , o n the one h an d, an d new grow th th e o ry on the o th er h an d , is the em p irical
m o d el b y E aton and K o rtu m (1999). T h is p ap er p ro vid e s a m o d el in w h ich in n o v ­
atio n and tech n o lo gy d iffu sio n are b oth d rivers o f co u n try -lev e l grow th . T h e m odel
is m o tivated b y em p irically observed trends, an d is estim ated w ith data o n tech n o l­
o g y in d icators (patents, R & D ) and grow th. T h e results o f the estim atio n s sh o w that
both en d o gen o u s R & D and the d iffu sio n o f k n o w led ge betw een co u n tries co n trib ­
ute to grow th , alth o u gh the m ix betw een these tw o sou rces d iffers greatly between
co u n tries and tim e perio d s. T h is ap p ro ach an d its co n clu sio n s also has m u ch in
co m m o n w ith earlier tech n o lo g y gap m o d els such as that o f N elso n (1968), as
su rveyed b y Fagerb erg and G o d in h o (ch. 19 in this v o lu m e ). F ag erb erg an d V erspa­
gen (2002) recently reassessed the p o st-w a r evidence fo r these typ es o f m o d els, and
co n clu d ed that, o ver tim e, in n o va tio n has b eco m e a m o re im p o rta n t source o f
grow th as co m p ared to the “ p u re” im itatio n o f fo reign tech n o logy. M o d e ls such as
that o f E ato n an d K o rtu m thus have great p ro m ise to gu id e n ew g ro w th th eo ry in a
d irection that has m u ch in co m m o n w ith the h isto rica lly -in sp ired e vo lu tio n a ry
ap p ro ach .

18.4 O u t l o o k for T h e o r e t i c a l R e s e a r c h
on I n n o v a t i o n a n d G r o w t h

N eoclassical w o rk in “ new gro w th ” or “ en d o gen o u s g ro w th ” recently has shifted

tow ard m o re “ realistic” m odels that can acco m m o d a te a range o f p h en o m en a
p re vio u sly o f interest o n ly in the e v o lu tio n a ry trad ition . H eertje d escribed this
convergence as follow s:
 i n n o v a t io n a n d e c o n o m ic gro w th 509

neo-Schumpeterians [i.eMthe evolutionary tradition] have been productive in their criticism

of the neoclassical scheme on the basic o f an evolutionary approach, but the (juestions they
have raised have been addressed more or less successfully by many scholars, who have close
links with the neoclassical tradition.. Л would not be surprised to see the present Schumpe­
terian mood to be part o f mainstream economics before the end of this century, (Heertje
19 9 3: 273-5)

Is further convergence o f the tw o traditions likely, as Heertje predicted for the end
of the (previous) century? O ne avenue for convergence is in the further analysis o f
the intertemporal variab ility o f grow th patterns. At least som e new growth m odels
(e.g. Aghion and H ow itt 1992) argue that tim e series o f econom ic growth show
variability and this is a m ain top ic in evolu tion ary m odels. The application o f
Pareto-type probability d istribution s, in w hich very large innovations have non-
negligible p ro b ab ility m ay b rin g the two approaches closer together, since they
provide an intuitive w ay o f m odelin g "stro n g uncertainty” (see e.g, Sornette and
Zajdenweber 1999).
Each o f the tw o approaches also contains a range o f im portant and interesting
lines o f research to be pu rsued. In the endogenous growth tradition, the returns to
purely theoretical w o rk seem to have slowed dow n, but im portant em pirical chal­
lenges remain open. The m ost fru itfu l avenue o f research here seems to be further
theoretical refinem ent indu ced b y em pirical w ork on technology and growth, with
the explicit aim o f developing em pirically relevant m odels instead o f new explor­
ations m otivated b y technical problem s w ith the existing models. For a long time,
empirical research has led the w ay in the m ainstream analysis o f technology and
growth, and this app roach still seem s to be the w ay forw ard.
Two main challenges confront the evolu tion ary tradition. The first is to develop a
research program that goes beyond ju st em ulating, although with a m ore plausible
micro-foundation, the results o f neoclassical analysis. Such an extension o f the
evolutionary research agenda could benefit from closer interaction with the non-
formal w ork in the evo lu tio n ary trad ition and greater reliance on historical research.
Evolutionary m odelers could seek to explain observed historical regularities in the
relation between grow th and technology.
A second challenge for evolutionary theorists is the development o f more practically
relevant models, for exam ple, with regard to specific policy advice. Evolutionary
theory rarely generates precise policy advice (see also Lundvall and Borras, Ch. 22 in
this volume), m ostly as a result o f the nature o f the theory that points to com plex
interactions and rather unpredictable dynam ics as im portant ingredients o f the
economic environm ent. To a certain extent, evolutionary theory will argue for a
change in the w ay p olicy is view ed, but m ore precise w ork on how this can be
implemented to achieve higher o r m ore sustainable econom ic growth remains
crucial
510 BART VERSPAGEN

N o tes _______ _______________________

1. The (older) data are necessarily rather imprecise, but the general trends are plausible on
the basis o f historical evidence. Note that since the vertical axis displays the logarithm o f
per capita income, a straight line would correspond to growth at a fixed rate, the slope o f
the line indicating the growth rate.
2. Solow (1957) is often quoted as the standard reference on growth accounting, but the
ancestry o f the method lies earlier (e.g. Tinbergen 1943 and Abramovitz 1956; for an
overview see Abramovitz 19 89 :13-15).
3. Well-known studies in this tradition are Denison (1962, 1966), Jorgensen (1967) and
Maddison (1987,1991): see Nadiri (1970) for an early overview o f the methodology.
4. Critical surveys o f the method can be found in Nelson (1973,1981) and Fagerberg (1988b).
5. Pasinetti 1993 analyzes growth and technology from a demand perspective.
6. An elaborate overview o f (empirical as well as theoretical) work on growth in the post-
Keynesian tradition is in McCombie and Thirlwall (1994)* A specific application to the
issue o f technology dynamics and growth is in Fagerberg (1988a).
7. Griliches (1992) provides a broad overview o f empirical studies estimating R8cD spill­
overs; Cincera and Van Pottelsberghe (2001) provide a survey on international spillovers,
Van Pottelsberge (1997) on intersectoral spillovers.
8. An early attempt to develop a heuristic similar to the ones cited by Silverberg is in Sahal
(1981).

R efer en c es

A b r a m o v it z , M. A . (1956 ), “ Resources and Output Trends in the United States since 1870,”
American Economic Review 46: 5 -2 3 .
------ (1989), Thinking About Growth, Cambridge: Cambridge University Press.
A g h i o n , P., and H o w i t t , P. (1992), “A Model o f Growth Through Creative Destruction,”
Econometrica 60 (1992): 323-51.
A r r o w , K. J. (1962), “ The Economic Implications ofLearning by Doing,” Review of Economic
Studies 29:155-73.
B a r r o , R. J. (1991), “ Economic Growth in a Cross-Section o f Countries,” Quarterly Journal of
Economics 106: 407-43.
B i j k e r , W. E., H u g h e s , T P., and P in c h , X (eds.) (1987), The Social Construction of
Technological Systems, Cambridge, Mass.: M IT Press.
C h e n e r y , H. B ., S y r q u i n , M., and R o b i n s o n , S. (1986), Industrialization and Growth: A
Comparative Study, Oxford: Oxford University Press.
C h i a r o m o n t e , E , and D osi, G. (1993), “ Heterogeneity, Competition, and Macroeconomic
Dynamics,” Structural Change and Economic Dynamics 4: 39-63.
C in c e r a , M., and V an P o t t e l sb e r g h e , B. (2001), “ International R & D Spillovers: A
Survey,” in M. Cincera, Cahiers Economiques de Bruxelles 169: 3-32.
C o e , D. I., and H elp m a n , E. (1995), International R&D Spillovers,” European Economic
Review 39: 859-87.
An Aggregate Model o f Technical Change,” Quarterly Journal of
* C o n l i s k , J. (1989),
Economics 104: 787-821.
 IN N O V ATIO N AND ECONOMIC GROWTH 511

C J. (i 977)> Modern Capitalism: Its Growth and Transformation, London: Martin

o r n w a l l ,

Robertson,
Crafts, N. E R. (1985), British Economic Growth During the Industrial Revolution. Oxford:
Oxford University Press,
Denison, E. (1962), The Sources of Economic Growth in the United States and the Alternatives
Before Us, Washington: Committee for Economic Development.
-----(1966), Why growth rates differ>Washington: Brookings Institution.
Dixon, R. J., and T h irlw a ll , A. P. (1975), “A Model of Regional Growth-Rate Differences on
Kaldorian Lines,” Oxford Economic Papers n: 201-14,
Dosi, G. (1982), ‘'Technological Paradigms and Technological Trajectories,” Research Policy
11:147-62.
Eaton, J., and K ortum , S. (1999), “ International Technology Diffusion: Theory and Meas­
urement,” International Economic Revim' 40: 537-70.
Fabricant , S. (1942), Employment in Manufacturing 1899-1939, New York: NBER.
Fagerberg , J. (1988a), “ International Competitiveness,” Economic Journal 98: 355-74.
----- (1988b), “ Why Growth Rates Differ” in G. Dosi, C. Freeman, R. R. Nelson, G. Silver-
berg, and L. Soete (eds)., Technical Change and Economic Theory; London: Pinter, 87-99.
----- and V e r s p a g e n , B. (2002), “ Technology-gaps, Innovation-diffusion and Transform­
ation: an Evolutionary Interpretation,” Research Policy 31:1291-304.
Freeman, C , and Lou^A, F. (2001), As Time Goes By: From the Industrial Revolutions to the
Information Revolution, Oxford: Oxford University Press.
----- and Soete , L. (1997), The Economics of Industrial Innovation, 3rd edn., London and
Washington: Pinter.
*------------ (1990), “ Fast Structural Change and Slow Productivity Change: Some Paradoxes
in the Economics o f Information Technology,” Structural Change and Economic Dynamics
1:225-42.
Griliches , Z. (1979), “ Issues in Assessing the Contribution o f Research and Development to
Productivity Growth,” The Bell Journal of Economics 10: 92-116,
----- (1980), “ R&D and the Productivity Slowdown,” American Economic Review 70:343-8.
----- (1984), R&D, Patents and Productivity>Chicago: Chicago University Press.
----- (1986), “ Productivity, R&D and Basic Research at the Firm Level in the 1970s,” American
Economic Review j 6 : 141-54.
* --- (1992), “ The Search for R&D Spillovers,” Scandinavian Journal of Economics 94:
S29-S47.
----- (1996), “ The Discovery o f the Residual: A Historical Note,” Journal of Economic
Literature 34:1324-30.
----- and M a i r e s s e , J. (1984), “ Productivity and R&D at the Firm Level,” in Z. Griliches
(ed.), R&D, Patents and Productivity>Chicago: Chicago University Press, 339-74*
* G r o s s m a n , G. M., and H e l p m a n , E. (1991), Innovation and Growth in the Global Economy;
Cambridge, Mass.: M IT Press.
G r u e b l e r , A, (1990), The Rise and Fall of Infrastructures: Dynamics of Evolution and
Technological Change in Transport, Heidelberg: Physica-verlag.
H e e r t j e , A , (1994 ), “ Neo-Schumpeterians and Economic Theory, in L, Magnusson (ed.),
Evolutionary Approaches to Economic Theory, Dordrecht: Kluwer, 265—76.

* Asterisked items are suggestions for further reading.

512 BART VERSPAGEN

H e l p m a n , E. (ed.) (1998), General Purpose Technologies and Economic Growth, Cambridge,

Mass*: M IT Press.
♦ Jones, С. (1995л), “ R& D Based Models o f Economic Growth,” Journal o f Political Economy
103:759 - 8 4 -
------ (1995 b), “ Time Series Tests o f Endogenous Growth Models,” Quarterly Journal o f
Economics 110: 495-525.
Jorgenson, D. W., and G riliches, Z. (1967), “ Explanation o f Productivity Change,” Review
of Economic Studies 34: 249-83.
♦ Kaldor, N. (1957), “A Model o f Economic Growth,” Economic Journal 67, December:
591- 624.
------ (1970), “ The Case for Regional Policies ,” Scottish Journal of Political Economy 6j: 591-624.
K a t z , B. G., and P h i l l i p s , A. (1982), “ Government, Economies o f Scale and Comparative
Advantage: The Case o f the Computer Industry” in H. Giersch (ed.), Proceedings of
Conference on Emerging Technology, Kiel Institute o f World Economics, Tuebingen: J. C.
B. Mohr.
K e l l e r , W. (1998), “Are International R& D Spillovers Trade-Related? Analyzing Spillovers
Am ong Randomly Matched Trade Partners,” European Economic Review 42:1393-612.
Kleinknecht , A. (1987), Innovation Patterns in Crisis and Prosperity Schumpeter s Long Cycle
Reconsidered, London: Macmillan.
K u h n , T. (1962), The Structure of Scientific Revolutions, Chicago and London: The University
o f Chicago Press.
L e v i n e , R., and R e n e l t , D. (1992), “A Sensitivity Analysis o f Cross-Country Growth
Regressions,” American Economic Review 82: 942-63.
L i c h t e n b e r g , E, and V a n P o t t e l s b e r g h e , B. (1996), “ International R& D Spillovers: A Re-
Examination,” NBER Working Paper 5668.
Los, B., and V e r s p a g e n , B. (2000), “ R&D Spillovers and Productivity: Evidence from U.S.
Manufacturing Microdata,” Empirical Economics 25:127-48.
L u c a s , R. E. B. (1988), “ On the Mechanics o f Economic Developm ent” Journal of Monetary
Economics 22: 3-42.
* M a d d i s o n , A. (1987)7 “ Growth and Slowdown in Advanced Capitalist Economies: Tech­
niques o f Quantitative Assessment,” Journal of Economic Literature 25: 649-98.
------ (1991), “ Economic Stagnation Since 1973, its Nature and Causes: A Six-Country Survey,”
De Economist 131: 585-608.
------ (1995)7 Monitoring the World Economy 1820-1992. Paris: OECD Development Centre.
------ (2001), The World Economy: A Millennial Perspective, Paris: O ECD Development
Centre.
M a l e r b a , E, N e l s o n , R., O r s e n i g o , L., and W i n t e r , S, (1999), “ 'History-Friendly Models
o f Industry Evolution: The Computer Industry,” Industrial and Corporate CJiangeS: 3-40.
M c C o m b i e , J. S. L., and T h i r l w a l l , A. R (1994), Economic Growth and the Balance-
of-Payments Constraint London: St. Martin s Press.
M e n s c h , G. (1979), Stalemate in Technology: Innovations Overcome Depression, Cambridge:
Ballinger.
M o r r i s o n , C. J. (1986), “ Productivity Measurement with Non-Static Expectations and
Varying Capacity Utilization,” Journal of Econometrics 33: 51-74.
N M. I. (1970), “ Some Approaches to the Theory and Measurement o f Total Factor
a d i r i ,

Productivity: A Survey,” Journal of Economic Literature: 1137-77.

N R. R. (1968), A Dillusion Model o f International Productivity Differences in
e l s o n ,

Manufacturing Industry” American Economic Review 58:1219-48.

 IN N O V A T IO N AND ECONOMIC GROWTH
513

-— (1973). "Recent Exercises in Growth Accounting; New Understanding or Dead End?”

American E co n om ic Review 63: 462-8,
----- (1981)* "Research on Productivity Growth and Productivity Differences: Dead Ends and
New Departures” Journal of Economic Literature 19:1029-64.
* --- and W i n t e r , $. G. (1982), An Evolutionary Theory of Economic Change, Cambridge,
Mass.: Harvard University Press.
*Pa sin e t t i , L. L. (i993)> Structural Economic Dynamics, Cambridge: Cambridge University
Press.
Perez, C. (1983b "Structural Change and the Assimilation of New Technologies in the
Economic and Social Systems,” Futures 15: 357-75.
Romer, P, (1986), “ Increasing Returns and Long Run Growth,” Journal of Political Economy
94:1002-37.
----- (1990), “ Endogenous Technological Change,” Journal of Political Economy 98: S71-S102.
Sahal , D. (1981), Patterns of Technological Innovation, New York: Addison Wesley.
Schumpeter, J. A. (1939). Business Cycles: A Theoretical, Historical and Statistical Analysis of
the Capitalist Process, New York: McGraw-Hill.
Shell, K. (1967) "A Model o f Inventive Activity and Capital Accumulation,” in K. Shell (ed,),
in Essays on the Theory of Optimal Growth, Cambridge, Mass,: MIT Press, 67-85.
Silvers erg , G. (2001), “ The Discrete Charm o f the Bourgeoisie: Quantum and Continuous
Perspectives on Innovation and Growth,” Research Policy 31:1275-89.
* --- and V e r s p a g e n , B, (1994), “ Learning, Innovation and Economic Growth: A Long-
Run Model o f Industrial Dynamics,” Industrial and Corporate Change 3:199-223,
----------- (1994), “ Economic Growth and Economic Evolution: A Modeling Perspective,” in
E Schweitzer and G. Silverberg (eds.), Selbsorganisation. Jahrbuch fiir Komplexitat in den
NaturSoziaT und Geisteswissenschaften, Berlin: Duncker 8t Humblot, 265-96.
----- (i957)> “ Technical Progress and the Aggregate Production Function,” Review of Eco­
nomics and Statistics 39: 312-20,
Solow, R.M. (1956), ‘A contribution to the Theory o f Economic Growth', Quarterly Journal of
Economics, vol. 70, 65-94
Sornette , D., and Z a jd e n w e b e r , D. (1999), “ The Economic Return of Research: The Pareto
Law and its Implications,” European Physical Journal B 8(4): 653-64.
S u m m e r s , R., and H e s t o n , A. (1991), “ The Penn World Table, Mark 5: An Expanded Set of
International Comparisons 1950-1988,” Quarterly Journal of Economics 6: 361-75.
Temple, J. (1999), “ The New Growth Evidence,” Journal of Economic Literature, March, 37(1):
112-56.
Tinbergen , J. (1943), “ Zur Theorie der Langfristigen Wirtschaftsentwicklung,” Welt-
wirtschaftliches Archiv 55: 511-49.
Uzawa, H. (1965), “ Optimum Technical Change in an Aggregative Model o f Economic
Growth,” International Economic Review 6:18-31.
V a n Pottelsberghe d e l a P o t t e r i e , B. (1997), “ Issues in Assessing the Effect o f Inter­
industry Spillovers,” Economic Systems Research 9:331-56*
V brdoorn, R l (1949), “ Fatten che Regolano lo Sviluppo della Produttivita del Lavoro,
Ulndustria 1:45-53.
V erspagen , B. (1992), “ Endogenous Innovation in Neo-Classical Growth Models. A
Survey” Journal of Macroeconomics 14: 631-62.
— - (1995). “ RM> and Productivity: A Broad Cross-Section Cross-Country Look,” Journal
o f Productivity Analysis 6 :117-35.
----- (1997)? “ Estimating International Technology Spillovers Using Technology Flow Matri­
ces” Weltwirtschaftliches Archiv 133: 226-48.
Y o u n g ,A, ( 1928 ), “ Increasing Returns and Economic Progress,” Economic Journal 38:527-42.
 C H A P T E R 19

INNOVATION AND
CATCHING-UP

JAN FAGERBERG
M A N U E L M. G O D I N H O

19.1 I n t r o d u c t i o n

T h e h isto ry o f cap italism fro m the In d u strial R evo lu tio n o n w ard s is on e o f increas­
in g differences in p ro d u c tiv ity and livin g co n d itio n s across d ifferen t p arts o f the
globe. A cco rd in g to o n e source, 250 years ago the d ifferen ce in in co m e o r p ro d u c ­
tiv ity p er head betw een the richest and poorest c o u n try in the w o rld w as a p p ro x i­
m ately 5 : 1 , w h ile to d ay this difference has increased to 4 0 0 : 1 (L an d es 1998).
H ow ever, in spite o f this lo n g -ru n trend tow ards d ivergen ce in p ro d u c tiv ity and
in com e, there are m a n y exam p les o f (initially) b ack w ard co u n tries that— at different
tim es— have m an aged to n a rro w the gap in p ro d u c tiv ity an d in co m e between
them selves and the fro n tier cou ntries, in o th er w o rd s, to “ catch up.” H o w d id they
do it? W hat w as the role o f in n o va tio n and d iffu sio n in the process? T h ese are am o n g
the q u estion s that we are go in g to discuss in this ch ap ter.1
The “ catch-up” question should be seen as distinct from the discussion o f
“ convergence ” although the two issues partially overlap. “ Catch-up” relates to the
ability o f a single country to narrow the gap in productivity and incom e vis-a-vis a
leader country, while “ convergence” refers to a trend towards a reduction o f the
overall differences in productivity and income in the world as a whole. The issue o f
 I N N O V A T I O N A N D C A T C H I N G - UP 515

convergence has been central to the econom ists research agenda, in part because
some prom inent theoreticians form ulated m odels o f long-run growth im plying
such convergence (Solow 1956),2 O f course, if all countries below the frontier catch
up, convergence w ill necessarily follow, But if only som e countries catch up (and
perhaps forge ahead), w hile others fall behind, the outcom e with respect to conver­
gence is far from clear (Abramovitz 1986), W hat the em pirics show is that, at best,
such convergence is confined to groups o f countries— or “convergence clubs”
(Baumol et a l 1989)— in specific tim e periods. Arguably, to explain such differences
in the conditions for catch-up through tim e, it is not enough to rely on general
mechanisms. Som e historical perspective is required.
During m ost o f the nineteenth century, the econom ic and technological leader o f
the capitalist w orld was the U nited K ingdom , with a G D P per capita that was 50 per
cent above the average o f other leading capitalist countries. However, during the
second h alf o f the century, the U nited States and G erm any both started to catch up
and substantially reduced the U K lead. T h ey did not achieve this growth b y merely
imitating the m ore advanced technologies already in use in the leading country, but
rather did so b y developing new ways o f organizing production and distribution, e.g.
by innovating (Freem an and Soete 1997; Freem an and Lou^a 2001). In the case o f the
US, this led to the developm ent o f a historically new and dynam ic system, based on
mass production and the d istribution and exploitation o f econom ies o f scale.
Germany introduced new w ays o f organizing production, particularly with respect
to R&D in the chem ical and engineering industries, that in the long run w ould com e
to have a very im portan t im pact. M ore recently, the very rapid catch-up o f Japan to
Western p ro d u ctivity levels d u rin g the first h alf o f the twentieth century was
associated with a n u m ber o f very im portant organizational innovations (such as
the Ujust-in-tim e system ” see B o x 19.1) that, am ong other things, totally transform ed
the global car industry, These innovations did not only benefit Japan, but diffused
(with a lag) to the established leader (the U SA ) and contributed to increased
productivity there.
As these b rie f exam ples show, successful catch-up has historically been associated
not merely w ith the adoption o f existing techniques in established industries, but
also with innovation, particu larly o f the organizational kind, and with inroads into
nascent industries. H ow ever, as is equally clear, this has been done in different ways
and with different consequences. I f we extend the perspective to the m ost recent
decades, as we w ill do in this chapter, this diversity in strategies and perform ance
becomes even m ore striking. In the next section, w e discuss som e o f the perspectives
that have em erged in the catching up literature. Section 19.3 extends the perspect­
ive to the m ost recent decades, com pares cases o f successful catch-up to less
successful ones, and considers the lessons that m ay be drawn. Finally, Section 19.4
raises, by w ay o f con clu sion , the question o f what present day developing countries
can learn, particu larly w ith respect to policy, from the literature on innovation and
catching up.
516 JAN F A G E R B E R G A N D M A N U E L M. G O D I N H O

19.2 L e s s o n s fr o m t h e L i t e r a t u r e

T h e p o ten tially relevant literatu re o n w h y grow th d iffers (an d w h y so m e countries

catch u p w h ile oth ers d o not) is arg u ab ly v e ry large. P ro v id in g an o v e rv ie w o f all the
literature w o u ld be b eyo n d the scope o f a sh ort e ssa y H ow ever, the w o rk in this area
fo cu sin g sp ecifically o n catch -u p — as d istin gu ish ed fro m e co n o m ic g ro w th m o re
gen erally— and, in p articu lar, o n concepts an d th eo ries that m a y b e h elp fu l for
u n d erstan d in g catch up (o r lack o f su ch ), is ap p reciab ly sm aller. W e w ill in the
fo llo w in g lim it the d iscu ssion to three central cases w ith in the latter. First, there are
the co n trib u tio n s o f T h o rstein Veblen, A le x a n d e r G ersch en k ro n , an d oth ers on
E u ro p ean catch -u p p rio r to W orld W ar L 3 T h e m ain p o in t o f interest here is the
in terp retatio n o f the G e rm an catch -u p w ith the U K , an d the ro le o f p o lic y and
in stitu tio n s in this context. Secon d, there is a large literatu re o n the A sia n catch-up,
p a rticu la rly Ja p an , b u t in creasin gly also So u th K orea, T aiw an, an d o th er countries
that, to v a ry in g degrees, have attem pted to fo llo w the Jap an ese route. T h e argu m en t
that an activist, “ d evelo p m en tal state” is an efficien t m ean to su ccessfu l ca tch -u p has
been a central fo cal p o in t in m u ch o f this literatu re. T h ird , there is a stran d o f
m a cro h isto rical and m a cro e co n o m ic analysis fo cu sin g o n in terp retatio n s o f long-
ru n data o n e co n o m ic grow th , and o n the role o f tech n o lo g y an d in n o v a tio n in this
context. A central co n trib u to r here has been M o ses A b ra m o v itz . In w h at follow s
next, these three perspectives w ill be b rie fly review ed.

19.2.1 Lessons from European Catch-Up

T h e d iscu ssio n o f co n tin en tal E u ro p e catch -u p illu strates n icely so m e o f the central
issues in the catch -u p literature. Veblen (19 15), w h o in itiated the d iscu ssio n , put
fo rw a rd the argu m en t that recent tech n o lo gical changes altered the c o n d itio n s for
in d u strializatio n in latecom er econom ies. In earlier tim es, he argu ed , the d iffu sio n
o f tech n o lo gy had been h am p ered by the fact that tech n o lo g y w as m o stly em b od ied
in person s, so that m ig ratio n o f skilled w o rk ers w as a n ecessary p rereq u isite fo r its
spread across d ifferent lo cation s. H ow ever, w ith the adven t o f “ m a ch in e tech n o l­
ogy,” this lo gic ch an ged (ibid. 19 1). In contrast to the co n d itio n s that h ad prevailed
p reviou sly, Veblen argu ed, this n ew type o f kn ow led ge “ can be h eld an d tran sm itted
in d efin ite an d u n eq u ivo cal shape, and the acq u isitio n o f it b y su ch tran sfer is no
lab o rio u s o r un certain m atter” (ib id .). A lth o u g h Veblen d id n o t use the term in o lo g y
that is n o w co m m o n ly ap p lied to the process he d escribed, it is p retty clear w h at he
had in m in d . Effectively, w h at he w as argu in g is that, w h ile tech n o lo g y w as p re v i­
o u sly “ tacit” an d em b o d ied in person s, it later becam e m o re “ c o d ifie d ” an d easily
tran sm ittable. H ence, catch -u p sh o u ld be expected to be relatively easy, an d w as,
 ______ I N N OVATION a n d C A T C H IN G -U P 517

under “otherwise suitable circ u m sta n c e s"4 largely “ a question o f the pecuniary
inducement a n d . . . opp ortu n ities offered by this new in du stry" (ibid, 192), Since
the latecomers could takeover the new technology ready-m ade, 1 without having to
share the costs o f its developm ent, this m ight be expected to be a very profitable affair
(ibid, 249)* This being the case, Veblen predicted that other European countries, e g
France, Italy, and Russia, w o u ld soon follow suit (he also m entioned the case o f
Japan).
While in Vebleffs interpretation, the G erm an catch-up was a relatively easy affair,
the economic historian A lexan d er G erschenkron (1962) took a different view,
emphasizing the d ifficu lty o f the m atter. W hile, he argued, technology at the time
Britain industrialized w as sm all scale, and hence institutionally not very dem anding,
these conditions were rad ically altered in the nineteenth century when Germ any
started to catch up. W hat G erschenkron particularly had in m ind was the seemingly
inbuilt tendency o f m odern tech nology to require ever larger and m ore com plex
plants (static and d yn am ic econom ies o f scale), with sim ilarly changing require­
ments with respect to the physical, financial, and institutional infrastructure. He
argued that, because o f the high potential rewards from successful entry, and the
heavy transform ation (m odern ization) pressure on the rest o f the econom y it helped
to generate, it was o f p aram ou n t im portance for the latecom er to target such
progressive, dynam ic industries, and to com pete globally through investing in the
most modern equipm en t/plan ts.5 H ow ever, to succeed, catching-up countries had,
in Gerschenkron s view , to build up new “ institutional instrum ents for which there
was little or no cou nterpart in the established industrial cou n try" (1962: 7), The
purpose o f these institutional instrum ents w ould be to m obilize resources to
undertake the necessary changes at the new and radically enlarged scale that m odern
technology required. H is favorite exam p le6 was the G erm an investm ent banks (and
similar examples elsew here in Europe), but he also adm itted that, depending on the
circumstances, other types o f institutional instrum ents, such as, for instance, the
government (in the R ussian case),7 m ight conceivably perform the same function.
Gerschenkroffis w o rk is often identified w ith his focus on the role o f banks in
industrialization, although as pointed out by Shin (1996), it is possible to see it as an
attempt to arrive at a m ore general th eo ry about catch-up, focusing on certain
requirements that m ust be m et fo r successful catch-up to take place, as well as
different, though “ fu n ctio n ally equivalent," institutional responses (or catch-up
strategies). A n im po rtan t chain in G erschenkron’s argum ent is the em phasis on
the advantages o f targeting rap id ly grow ing, technologically advanced industries. It
should be pointed out, how ever, that for h im this was a generalization based on
historical evidence. T h u s it is not o bviou s that his recom m endations w ould be
equally relevant for later tim e periods/technologies. N either did he rule out that
there might be other paths to successful industrialization than the one he recom ­
mended, although he held that to be rather exceptional. For instance, he pointed to
Denmark as an exam ple o f a co u n try that m anaged to catch up w ithout targeting the
5l 8 I A N F A G E R B E R G A N D M A N U E L M. G O D I N H O

p ro gressive in d u stries o f the tim e, an d exp lain ed this w ith its close lin k s to the
rap id ly gro w in g B ritish m arket fo r ag ricu ltu ral p ro d u cts.
W e m a y use Veblen s an d G ersch en k ro n s accou n ts o f the G e rm a n catch -u p to
m ake a p re lim in a ry classificatio n o f catch -u p strategies. T h e ty p e d escribed by
Veblen assum es that tech n o lo g y is easily available/transferable, n o t v e ry d em an d in g
in term s o f skills o r in frastru ctu re, and that m ark et forces are able to take care o f the
necessary c o o rd in a tio n w ith o u t the large-scale in vo lvem en t o f extern al “ change
agents/’ In contrast, there is the G ersch en k ro n ian case in w h ich te c h n o lo g y tran sfer
is so d em an d in g in term s o f skills/in frastru ctu re that m ark et fo rces, i f left alon e, are
co n sid ered u n lik ely to lead to success, an d so so m e degree o f active in terv en tio n in
m arkets b y ou tsid ers, w h eth er private o rg an izatio n s o r p arts o f go vern m en t, is
d eem ed necessary.

19.2.2 The Asian Experience

S im ila r perspectives to those o f Veblen and G e rsch en k ro n have also p layed a role in
the d iscu ssion s o f the A sian catch -u p in the p o st-W orld W ar II p erio d . T h e p rim a ry
exam p les are, in ad d itio n to Jap an , the cases o f So u th K o rea, S in g ap o re , an d Taiwan.
A lth o u g h so m e o bservers have attem pted to classify these as V eb len -typ e catch-up
stories (W orld B a n k 1993), there is b y n o w an ab u n d an t literatu re sh o w in g that the
catch -u p strategies app lied are m u ch closer to the G ersch e n k ro n ia n sch em e (Jo h n ­
so n 1982; A m sd en 1989; W ade 1990; Shin 1996).
T h ere are m a n y accounts o f the Jap an ese catch -u p. T h e so -called M e iji restoration
in 1868 p ro vid es a n atu ral startin g p oin t. W hat h ap p en ed in 1868 w as that a fraction
o f the ru lin g elite established a n ew regim e, w ith the exp licit p u rp o se o f
stren gth ening the e co n o m y an d the m ilita ry strength o f the state, w h ich at the
tim e w as stron gly challenged b y W estern im p erialism (B easley 1990). “A rich society
an d a stron g a rm y ” w as the slogan o f the day. Since Ja p a n lacked o th er “ m o d e rn iz a ­
tio n ” agents, the go vern m en t (b u reau cracy) to o k on the challenge. It m odern ized
the legal system , the physical in frastru ctu re, an d the ed u catio n al system , in iti­
ated new businesses (that later on w ere privatized) in in d u stries that w ere deem ed
strategically im p o rtan t, etc.8 U n iversities, colleges, an d research centers w ere also
fo u n d ed , often w ith a bent tow ard s en gin eerin g and app lied science. W hile, p a rtic u ­
larly in the initial phase, the pu blic sector played a vital role, p riva te in itiatives and
co o p eratio n betw een pu b lic an d p rivate actors becam e g rad u a lly m o re im p o rtan t.
M u ch o f the in itiative cam e to rest w ith a n u m b e r o f em erg in g fam ily-o w n ed
business grou p s, the Z aib atsu s, in in teraction w ith the b u rea u cracy an d the m ilitary.
Initially, the d o m in an t in du stries w ere fo o d processin g and textiles, but, d u rin g
W orld W ar I and the p erio d that fo llo w ed , the Jap an ese e c o n o m y u n d erw en t a rapid
tran sfo rm atio n , w ith m ach in ery and other “ h eavy in d u stries” tak in g o ver as
 IN N O V A T IO N AND CATCH IN G-UP
519

leading sectors. R & D activity also flourished, partly for m ilitary needs, and was,
according to one source (Odagiri and Goto 1996), well above 1 percent of GDP in the
early 1940s.
The defeat o f Jap an in W orld W ar II changed the pow er structure in Japanese
society by eliminating two of the three contending power centers, the military and
the (owners of the) Zaibatsus, hence giving a boost to the bureaucracy that once
more took on the challenge of gearing the economy and the society at large towards
economic catch-up with the West. The sequence of events from the late nineteenth
century somehow repeated itself, with a very important role for the state (and—in
particular—the Ministry for Trade and Industry, MITI) in the early phase, and a
growing role for private initiatives (and business groups) as the economy grew
stronger (with no role left for the military). The new business groups that emerged,
the Keiretsus, were in som e cases based on the pre-w ar groupings (which had been
dissolved b y the A m erican o ccu p atio n forces), w hile in other cases they were totally
new constructions. T h is reorientation differed from the pre-w ar groupings in
respect o f a stronger role fo r banks, and a sm aller role for private investors/family
ownership. In the early phases, the banks were very dependent on credit from the
state, reinforcing the power of the bureaucracy in getting business to cooperate with
the government in its preferred catch-up path. A fter a few decades o f rapid growth,
the banks (and business m o re generally) grew m ore independent, and the role o f the
state diminished and to o k on m ore “ norm al,” W estern proportions.
The exact role o f the govern m en t versus private actors in the various phases o f
Japanese econom ic grow th is a m atter o f considerable controversy, and we shall not
attempt to resolve it here. Suffice it to say that governm ent/bureaucracy interven­
tion, through activist econom ic, industrial, and trade policy (protectionism ), was
very im portant, especially in the early phases. A lthough not everything it touched
turned in to 41gold,” and som etim es its interventions were strongly resisted by private
business (and for perfectly go od reasons), there is no doubt that it contributed
significantly to focusing the attention o f private business to catch-up with the West.
An important elem ent in this catch-u p process (and the policies that were pursued)
was a very rapid but ord erly process o f structural change, through which industries
“of the past” were grad u ally phased out in favor o f technologically m ore progressive
industries, em phasizing in particu lar the com bination o f econom ies o f scale,
product differentiation, and rap id ly grow ing dem and, on the one hand, and con­
tinuous im provem ents o f produ cts and processes through learning, on the other. In
this way, Japanese in d u stry soon rose to the pro d u ctivity frontier in its chosen fields,
first in the steel in d u stry and in sh ip-bu ild ing, and later in cars and (consum er)
electronics.9 A lth ou gh Japan ese in n ovation in the catch-up phase also included a
large num ber o f pro d u ct in n ovation s, especially o f the m in or type (adaptations to
demand), the m ain em phasis w as on process innovations, particularly o f the organ­
isational type, that allow ed for sim ultaneous exploitation o f scale econom ices and
flexibility, leading to high throu gh -pu t, efficient inventory m anagem ent, high
520 JA N F A G E R B E R G AN D M A N U E L M. G O D I N H O

q uality/reliability, and a p ro ven ab ility to ad ju st to the n eed s o f the en d -u ser

(see B o x 19 .1).
N o t su rp risin gly, the Jap an ese exp erien ce generated a lo t o f in terest in other
d evelo p in g cou n tries, p a rticu la rly in A sia, that co n sid ered the p o licies an d practices
p u rsu ed as a possib le m o d el fo r their o w n catch -u p to w ard s W estern levels. The
p rim e exam p les are, as noted, So u th K orea, S in gap o re, an d T aiw an , alth o u gh the
Jap an ese in fluen ce is also recognizable in oth er cou n tries. F o cu sin g in p a rtic u la r on
the form er, w h at these co u n tries have in co m m o n is that th ey h ave cau g h t u p very

Box 19.1 Organizational innovation in Japan

Henry Ford allegedly once said that a customer could get a car in any color he wanted as
long as it was black. This was the quintessential logic o f the Am erican system o f
manufacturing, based on standardized products, produced in long series for mass
consumption, by low-skilled (often immigrant) labor, controlled by a hierarchy o f
foremen, engineers, and managers.
The attempt to adapt this system to Japanese conditions after World War II led to
important modifications. First, the Japanese market was much smaller, so critical mass
could only be reached through exploiting demand diversity. Second, the Japanese labor
force was well educated, trained, and culturally homogeneous, and the differences in
status and pay between blue and white collar workers small. As a result of these
differences, the production system that evolved in Japan came to look very different
from that o f the USA (Freeman 1 9 8 7 ).
The kanban or “just-in-time” system, developed by the Japanese auto industry,
combines the advantages of mass production with flexibility in adjusting to changes
in the composition and level of demand (Aoki 1 9 8 8 ). What is going to be produced
(and when) is decided by the part of the firm close to the end users (market). Orders
are placed on a daily basis at the firm's production units, which have to deliver the
requested products “just-in-time.” This also holds for suppliers of parts, and the
system is referred to as the “zero inventory” method. However, “zero inventory”
implies that defect parts cannot be tolerated (because otherwise production would
be halted). To increase quality and eliminate defects, organizational practices such as
“total quality control,” originally borrowed from US industry, and “ quality circles”
were introduced. Eventually, a new organization of work emerged, with workers
rotating through different tasks and a much greater role for the individual worker
(and work-team) in surveying production and quality, than what was common in the
US auto industry. This new organization of work also meant more competent, com­
mitted, and motivated workers.
Important efficiency improvements stemmed from these organizational innov­
ations. By the late 1 9 8 0 s, Japanese manufacturing, particularly in the car industry,
was unrivalled in its efficiency (Womack et a l 1990). The time needed to produce a car
in Japan in 1989 was 16.8 hours, while the equivalent figures for the US and Europe
were, respectively, 25.1 and 36.2 hours (The Economist, 17 October 1992).
 IN N O VA TIO N AND CATCH IN G-UP 52!

rapidly, undergone extensive structural change, and— finally— established them ­

selves as am ong the m ajor producers (and exporters) internationally in the most
technologically progressive in d u stry o f the day, electronics (broadly defined). The
government appears to have played a very im portant role in these processes.10
Everywhere, em phasis has been placed on the expansion o f education, particularly
that o f engineers (ball 2000). In the early phases, governm ents in South Korea and
Taiwan intervened h eavily w ith ta riff protection, quantitative restrictions, financial
support, etc. to benefit the grow th o f indigenous industries in targeted sectors.
Singapore is a special case, since its governm ent has relied heavily on inward Foreign
Direct Investment (FD I) in its industrialization efforts, and thus targeting has had to
be achieved through selective FD I policies (Lall 2000). In all countries, targeting
production for exports and rew arding successful export perform ance was very
important. M ore recently, all countries have placed em phasis on policies supporting
R&D and innovation. H ow ever, w hen it com es to the industrial structure, there Is a
considerable elem ent o f diversity. In South Korea, large, diversified business groups
(chaebols), sim ilar to the fam ily-ow n ed groups in pre-w ar Japan, have been and
continue to be very im portan t, w hile in Singapore, foreign m ultinationals dom inate
the scene. Taiwan, b y contrast, is characterized b y an industrial structure dom inated
by small and m edium -sized private firm s.
With respect to the G erschenkron ian schem e, the experiences o f these Asian
economies fit well w ith the em phasis on targeting the technologically most progres­
sive industries. On a general level, in all fo u r countries, the state (bureaucracy) has
played a very im portan t role in the early stages. However, as noted above, this has
been achieved in different w ays in different countries. For instance, in both Japan
and South Korea, credit ration in g b y the state (so-called “ directed credit” ) was
extensively used to persuade private business to go along with the governm ent’s
objectives, while this m echanism played virtu ally no role in Taiwan (which under­
went a financial liberalization early on). In the Taiwanese case, the governm ent had
to rely on other instrum ents such as state-ow ned firm s (which came to play an
important role) and, in particular, heavily supported “ interm ediate institutions”
(R&D infrastructure etc.) w ith m ixed public/private sector participation. M oreover,
while industrialization in Japan , and in the U SA and G erm any before it, was
mainly geared tow ards the h om e m arket, exports played a sim ilar role in the
catch-up strategies o f the three “ tigers.” T his may, arguably, have to do with
the fact that the dom estic m arkets in the latter were in m any cases too small to
support large-scale indu strialization efforts, but the gradual reduction in barriers
to trade during the post-W orld W ar II period also played an im portant role
(Abramovitz 1994).
The freeing o f international capital flow s, and the deregulation o f financial
markets towards the end o f that period , placed the “ tigers and other late-latecom ers
in a somewhat different situ ation from that o f Japan 50—100 years earlier, with a
greater potential role fo r external finance, whether in the form o f FD I or lending.
522 JAN F A G E R B E R G A N D M A N U E L M. G O D IN H O

F o r instance, w hile Jap an ese catch -u p w as largely self"fin an ced , the S o u th K orean
catch -u p cam e to d epend h eavily on fo reign len d in g. H ow ever, su ch increased
debt exp o su re, w h ile p ro v id in g o p p o rtu n ities fo r catch -u p , m a y also m ak e countries
vu ln erab le, as sh ow n b y the fin an cial crisis in K o rea (an d to so m e extent in
o th er A sia n cou n tries) tow ard s the end o f the 1990s (see B o x 19 .2). A lth o u gh
co n tro versial (Sh in an d C h an g 2003), the crisis m a y also be seen as an illu stration
o f the im p o rtan t p o in t that p o licies an d in stitu tion s that w o rk ed w ell d u rin g the
catch -u p phase m a y n ot be equ ally w ell suited w h en this ph ase is co m p leted and
the fo rm er catch -u p c o u n try has to co m p ete w ith o th er d evelo p ed co u n tries on an
equal fo o tin g. A n o th er exam p le o f this, also fro m the fin an cial sector, com es
fro m Jap an . T h e Jap an ese fin an cial system w as d esign ed to generate large savings
a m o n g the general p u b lic and to fu n n el these to the large in d u strial con glom erates,
w h ich w ere the van g u ard s o f the catch -u p process, on p referen tial term s, and
as such it w as v e ry effective. H ow ever, w h en the catch -u p w as co m p leted , this
fin an cial m ach in e co n tin u ed to generate large savin gs, even th o u g h the profit
o p p o rtu n ities created b y the p o ten tial fo r catch -u p w ere largely gon e. T h is led
to excesses, crises, and d ep ression . H ence, fro m b ein g a v e ry v a lu a b le asset,
the c o u n try ’s fin an cial system actu ally tu rn ed into a co n sid erab le b u rd e n for the
Jap an ese econom y.

Box 19,2 The financial crisis in Korea

South Korea went through a deep financial crisis in 1 9 9 7 - 8 . A factor contributing to

this was the, by international standards, very high debt-exposure of Korean chaebols,
which had to do with the way catch-up in Korea traditionally had been financed (as in
Japan, through loans, often by state-controlled banks on concessionary terms, rather
than equity). As long as international capital flows were subject to strict government
control, as occurred in most of the post-World War II period, the system may have
been said to fulfill its purpose (rapid catch-up). But when these restrictions eased,
several Korean chaebols and financial institutions exploited their new freedom to
increase their financial exposure and, as a result, substantially increase the country’s
national debt. This paved the way for the crisis. To resolve it, the Korean government
went into a settlement with the IMF, involving, among other things, expectations o f
extensive "structural reforms/5 intending to bring the Korean system closer to the
Anglo-American model (see Shin and Chang ( 2003 ) for an extended discussion). The
crisis was relatively short-lived, as the Korean economy underwent a rapid recovery in
1999. It had some repercussions elsewhere in Asia, although Taiwan and Singapore
were much less exposed, simply because debt financing and foreign lending did not
play the same prominent role as in Korea.
 IN N O VA TIO N AND CATCHING-UP 523

19.2.3 A Macro-View
The third strand of catch-up research mentioned above operates on the macro-level
and asks questions of the extent to which the catch-up or convergence actually
occurred, for whom and how this may be explained. As mentioned in the introduc­
tion, an important finding in this literature is that the long-run trend since the
British Industrial Revolution points to divergence, not convergence, among capital­
ist economies. It has also been shown that these trends differ considerably between
time periods. For example, one such period in which the conditions for catch-up
appear to have been especially favorable (and during which many countries man­
aged to narrow the gap in productivity and income vis-a-vis the leader) is comprised
of the decades following the end of World War II, what Abramovitz (1986,1994) has
called “the post-war catch-up and convergence boom.” He suggested that such
differences in performance over time and across countries might, to some extent,
be explained with the help of two concepts, technological congruence and social
capability. The first concept refers to the degree to which leader and follower country
characteristics are congruent in areas such as market size, factor supply, etc. For
instance, the technological system that emerged in the USA around the turn of the
century was highly dependent on access to a large, homogeneous market, something
that hardly existed in Europe at the time, which may help explain its slow diffusion
there. The second concept points to the various efforts and capabilities that develop­
ing countries have to develop in order to catch-up, such as improving education,
infrastructure and, more generally, technological capabilities (R&D facilities etc.).
Abramovitz explained the successful catch up of Western Europe in relation to the
US in the first half of the post-war period as the result of both increasing techno­
logical congruence and improved social capabilities. As an example of the former, he
mentioned the manner in which European economic integration led to the creation
of larger and more homogeneous markets in Europe, facilitating the transfer of
scale-intensive technologies initially developed for US conditions. Regarding the
latter, he pointed to, among other things, such factors as the general increases in
educational levels, the rise in the share of resources devoted to public and private
sector R&D, and the ability of the financial system in mobilizing resources for
change.
There have also been attempts to develop testable models of cross-country
differences in growth performance (or productivity) that includes the potential
for catch-up as one of the explanatory factors. Classical papers on the subject are
Nelson (1968) and Gomulka (1971). In a highly innovative contribution, Cornwall
(1977) analyzed economic growth in the first half of the post-World War II period as
driven by catching-up processes, the ability to mobilize resources for change (invest­
ment), demand, and endogenous technological change (through the working of the
so-called “Verdoorn’s law” ). Baumol et al. (1989) presented and tested a model of
524 JAN F A G E R B E R G A N D M A N U E L M. G O D I N H O

cro ss-c o u n try grow th , including the potential for catch-up and social capability
(p ro xied b y e d u catio n ), fo r a large n u m b er o f co u n tries an d d iffe re n t tim e-span s,
and since then there has been a p leth ora o f su ch exercises c o n firm in g (o r q u estio n ­
ing) the im p o rta n ce o f such factors (see F agerb erg 1994 an d T em ple 1999 for
overviews)* H ow ever, m ost o f these studies have ig n o red Abramovitz’s em phasis
on the im p o rtan ce o f tech n o logical co n gru en ce as w ell as the role o f innovation*
M o re to the latter, Fagerb erg (19 8 7,19 8 8 ) has suggested an e m p irica l m o d e l b ased on
Sch u m p eterian lo gic that inclu des in n o va tio n , im itatio n , an d o th er effo rts related to
the co m m ercial exp lo itatio n o f tech n o lo g y as d riv in g forces o f growth* F o llo w in g
this ap p ro ach , catch -u p o r convergence is b y no m eans gu aran teed , as it dep en d s on
the balan ce o f in n o va tio n and im itatio n , h o w ch allen gin g these activities are, and the
extent to w h ich co u n tries are eq u ip p ed w ith the n ecessary cap ab ilities. A cco rd in g to
V erspagen (19 9 1), w h o im plem en ted sim ilar ideas into a n o n -lin e a r setting that
allow s fo r b o th catch -u p an d a “ lo w -g ro w th trap,” p o o r co u n tries w ith a lo w “ social
ca p a b ility ” are the ones at risk o f b ein g “ trap p ed ”
A b r a m o v itz s w o rk has been criticized b y Sh in (1996) fo r n o t b ein g sufficiently
h isto rica lly specific. In p articu lar, he argu es that the “ so cial c a p a b ility ” concept is
v e ry d ifficu lt to o p eratio n alize, a fact ad m itted b y A b ra m o v itz h im self. H ow ever,
A b ra m o v itz ’s em ph asis on tech n o logical co n gru en ce d e a r ly p o in ts to an awareness
o f the im p o rtan ce o f changes in tech n o logical d yn am ics o ver tim e, alth o u gh it is
clear that he h im se lf did little to su bstantiate it. In this he d e a r ly sided w ith
G ersch en kron , w h o also focused alm ost exclu sively o n ca tch -u p in scale-based
tech nologies. T h ere is, how ever, no scarcity o f c o n trib u tio n s that argu e that the
d yn am ics o f the scale-based system is w a n in g (N elso n an d W righ t 1992; Fagerb erg et
al. 1999)* I f so, this m a y have stron g im p lica tio n s fo r the c o n d itio n s fo r catch -u p . We
retu rn to this issue in the final section o f this chapter.

19.3 C a t c h i n g u p : A R e v i e w of
Recent Evidence

We w ill n o w take a closer lo o k at the glo bal catch in g -u p p ro cess (o r lack o f such)
d u rin g the fo u r last decades. W hile oth er studies have p ro v id e d h ig h ly aggregated
analyses o f differences in grow th across large sam p les o f co u n tries (fo r overview s, see
Fagerb erg 1994 an d Tem ple 1999), w e w ill in this section lim it the analysis to a
selection o f cou n tries that we find p a rticu la rly relevant fo r the stu d y o f such
processes, and fo r w hich g o o d data on relevant factors, such as R&D an d innovation,
are available. This includes the cou n tries discussed so far, such as the p re v io u s— and
 IN N O V A T I O N A N D C A T C H IN G - UP 52.5

present world leaders (U S and U K ) and Gerschenkrons favorite object o f study,

Germany (we also include data for two other countries he studied, France and Italy,
for com parison), M oreover, we include a group o f Asian countries, which, in
addition to those m entioned so far, also contains China, H ong Kong, India, M alay­
sia, and the Philippines. Finally, we introduce two other country groupings, with
which the experience o f the catch-up countries o f Asia m ayb e com pared, a group o f
European catch-up countries (Finland, Greece, Ireland, Portugal, and Spain) and a
group o f potential catch-up countries from Latin A m erica (Argentina, Brazil, Chile,
and M exico).
Table 19.1 ranks the countries in o u r sam ple by initial G D P per capita level (i960).
The countries that industrialized a century or m ore ago are, not surprisingly, at the
top, headed by the U SA , w hile the seven countries at the bottom o f the list are all
Asian. In the m iddle, we find the rem aining European and Latin A m erican coun­
tries, joined by two A sian econom ies, H ong Kong and Japan. Figure 19.1 illustrates
the way in which these changes in the distribution during the last four decades came
along. Evidently, there is a group o f Asian countries that have caught up very rapidly.
In fact, the top seven perform ers in term s o f per capita grow th are all from Asia.
Annual per capita grow th for these countries ranges from 6.5 per cent (South Korea)
to 4.2 per cent (Japan ). The European catch-up countries follow, headed by Ireland
(4.1 per cent) and Portugal (3.9 per cent). The m ore established countries, which
were in the lead in the early 1960s, cluster towards the low er h a lf o f the distribution,
with growth-rates in the 2 -3 per cent area. At the very bottom , we find three potential
“catch-up” countries that have experienced very dism al perform ance, the Ph ilip­
pines, Argentina, and M exico, each o f w hich evidently has “ fallen behind/' to use the
terminology suggested b y A bram ovitz. The rem aining potential “ catching-up”
countries o f Latin A m erica, plus India, although perform ing slightly better, also
failed to reduce the gap v is-a -v is the leader, the U SA . The result o f this dynam ic is
that, while the seven A sian countries at the top o f the list all im proved their relative
positions during this period, by m o vin g to a higher quartile o f the distribution, all
the Latin A m erican countries m oved dow n one o r m ore quartiles (Table 19.1).
We will now explore the m anner in which these differences in perform ance relate
to differences in relevant “ social capabilities.” A lthough, potentially, there m ay be
many variables o f Interest, we have in the present context chosen to lim it the
discussion to three that we believe are o f particular relevance, these being skills
(education), R8cD, and in n ovation (as reflected In patents). Traditionally, m any
analyses o f differences in cross-co u n try grow th, based on large cross-country
samples, have focused on differences in the extent o f prim ary and secondary educa­
tion as a possible factor behind the observed differences in perform ance (see e.g.
Baumol et al. 1989). H ow ever, w hile relevant for understanding the failures o f som e
developing countries, in Sub-Sah aran A frica for instance, to enter the catch-up
phase, these variables discrim inate less well between the countries in our sample,
which, with very few exceptions, all have relatively extensive primary and secondaiy
>26 JAN F A G E R B E R G A N D M A N U E L M. G O D I N H O

Table 19.1 Income groups, 1960- 1999 (GDP per capita, 10 3 $US, 1990 constant
PPPs)

I960 GDPpc 1999 GDPpc

1st Guartile US 11.3 us '28.1 " -v'

(West) Germany 10.1 Japan 21.0
UK 8.6 Singapore 20.7
France 7.5 France 20.1
Finland 8.2 Hong Kong 19.9
Italy 5.9 Ireland 19.7

2nd Guartile Argentina 5,6 UK 19.2

Chile 4.3 Finland 19.1
Ireland 4.2 (Unified) Germany 19.0
Japan 3.9 Italy 18.2
Spain 3.4 Taiwan 16.6
Mexico 2.2 Spain 14.6

3rd Guartile Greece 3.1 Portugal 13.5

Hong Kong 3.1 South Korea 13.2
Portugal 3.0 Greece 11.5
Brazil 2.3 Chile 10.0
Singapore 2.1 Argentina 8.7
Malaysia 1.5 Malaysia 7.7

4th Guartile Taiwan 1.5 Mexico 6.9

Philippines 1.5 Brazil 5.4
South Korea 1.1 China 3.3
India 0.8 Philippines 2.3
China 0.7 India 1.8

Source: Calculations based on Angus Maddison/Groningen Growth and Development Centre and The
Conference Board, Total Economy Database, July 2003, http://www.ggdc.net.

education systems.11 We have, therefore, chosen to focus on third-level education

(universities, colleges, etc.).
Figure 19.2 confirms that the established industrialized leaders, with the USA in a
comfortable lead, place strong emphasis on higher (third level) education.12 But
some catching-up economies also figure relatively high. Finland, for instance, is
 IN N O V ATION AND C ATCH IN G-U P 527

So u th Korea
T a iw a n *
S in g a p o re *
Hong K ong*
M a la y s ia *
C h in a
Jap a n
Ireland
Po rtugal
S p a in
G reece
Finland
Italy
France
B ra zil*
United S ta te s
In d ia *
C h ile *
United K in g d o m
G e rm a n y**
M e x ic o
A rg e n tin a *
P h ilip p in e s*

1 2 3 4 5 6

Figure 19.1 GDP per capita growth 1960-2001

Notes: All calculations based on 1990 constant prices. For countries with (*) the calculation period is 1980-1999. The
German growth rate (**} refer to W est Germany, 1960-1997.

Source: Calculations based on A n gus Maddison/Groningen Growth and Development Centre and The Conference Board,
Total Economy Database, July 2003, http://www.ggdc.net.

second, followed by South Korea. However, one should not

differences, became today, in sharp contras, to the situation thirty “
the great r a r i t y of countries under study ^
education. The increase m higher education is espe у Finland, South
catching-up economies in Asia and Europe, sue as, ^ ^ education
Korea, and Spain. The deviants from this strong emp - (Mexico,
consist of a group of (low-income) Latin Am 1qw levels Qf higher
Brazil, Malaysia, India, and China) that continue
education. and not every element is necessarily
Highereducationis, however, amixedbag, wefocus
equallyessentialfor innovationorcatch-upmtec no gy-
 28 JAN FAGERBERG AND MANUEL M. GODINHO

■ 1995
□ 1965

Figure 19.2 Third level enrollment in relation to age group,

2 0 -2 4 years old (1965-1995)
Source: UNESCO, Education Statistics, various years.

more narrowly on the production of (undergraduate) university degrees in natural

sciences and engineering (as a percentage of 24 year-olds in the population). In this
case, we see a much clearer divide between the countries in the upper half of the
distribution, in which between 6 and 9 per cent of the cohort take such education,
and the countries in the lower half, in which— in all but one case— less than 3 per
cent of the cohort get such degrees. As is evident from the figure, the countries that
place most emphasis on education in natural sciences and engineering are the
 i n n o v a t io n AND C A T C H IN G - U P 529

10 11 12
UK ~27j
Finland ---------- ------------ ----------- ------- ---- --------- ------------;______ ---------
—---- 30,01
Korea ■---------- 1-----------1----------- 1-----------,______ ! _ ____ L_ ~ T ,
36,4
1 1 1
Singapore 3 37,8
France 3 79,6
Japan 1 25,6
Taiwan 1 І3 .9
Germany 33,8
Ireland 26,7
Spain 3 19,
US 17,0
Italy 25 ,6
Portugal 17,2
Greece 23,5
Chile 1 26,3
Brazi
Thailand
Mexico
Argentina
China
India
Malaysia

Figure 19.3 Ratio of first university degrees in natural sciences and engineering to
24-year-olds in the population, 1999 (all values in °/o)
Notes:Alt figures for 1999 or most recent year. The numbers to the right of the horizontal bars refer to the percentage share
of "1st University Degrees in Natural Sciences and Engineering*' in "Total 1st University Degrees" (French and Greek
numbers refer only to "lo n g " degree courses, and are therefore not directly comparable to the numbers for other countries).

Source: NSF, Science and Engineering Indicators 2002, http://www.nsf.gov/sbe/srs/seind02/start.htm

developed countries (the early industrializers) and the four Asian “ U” ^ ^ ^ S

cussed in the orevious section, joined by some o f the catch-up economies m Europe
cussed in the previous section, join у , ,f fthe distribution, those
(Finland, Ireland, an d S p ain , in particular). The - th e iess-
with low investments in this area, includes all the Latin American countr^ t h e
« „ p e * c o p i e s of Asia a n d - c t a .0 ih«
catch-up co u n trie s inEurope(PortugalandGreecUi or engineering) in
in the USA, one out of six students graduate 1 ^ ^ singapore two-thirds.
South Korea the equivalent number is one- no/only place strong
Hence, countries such as Korea, Taiwan, an ДР {han most other
emphasis on higher edu cation m general, but g
530 I A N F A G E R B E R G A N D M A N U E L M. G O D I N H O

co u n tries d irect their ed u catio n al investm ent tow ard s typ es o f e d u c a tio n o f p a rtic u ­
lar im p o rta n ce fo r tech n o logical catch -u p (and in n o v a tio n ).
It sh o u ld be n oted , how ever, that there are so m e exam p les o f co u n tries that have
fallen b eh in d despite q u ite su bstan tial investm ents in h igh er e d u ca tio n , fo r exam ple,
in the presen t sam ple, A rg en tin a and the P h ilip p in es. A rg u a b ly , im p o rta n t as
ed u catio n is, w h at m atters fo r grow th in the lo n g ru n is h o w it is p u t in to use, and
the failu re to exp an d the em p lo ym en t o p p o rtu n ities fo r h ig h ly ed u cated la b o u r m ay
se rio u sly im p ed e the p o ten tial gro w th effects fro m in vestm en ts in h ig h e r education.
In fact, o n e o f the reasons the A sia n N IC s m an aged to e xp a n d h ig h er technical
ed u catio n so rap id ly w as the sim ilar rap id increase in e m p lo y m e n t o p p o rtu n itie s fo r
engineers (an d scientists). T h u s, fo r these co u n tries, in d u strial, tech n o logy, and
ed u catio n p o licies w ere co m p lem en ts, n ot substitutes, and the a b ility to c a rry out
these p o licies in a su stain ed and co o rd in ated fash io n p ro b a b ly exp lain s a go o d deal
o f th eir eco n o m ic success. Sim ilarly, attem pts to target h ig h -g ro w th , strategic in d u s­
tries w ith o u t investing su fficien tly in c o m p le m e n tary assets, such as h ig h er ed u ca­
tio n , o r w ith o u t p ro v id in g sufficient incentives fo r tech n o lo g ical u p g ra d in g (a
“ d y n a m ic ” co m p etitive en viro n m en t), are also b o u n d to fail, as the evidence o f
som e co u n tries in, fo r instance, Latin A m erica show s.
O ne im p o rta n t use o f h ig h ly com peten t la b o u r is, o f co u rse, in R & D . Figu re 19.4,
w h ich focuses o n R & D as a share o f G D P, show s that, in the early 1960s, o n ly a few o f
the co u n tries in o u r sam ple, w ith the U S A , the U K , and Fran ce in the lead, devoted a
sign ifican t share o f G D P to R & D activ ities.13 A p art fro m these three co u n tries, and
G e rm a n y and Ja p an , all co u n tries in o u r sam ple d evoted less than 1 p er cent o f G D P
to R & D . Today, the U S A has been replaced b y Ja p a n as the c o u n try that em p lo ys the
largest share o f its in com e o n R & D activities, and the clu b o f h ig h R & D p erfo rm ers
has been enlarged b y a n u m b er o f new m em b ers, w ith S o u th K o rea, Finland,
and T aiw an d eservin g o f p articu lar m en tion . H ow ever, S in g ap o re, Irelan d , and
Italy h ave also increased exp en d itu res o n R & D b ey o n d the 1 p er cent o f G D P level.
T h e rem a in in g cou n tries, in clu d in g those fro m L atin A m erica, m a n y A sian and
m o st o f the catch in g -u p econ o m ies in E u rop e, rem ain lo w R & D p erfo rm ers,
alth o u gh R & D investm ents h ave in several cases increased sig n ifica n tly co m p ared
to the situ ation a few decades ago. D ata on patents reveal a v e ry sim ilar pattern
(Figu re 19.5).
A n o th er in d icato r that is often in vok ed in analyses o f catch in g -u p an d tech n o logy
tran sfer is inw ard fo reign direct investm ents (F D I), on the gro u n d s that th ose w ho
do such investm ents are assu m ed to co n tro l, and are w illin g to share, su p erio r
tech nology. T h e available evidence, how ever, indicates that the d istrib u tio n o f FD I
is h ig h ly skew ed, w ith a d isp ro p o rtio n a tely h igh a m o u n t invested in tw o sm all
econ o m ies, H o n g K o n g and Sin gap ore, and, m o re recently, an d to a lesser extent,
In a n u m b er o f oth er lo w er in com e cou n tries, such as Irelan d , C h ile, M alaysia, and
C h in a. H ow ever, som e o f the m ost successful catch in g-u p eco n o m ies, such as Jap an ,
Taiwan, and Sou th K orea, have received very little in w ard F D L T h is d oes n o t im ply,
 1 N N ОV А ТIО N a n d
С А Г с И I КГ G " U p

и.ь 1 1,5 2 2,5 3

р§
А "Ol
Шщтт
'.У.* '' . :

Figure 19.4 R&D as percentage of GDP, 1960s and 1990s

N o te : Countries with 0 - 1 970 s; Countries with 0 - 1 980s.

So u rce : Calculations based on 0ECDt UNESCO and national statistics.

ofcourse, that these countries did not benefit from “ *e™ f* more efficient—wavs of
h just that they found other— and perhaps equ У

absorbing foreign technology (see Box i9 -3;- Q£ Gerschenkronian

The evidence presented here confirms the re ev _ ^ ffigt have been
іи и е referred to earlier, in the sense th a t1 ^ 'W??'".
successful in catching up, namely South Korea, aiw^ ’ __ ваД І ?
before them), have all— after initially having acquir
 l$2 J A N E A G B R S E R G A N D M A N U E L M. G O D I N H O

U n ite d S t a t e s

Japan

Taiwan*

G e rm a n y **

Finland

France

UK

South Korea

Hong Kong*

Italy

Singapore*

Ireland
*95-01
* 01*8-7
Spain

Greece

Malaysia*

Argentina*

Portugal

Mexico

Chile*

Brazil*

Philippines*

India*

China*

0,01 0,1 1 10 100 1000

Figure 1 9 .5 US patenting per million inhabitants (log scale)

N am : f) For these countries the second period is 1995-1999; Г ) Up to 1989 the data refers to the Federal
Republic of Germany.

Source. U S P T O , P a te n t c o u n t s by c o u n t r y / s t a t e a n d y e a r —A ll p a t e n t s , a ll ty p e s,

h t tp :/ / w w w .u s p to .g o v / w e b / o f fic e s / a c / id o / o e ip / t a f / r e p o r ts .h tm
 INNOVATION a n d c a t c h i n g "UP 533

Box 19.3 How to access foreign technology? The OEM system

Asian catch-up has benefited greatly from technology developed elsewhere, However,
the mechanisms used to tap foreign technology sources differ. One central mechanism!
used extensively by Singapore, is inward Foreign Direct Investment (FDI). By contrast!
Taiwan and especially South Korea relied mostly on a form o f subcontracting, “ Ori­
ginal Equipment M anufacturing” (O EM ). As suggested by Hobday (2 0 0 0 ), OEM
might be seen as an organizational innovation, facilitating learning and technological
upgrading in latecomer firms.
Under an OEM contract, a product is produced according to a customer's specifica­
tions, normally a transnational corporation (TN C ), that markets and sells the product
under its own brand-nam e (such as, for instance, “NIKE” or “ IBM ” ). From the 1 9 7 0 s
onwards, many US and Japanese firms, particularly in the ICT sector, used this
mechanism to contract out their production to Korean and Taiwanese firms. This
allowed the latter to acquire basic producing capabilities in electronics, since the TNCs
normally “ helped with the selection o f equipment; the training o f managers, engineers
and technicians; and advice on production, financing and management. . . . Local
learning was encouraged because the T N C depended on quality, delivery, and price o f
the final output” (H obday 2 0 0 0 ; 1 3 4 ).
As successful OEM arrangements evolved into closer long-term relationships, the
Korean and Taiwanese firm s gradually acquired more advanced capabilities, first in
process engineering and later in product design. This led OEM to evolve into a more
advanced stage, O DM (Own Design and Manufacturing), with a greater emphasis on
R&D. A next step, O BM (Own Brand Manufacturing), occurs if a firm uses the
acquired capabilities to produce and to market products under its own brand-name.
This requires new capabilities in marketing, and very substantial investments in
distribution. Hence, it is a difficult step, but potentially very rewarding, since a lot o f
the value added is generated at this stage. Several Korean and Taiwanese firms have
tried, with mixed success, although a few (e.g. Samsung) have managed quite well.

more traditional activities— aggressively targeted the most technologically progres­

sive industries of the day, in which they today play an important role. This trans­
formation of the economy has been accompanied by extensive investments in higher
education, particularly in engineering and natural sciences, and big increases in the
resources devoted to R&D and innovation. As discussed in the previous section,
proactive governments— and policies— have played an instrumental role in these
processes, though in different ways, reflecting different historical backgrounds and
conditions. However, not far behind these success stories, measured in terms of
economic performance, we have a more diverse group of countries that also have
managed to reduce substantially the gap vis-a-vis the frontier. Some of these, such as
Finland, Ireland, and Malaysia, share the focus on targeting the technologically most
progressive industries of the day (ICT), although with considerable differences
between them with respect to the instruments pursued in achieving that goal.
534 J A N F A G E R B E R G A N D M A N U E L M. G O D I N H O

Others, such as Portugal, Spain, and Greece, have preferred to pursue catch-up
without sim ilarly am bitious goals for changing the industrial structure, and, argu­
ably, with m ore modest results, both in terms o f econom ic perform ance, accum ula­
tion o f skills, and technological capabilities. Still others, such as the Latin Am erican
countries considered here, have failed to invest sufficiently in skills and technological
capabilities, and have as a consequence fallen further behind.

19.4 C a t c h in g -u p a n d P o l ic y

The literature on catch-up processes in Europe, particularly in Germany, led to a

strong focus on the relationship between catch-up, “ institutional instruments,” and
policy. Similarly, the m ore recent idea o f a “ developm ental” state, m odeled on the
experiences o f Japan, has brought increased attention to the im portant role played
by policy in catch-up processes. This led, am ong other things, the W orld Bank to
publish a study on East Asian catch-up in which it sought to emphasize the
advantages o f its so-called “ market friendly” approach, and to dow nplay the role
that interventionist politics had played in the catch-up o f these countries (World
Bank 1993, for rebuttals see Rodrik 1994 and Cappelen and Fagerberg 1995). H ow­
ever, the discussion o f catch-up and policy arguably has older roots (Chang 2002).
Two hundred years ago, when the Americans began to consider how to reduce the
gap vis-a-vis the U K, this issue was lively debated. Some, basing themselves on the
great authority o f Adam Smith, argued that the best thing would be to practice free
trade, refrain from governmental intervention in econom ic affairs, and stick to
Am erica s acquired advantage in agriculture. Others, such as the first Secretary o f
the US Treasury, Alexander Hamilton, doubted the wisdom in this approach, and
advocated an industrialization policy based on so-called “ infant industry protec­
tion.” The German economist Friedrich List, who came to be known as the chief
protagonist o f this approach, pointed out that Britain— the m ain advocate o f free
trade at the time— had itself used infant industry protection intensively during its
rise to econom ic and technological leadership (List 1841). Its m ore recent advocacy
o f free trade, List argued, was sim ply an attempt to “ kick away the ladder” the
country itself had used to industrialize,
Since that time, the issue o f catch-up and policy has been highly contentious.
Recently Chang (2002) has taken a fresh look at the evidence on this issue. Based on
an extensive overview o f policies carried out in the industrialization phase o f various
(developed) countries, he demonstrated that the interventionist policies applied by
Japan and other Asian countries during their catch-up were not historically unique.
On the contrary, most (though not all) present-day developed countries applied
 such policies when in the same situation, Many different policy instruments were
used to support the growth of new industries, trade protection (tariffs, etc.) being
only one, and not always the most important. However, during the last few decades,
Chang points out, there has been a concerted effort, led by the USA, with inter­
national organizations such as the World Bank, the IMF, and the WTO as central
players, to reduce the room o f maneuver for such interventionist politics by catch­
ing-up countries.14 He argued, as List before him, that these efforts maybe seen as an
attempt by the present economic and technological leaders to “kick away the ladder”
their own countries used to arrive at their present levels of development.
Do the changes in international rules and regulations during the last few decades,
making certain types o f policies (or practices) previously applied by catching-up
countries more difficult (or even impossible) to pursue, imply that catch-up is
becoming progressively more difficult? This is an important question that deserves
a place on the research agenda in this area. But it needs to be emphasized that what is
a suitable policy nowadays depends not only on the characteristics of the policies
that seemed to work well in the past, but also on the economic, technological,
institutional, and social context today (which may be quite different from those of
previous times). However, a cautionary look at the empirical evidence from the
last few decades suggests that many developing countries have found it increasingly
difficult to exploit the potential for catch-up. This reading of the evidence is also
confirmed by a recent empirical study by Fagerberg and Verspagen (2002), which
found that the conditions for catch-up have become more stringent over time, with
ever-greater demands on the technological capabilities and innovative efforts of
countries striving to narrow the gap vis-a-vis the frontier. While in the 1960s and
1970s, the main factors supporting catch-up were found to be capital accumulation
and a sufficient manufacturing base, in the 1980s and 1990s, the accumulation of
technological capabilities and specialization in services were shown to be more
relevant. These findings indicate that what has happened cannot be explained solely
by changes in institutions and policies, but also has to do with a shift in the
underlying technological conditions, an area clearly in need of further research.
Fagerberg and Verspagen suggested that the observed shift in the conditions for
catch-up “may be a reflection o f the radical technological change in the last decades,
with ICT-based solutions substituting earlier mechanical and electromechanical
ones, and the derived change in the demand for skills and infrastructure (2002:
1303). This is, of course, the kind o f change that Abramovitz hinted at with his
concept of “ technological congruence.1' Following this, one might hypothesize that,
compared to the situation three or four decades ago, the progressive technologies
have become less “ congruent15with the economic conditions (particularly skill-base
^ d R8cD infrastructure) that prevail in many developing countries. In fact, as
shown in the previous section, today only countries that have invested massively
in the formation o f skills and R & D infrastructure seem to be able to catch up (while
those that have not fall further behind).
53^ JAN F A G E R B E R G AND M A N U E L M. GODINHO

A weakness o f m uch o f the existing discussion on catch-up and policy has been an
excessive focus on the policy level (government) at the expense o f the recipients o f
these policy initiatives, e.g. the firms o f the potential catching-up country. As
pointed out by Teece (2000:124): “ If firms are indeed the instrum ents o f develop­
ment, the study o f econom ic development cannot take place separate from the study
o f the theory o f the growth o f the firm ." We suggest this as an im portant area for
further research, theoretical as well as applied. Although an extensive treatment o f
the role o f firm s in catching-up processes is beyond the scope o f this chapter, we will
nevertheless try to emphasize a few points that we believe m ay be useful for further
w ork.15 Research on the role o f firms in innovation and long-run econom ic change
com m only stress that, in most cases, firms only have im perfect knowledge o f the
relevant options in front o f them, and that they tend to be m yopic, searching in the
neighborhood o f their existing competence for relevant inform ation, suggestions,
and solutions (Nelson and Winter 1982; Dosi 1988; Fagerberg Ch. 1 and Lam Ch. 5,
both in this volum e). These characteristics are, o f course, com m on for developed
and developing country firms, but, being far from both the technology frontier and
the potential market, greatly accentuate these problems. Moreover, the developing
country firm m ay be, to a much larger extent than developed country firms,
constrained by its environment: it m ay have a wish (and perhaps even the capability)
to introduce a new product or process, but the possibility to do so m ay depend on
capabilities in other firms or skills that are sim ply not there (or require substantial
investments to occur). Arguably, to avoid being stuck along an inferior path and
never catch up, “ institutional instrum ents" m ay be needed to compensate for some
o f these “ latecomer disadvantages," to use a Gerschenkronian term. In particular,
what the developing country firm m ay need are “ institutional instrum ents" that
improve:
* links with the technology frontier,
* links with markets (and sophisticated users),
* supply o f needed skills, services and other inputs,
* the local innovation system/network.
Arguably, much o f what firms and governments in catching-up countries have done
can be understood from this perspective. For instance, the diversified business-
groups that developed in Japan and South Korea might be seen as “ institutional
instrum ents" fulfilling some o f those needs (Shin 1996). The O EM system (original
equipm ent manufacture, see Box 19.3) that has developed in the electronics indus­
tries o f East Asia m ay also be seen as an “ institutional instrum ent"— or “ organiza­
tional in n ovation ’ (H obday 2000)— geared towards sim ultaneously im proving
links with the technology frontier and the market. Similarly, attracting inward FDI
may be seen as a “ functional equivalent" to OEM , which, however, judged by the
empirical evidence, seems to be less favourable for indigenous innovation. Other,
more demanding, but perhaps also more rewarding ways— since it allows the
 INNOVATION AND C ATC H IN G -U P 537

latecomer firm to reap a larger share of the profit generated— include technology
licensing, investments in own brands (OBM), etc. Improving the supply of needed
skills has, of course, been a central preoccupation of many latecomer governments,
as illustrated in Section 19.3 above. Moreover, we have witnessed sustained efforts by
several latecomer governments in accommodating the needs of firms for a high
quality RM ) infrastructure (innovation system).
What can the extraordinary success that some catching-up countries have had,
and the failure of others, teach present-day developing countries? One important
lesson is that there is no one unique way to successful catch-up that every country
has to emulate. Every country has to find its own way based on an understanding of
(a) the contemporary global technological, institutional, and economic dynamics,
(b) the behavior (and needs) of the relevant agents (of which the firm arguably is the
most important), and (c) the specific context in which the catch-up takes place and
the broader factors that influence it, being economic, technological, institutional,
political, or cultural (Freeman and Lou<;a 2001). There are, one may suggest, big
potential rewards in following the Gerschenkronian strategy of targeting techno­
logically progressive sectors, as part of a broader attempt to transform the economy,
stimulate learning, and enhance the creation of new skills (or assets). However, not
every country is equipped with the necessary capabilities to pursue such a strategy.
For instance, when, in the mid-nineteenth century, Japan initiated its efforts to catch
up with the West, the technology gap vis-a-vis the more advanced countries was
much smaller (compared to what developing countries face today), and its popula­
tion already had a level o f education that compared favourably with most other
countries at the time (Odagiro and Goto 1996). Given that educational standards
have been rising ever since, investing in education may be a good place to start for
countries that have not succeeded in catering to those needs already. For those that
have, there is a wider set of options, and for them some of the experiences outlined in
this chapter may be highly relevant (see Box 194).

Box 19.4 A tale of two countries

Ireland and Portugal are two small countries on the western periphery of Europe. Both
failed, for various reasons, to exploit fully the possibilities for catch-up and industri­
alization during the first half of the twentieth century, and were, therefore, at the
beginning of our period of study, among the poorest in Europe, their economies
Characterized by an industrial structure dominated by traditional, low-skill activities.
However, during the second half of the century, both countries took steps to
integrate their economies with the more dynamic economies of Western Europe, first
through membership of EFTA (1961), and later by joining the EU (1981 and 1986 ,
respectively). As members of the EU, Ireland and Portugal received substantial
economic support through the so-called “ Structural Funds, designed to induce
53$ JAN F A G E R B E R G A N D M A N U E L M. G O D I N H O

technological and economic catch-up* and to facilitate the necessary structural

changes* in poorer areas in the Union (Cappelen et at 2 0 0 3 ). During the 1980 s and
1990 s* both countries grew rapidly* and the gaps in G D P per capita between them and
the more advanced Western economies decreased substantially. In tandem with this*
both economies underwent extensive structural changes, with traditional (“ low-tech” )
activities contracting and more advanced activities expanding* In Portugal, as a result
o f inward FDI, a rapid expansion o f the automobile sector occurred. In addition* one
o f the traditional activities, footwear, underwent substantial technological upgrading,
and managed to expand in spite of raising relative labor costs (Godinho 2000). Ireland,
by contrast, leapfrogged into one o f tod ays most progressive technologies, ICT, almost
exclusively due to inward FDI (O’Sullivan 2002). As a consequence, Ireland has become
one o f the most “ high-tech” economies in the world today, comparable to the Asian
“ tigers,” with an export structure strongly geared towards IC T and so-called “ high-
tech” manufacturing.
W hy did the two countries develop differently? A com m on language, and traditional
strong ties, may have helped to focus the attention o f US multinationals on Ireland.
Another commonality is education. Although Irish educational efforts were not
particularly high by developed country standards, Portugal lagged much further
behind. This was the legacy o f the authoritarian regime o f Salazar, which did not
wish to equip its people with more skills than strictly necessary. As a consequence, two-
thirds o f the Portuguese today have prim ary school as their highest level o f education,
and only one out o f ten has higher education o f some sort. Am ong the OECD
countries, only Turkey is at a similarly low level.
Both countries face challenges today, but these differ, and so will, arguably, the
adequate policy response. In the Irish case, worries have emerged about the weak links
between the foreign-owned IC T firms and the economy at large (in spite of, it might be
argued, a blooming software sector), and the allegedly negligent attitudes of many of
these firms when it comes to developing technological capabilities and undertaking
R&D in Ireland (O ’Sullivan 2002). The Portuguese, naturally, are concerned about the
increased competition for its products stemming from the enlargement of the Euro­
pean Union to many Eastern European countries, many with a better educated labor
force.

N o tes

1. We thank Fulvio Castellacci, Sandro Mendonca and the authors and editors of this
volume for helpful comments and suggestions, retaining sole responsibility for remaining
errors and omissions.
2. More recently, economists have formulated theories (so called “ new growth theory” ) that
do not necessarily predict convergence. See Fagerberg (1994, 2000) and Verspagen, Ch. 18
in this volume, for an extended account.
3* I w instance, although there is a sizeable literature on US economic growth, most of it is
not written from a catch-up perspective, e.g. focusing on what the US experience may tell
us about what other countries should do (or not do) to succeed in catching up. See,
 IN NOVATION AND CATCH IN G -U P 539

however, the recent book by Chang (2002) and the discussion in Section 19.4 of this
chapter,
4. Veblen mentions factors such as the “funds available for investment” ! Veblen 1915:186), a
sufficient supply of “educated men” (ibid. 194 ), as well as a “sufficiently well-instructed
force of operative workmen” (192). The latter, he noted, did not have to be particularly
well educated or trained (188).
5* To the extent that industrialization took place, it was largely by the application o f the
most modern and efficient techniques that backward countries could hope to achieve
success, particularly if their industrialization proceeded in the face of competition from
the advanced country” (Gerschenkron 1962; 9).
6. Surprisingly, perhaps, he did not (nor did Veblen) put much emphasis on the achieve­
ments made by Germany in other areas, such as the educational sector, and in pioneering
the development of an R&D infrastructure.
7. See Gerschenkron (1962:16-20),
8. For instance, as a result of these efforts, illiteracy—which was relatively low by inter­
national standards even before the Meiji-restoration—was almost eliminated among
Japanese youth by the turn of the century, and by 1920, more than half of the children
graduating from elementary school proceeded to secondary schools (Odagiri and Goto
1996).
9. Productivity (and productivity growth) remained low, however, in sheltered industries
(agriculture and services), and this imposed growing burdens on the economy during
the 1990s and the early twenty-first century.
10. See Johnson (1982) on Japan, Amsden (1989) on Korea and Wade (1990) on Taiwan.
11. Most countries in our sample have between 80 per cent and 100 per cent of their youth
attending secondary schooling. However, some (but not all, Malaysia, for instance, has
93 per cent) of the poorer economies have less. For instance, Argentina and Chile were
reported to be in the 70-75 per cent range, and China and Mexico lower still, between
50 per cent and 60 per cent, Brazil is reportedly very low, only 15 per cent. Note,
however, that the data for Argentina and Brazil quoted here were from the 1985-7 period,
while for the other countries it was 1998 (source UNDP, Human Development Report
2002).
12. UNESCO defines the “gross enrolment ratio in tertiary education,” which is the indica­
tor used in Figure 19.2, as “total enrolment in tertiary education, regardless of age,
expressed as a percentage of the population of the five year age group following on from
the secondary-school leaving age.” If many tertiary students are older than this, as may be
the case in several advanced countries (with a high emphasis on “long” degree courses),
this indicator may give a too optimistic view on the actual share of an age group enrolled
in tertiary education. Comparisons with other available indicators of tertiary education,
based on educational attainment, indicates that such a bias may be present, and more so
for the USA than other countries. Hence, the figure probably exaggerates the difference
between the USA and other countries in the emphasis on tertiary education.
U. Note that the data reported here refer to total R&D, including the public part. If we had
focused on only the part undertaken (and/or financed) by the business sector, the
ranking of the countries would have been approximately the same, but the differences
between the top and the bottom would have increased (in general, the more a country
invests in R&D, the higher the share financed by business).
40 JAN F A G E R B E R G AND M A N U E L M . GODINHO

4, A recent example, mentioned by Chang, is the so-called TRIPS agreement, forcing

developing countries, which used to have a very lax attitude to protection of intellectual
property rights, to accept developed countries' standards and institutions in this area.
See Granstrand (Ch. 10 in this volume) for an extended discussion.
5. See Granstrand (1999: ch. 6), and references therein, for an attempt to link discussion of
catch-up and firm strategies.

Ie f e r e n c e s

A b r a m o v it z , M. (1986), “ Catching Up, Forging Ahead, and Falling Behind," Journal o f

Economic History 46: 386-406.
---- (1994), “ The Origins o f the Postwar Catch-Up and Convergence Boom ” in J. Fagerberg,
B. Verspagen, and N. von Tunzelmann (eds.), The Dynamics o f Technology, Trade and
G row th Aldershot: Edward Elgar, 21-52.
lOk i , M. (1988), Information , Incentives and Bargaining in the Japanese Economy , Cambridge:
Cambridge University Press.
lMSd en , A. H. (1989), Asia's Next Giant: South Korea and Late Industrialization , New York:
Oxford University Press.
►au m o l , W. J., B la c k m a n , S. A. B., and W o lff , E. N. (1989), Productivity and American
Leadership , Cambridge, Mass,: MIT Press.
►e a s l e y , W. G. (1990), The Rise o f Modern Japan, New York: St. Martins Press.
A p p e l e n , A., C a s t e l l a c c i , F., F a g e r b e r g , J., and V e r s p a g e n , B. (2003), “ The Impact of
EU Regional Support on Growth and Convergence in the European Union," Journal of
Common Market Studies 42: 621-44.
---- and F a g e r b e r g , J. (1995), “ East Asian Growth: A Critical Assessment," Forum for
Development Studies 2: 175-95 (repr. as ch. 3 in J. Fagerberg (2002), Technology, Growth
and Competitiveness: Selected Essays, Cheltenham: Edward Elgar)
C h a n g , H a -J oon (2002), Kicking Away the Ladder: Development Strategy in Historical
Perspective, London: Anthem Press.
Co r n w a ll , J. (1977), Modern Capitalism: its Growth and Transformation , London:
St. Martin’s Press.
)osi, G. (1988), “ Sources, Procedures and Microeconomic Effects of Innovation," Journal of
Economic Literature 26:1120-71,
F a g e r b e r g , J. (1987), “ATechnology Gap Approach to Why Growth Rates Differ," Research
Policy 16: 87-99 (repr. as ch. 1 in J. Fagerberg (2002), Technology Growth and Competitive­
ness: Selected Essays, Cheltenham: Edward Elgar).
---- (1988), “ Why Growth Rates Differ," in G. Dosi, et al, (eds,), Technical Change and
Economic Theory; London: Pinter, 432-57.
----- (i 9 9 4 )* “ Technology and Intern ation al D ifferences in G ro w th R ates," Journal o f
Economic Literature 32:1147-75.
— (2000), “Vision and Fact: A Critical Essay on the Growth Literature," in J. Madrick (ed.),
Uncovential Wisdom: Alternative Perspectives on the New Economy New York: The Century

* Asterisked Items are suggestions for further reading.

 IN NOVATIO N AND C A T C H IN G -U P 541

Foundation Press, 299-ЗЗО, 350-4 (repr. as ch. 6 in J. Fagerberg (2002), Technology, Growth
and Com petitiveness: Selected Essays, Cheltenham: Edward Elgar).
----- G uerrieri, P., and V erspagen , B. (eds.) (1999), The Econom ic C hallenge fo r Europe:
Adapting to In n ovation B ased G row th , Cheltenham: Edward Elgar.
and V erspagen , B. (2002), lechnology-Gaps, Innovation-Diffusion and Transform­
ation: An Evolutionary Interpretation” Research P o lic y 31:1291-304.
* F reem an , C. (1987)» Technology Policy a n d Econom ic Perform ance : Lessons fro m Ja p a n ,
London: Pinter.
------ and Lou<pA, F. (2001), As Tim es Goes By. From the Industrial Revolutions to the Infor­
mation Revolution, Oxford: Oxford University Press.
----- and Soete , L. (1997). The Econom ics o f In du strial Inn ovation , 3rd edn., London: Pinter.
*Gerschenkron , A. (1962), Econom ic Backw ardness in H istorical Perspective, Cambridge,
Mass.: Belknap Press.
Godin но, M. (2000), “ Desen.volvimento competitive do sector de cal^ado em Portugal:
Li^oes de um caso de clu sterin g? in I. Salavisa Lan^a (ed.), Trajectories com petitivas na
industria portuguesa, Oeiras: Celta.
Gomulka, $. (1971), “ Inventive Activity, Diffusion and the Stages of Economic Growth,”
Report No. 24, Institute of Economics, Aarhus University, Aarhus.
Granstrand, O. (1999), T he Econom ics o f M an agem en t a n d o f Intellectual Property, Chelten­
ham: Edward Elgar.
Hobday, M, (2000), “ East versus Southeast Asian Innovation Systems: Comparing OEM-
and TNC-led Growth in Electronics,” in L. Kim and R. Nelson (eds.), Technology Learning
& Innovation: Experiences o f N ew ly Industrializing Econom ies , Cambridge: Cambridge
University Press, 129-69.
J ohnson, C. A, (1982), MITI a n d the Japa nese M iracle: The Growth o f Industrial Policy ;
1925-1975, Stanford: Stanford University Press.
*Lall, S. (2000), “Technological Change and Industrialization in the Asian Newly Industri­
alizing Economies: Achievements and Challenges,” in L. Kim and R. Nelson (eds.),
Technology, L ea rn in g & In n ovation : Experiences o f N ew ly Industrializing Economies, Cam­
bridge: Cambridge University Press, 13-68.
Landes, D. (1998), T he W ealth a n d Poverty o f Nations, London: Abacus.
List , F. (1885), The N a tio n a l System o f Political Economy, trans. from the original German edn.
(1841), London: Longmans, Green and Company.
Nelson, R. R. (1968), “A ‘Diffusion Model of International Productivity Differences in the
Manufacturing Industry,” A m erican Econom ic R ev iew 58:1219-48.
-----and W inter , S. G. (1982), A n E vo lu tio n a ry Theory o f Econom ic Change, Cambridge,
Mass,: Harvard University Press,
* --- and W right , G. (1992), “The Rise and Fall of American Technological Leadership:
The Postwar Era in Historical Perspective,” Jo u rn a l o f Econom ic Literature 30:1931-64,
Odagiro, H„ and G oto, A. (1996), Technology a n d In du strial D evelopm ent in Ja p a n , Oxford:
Clarendon Press.
O'Sullivan , M. (2002), “Industrial Development: A New Beginning?” in John O'Hagan
(ed.), The E con om y o f Irelan d : Policy a n d Perform ance o f a European Region , Gill and
Macmillan, 260-85.
Rodrik, D, (1994), “ King Kong Meets Godzilla: The World Bank and the East Asian Miracle,”
in A. Fishlow et al. (ed.), Miracle or Design? Lessons from the East Asian Experience, Policy
Essay No. 11, Washington, DC: Overseas Development Council, 13- 53-
>4* JA N F A G E R B E R G A N D M A N U E L M. G O D I N H O

*Sh in , J ang -S up (1996)5 The Economics o f the Latecomers: Catching-up, Technology Transfer
and Institutions in Germany, Japan and South Korea, London: Routledge.
----- and Ha -J oon C hang (2003), Restructuring Korea In c , London: Routledge.
>o l o w >R. M. (1956), “A Contribution to the Theory ofEconom ic Growth,” Q uarterly Journal
o f Economics 70: 65-94.
Femple , J. (1999)5 “ The New Growth Evidence,” Journal o f Economic Literature 37:112-56 .
Feece , D. J. (2000), “ Firm Capabilities and Economic Development: Implications for the
Newly Industrializing Economies,” in L. Kim and R. Nelson (eds.), Technology, Learning &
Innovation: Experiences o f Newly Industrializing Economies, Cambridge: Cambridge Uni­
versity Press, 105-28.
^eblen , T. (1915), Im perial Germ any and the Industrial Revolution, New York: Macmillan.
/erspagen , B. (1991), “A New Empirical Approach to Catching Up or Falling Behind,”
Structural Change and Economic Dynamics 2:359-80.
W ade , R. (1990), Governing the M arket: Economic Theory and the Role o f Governm ent in East
Asian Industrialization, Princeton: Princeton University Press.
Vomack , J., J ones , D. T., and R oos, D. (1991), The M achine that Changed the World,
Cambridge, Mass.: M IT Press.
Vorld B ank (1993), The East Asian M iracle: Economic Growth and Public Policy, New York:
Oxford University Press.
 C H A P T E R 20

INNOVATION AND
COMPETITIVENESS

JOHN CAN TW ELL

20.1 In tr o d u c t io n

Traditionally , econom ists and econom ic historians since Adam Smith have
discussed econom ic growth principally in the context o f the national level— why
some countries grow faster (in modern terms, acquire the capabilities for sustained
growth that make them more competitive) and so become wealthier than others.
While in neoclassical economics questions o f national competitiveness came to
assume a lesser degree o f importance, as attention was shifted away from issues of
growth towards those o f static resource allocation and efficiency, there was even less
concern with the notion o f competitiveness at the firm level. The theory of the
(comparative) growth o f the firm was a minority interest o f those such as Downie
(1958), Penrose (1959), and Marris (1964), typically treated as a rather esoteric sub­
branch of industrial economics, that was to be accorded a lesser status in the
discipline than the conventional theory of the firm (which was really a theory of
the relationship between the firm and markets). In recent years two related changes
in economics and allied areas o f research have been under way. One is a revival of a
more widespread interest in the classical issues of competitiveness at a national level,
and the other is the growing attention now paid to competitiveness at the level o f
industries, regions, and firms, in which fields o f research a substantial new literature
544 JOHN C A N T W E L L

has emerged. In Section 20.2 the contribution is assessed o f the new literature on
com petitiveness across countries. Section 20.3 exam ines innovation and com peti­
tiveness at the industry level that connects together firm s and their environm ent,
and Section 20.4 looks at the regional and firm level. Section 20.5 draws some
conclusions with respect to the interaction between innovative actors, between the
different levels o f analysis o f competitiveness, and opportunities for future research.
C o m p etitiven ess is here taken to mean the possession o f the capabilities needed
for sustained econom ic growth in an internationally com petitive selection environ­
ment, in w h ich e n viro n m en t there are oth ers (cou n tries, clu sters, o r in d iv id u a l
firms, depending upon the level o f analysis) that have an equivalent but differen­
tiated set o f capabilities o f their own. The term competitiveness is also sometimes
taken to necessarily im ply as a result a continuing rise in the living standards o f the
individuals that are members o f a social group with the required capabilities
(notably in this context, to im ply a sustained increase in the living standards o f
the citizens o f the country that is suitably competitive in world markets— see Tyson
1992). W hile it is indeed necessarily true that productivity growth increases incomes
on average (i.e. per capita income), it m ay well be that the process o f capability
generation and growth also has a disruptive effect on the distribution o f incomes.
This issue is not addressed directly here, since the way in which innovation affects the
em ploym ent opportunities o f individuals, which is a m ajor influence upon their
respective earning capacities, is the subject o f Pianta (Ch. 21 in this volum e).
The winners from innovation are those that construct appropriate capabilities,
but capabilities are localized and nationally differentiated (as explained by Edquist,
Ch. 7 this volum e), and so there can be m any successful players in the competitive
game, each to some extent learning from and interacting with the somewhat
alternative paths to capability creation being taken by others. Put in these terms
few could object that the pursuit o f competitiveness through innovation is a
laudable objective o f national policy, and indeed an increasingly im portant objective
as the role o f innovation has risen in the modern knowledge-driven economy, even
for (actually especially for) countries that start behind and wish to catch up (Fager-
berg and Godinho, Ch. 19 in this volume).
To be meaningful, competitiveness must be thought o f as entailing a relative
com parison o f growth rates or benchm arking o f perform ance to assess how well
each participant has done in developing the capabilities for innovation and growth,
and not be about the mutual potential for dam aging one another (a misleading
interpretation o f competitiveness criticized by Krugm an 1994a, 1996). It is reason­
able to expect that, at least on average, the spillover benefits for others o f a good
perform ance in one location or by one agent tend to outweigh the costs o f that good
performance for others. This argument is largely applicable whether the unit o f
analysis is countries in the world econom y or firms in an industry. At a country level
the efforts o f each national system o f innovation to prom ote the competitiveness o f
businesses sited locally are increasingly com plem entary as scientific and engineering
 __ INNOVATION a n d c o m p e t i t i v e n e s s 545

communities become more international, and cross-border knowledge flows are

more common (as discussed by Narula and Zanfei, Ch. 12 in this volume). Likewise,
much of the growth achieved by the leading corporations in an industry reflects the
wider growth o f that industry. The competitive race between firms stimulates
innovation, and this innovation lowers costs and improves product quality in the
industry, and thereby increases industry demand. All firms benefit that contribute
successfully to what is often a com bined and interactive process o f innovation.

20.2 C o m p e t it iv e n e ss at the
N a tio n a l L evel

When looking at the country level, competitiveness is about the way in which the
pattern of international trade evolves over time to reflect changing patterns of
capabilities and hence competitive advantage (what might be thought of as the
evolution in the comparative advantage of countries), rather than about the estab­
lished pattern o f comparative advantage which is the usual focus of trade theory.
While the earliest theories o f trade and growth can be traced back to the classical
economists, such dynamic accounts of the paths of international trade and invest­
ment were revived in recent times by the technology gap approach (Posner 1961) and
the product cycle model (Vernon 1966). However, a major shortcoming of the
product cycle model was its reliance upon an overly simplistic demand-driven
theory o f innovation (which reflected the spirit o f the 1960s, when it was devised),
through which the firm was assimilated to the product, and innovation was sup­
posed to be concentrated in just one leading country— the US (see Cantwell 1989, for
a further discussion o f the m odel). Sadly, when the product cycle model broke down
in the 1970s, in large part ow ing to the reemergence o f multiple centers for innov­
ation in a num ber o f international industries, the amended versions o f the model
(Vernon 1974, 1979) focused upon considerations o f oligopolistic strategy rather
than revisiting the underlying theory o f innovation and competitiveness. It was only
in the 1980s that scholars based at Sussex once again wedded an analysis o f structural
shifts over time in the pattern o f international trade to a more realistic approach to
innovation: see Soete 1981; Dosi and Soete 1988; Dosi, Pavitt, and Soete 1990; and
Fagerbergs 1987 paper on structural changes in international trade (repr. as ch. 7 in
Fagerberg 2002).
Part o f the inspiration for Fagerberg’s research had been that economists some­
times use the term “ com petitiveness” in various different ways, and especially in
macroeconomic policy discussions not always in the way that has been defined here.
T h is ch ap ter is co n cern ed w ith in n o va tio n an d co m p etitiven ess, an d th is is so m e ­
tim es d istin gu ish ed as b ein g ab o u t lo n g er-term tech n o lo g ical co m p etitiven ess, as
o p p o sed to sh o rter-term p rice com petitiven ess. T h ere are tw o d ifferen t w ays o f
d iscu ssin g co m p etitiven ess in the latter sense o f sh o rter-te rm p rice co m p etitiven ess.
In the context o f co n ven tio n al d em an d m an agem en t p o lic y d iscu ssio n s, i f (say)
lo w er g o vern m en t b o rro w in g m eans a fall in interest rates an d so a rise in net
o u tw ard investm ent, and i f this leads to a decline in the v alu e o f the d o m estic
currency, then the price "'competitiveness” o f dom estically produced goods and
services can be said to have increased, as export prices fall in foreign currency terms
while im port prices rise in domestic currency terms. However, this type o f com peti­
tiveness is unlikely to be sustainable, especially if (for example) the rise in im port
prices sparks o ff domestic inflation, or if lower net inward investment has adverse
consequences for domestic productivity growth. The second and for our purposes
m ore substantive context is the conventional cost-based account o f competitiveness,
in which a fall in relative unit labor costs means lower prices (or a lower rate of
inflation), which in turn leads to a rise in exports and fall in im ports, and so an
increase in the value o f the domestic currency.
Longer-term technological competitiveness is more akin to the second o f these two
versions o f price competitiveness, in supposing that a faster growth o f (output and)
exports drives up the domestic currency, rather than it being a falling currency that
promotes net exports. In the context o f what is sometimes termed “ non-price”
competitiveness to distinguish it more clearly from the kind o f cost-based competi­
tiveness just referred to, innovation and new lines o f value creation m ay mean higher
average prices as an indicator o f higher quality, but in any event they lead to a faster
growth o f productivity and trade, and thus an upward trend in the value o f the
domestic currency. Given what has been said already, it is worth stressing here that
in this perspective the rise in the value o f the currency is simply the reflection o f
competitiveness, defined as a relatively rapid growth in productivity and the value o f
(output and) exports. The rise in the value o f the domestic currency is not itself the
achievement o f competitiveness (an improvement in the terms o f trade that is
essentially a potential side effect resulting from competitiveness). It is also worth
making explicit that the departures from comparative advantage associated with trade
imbalances are merely a temporary result o f competitiveness in this framework, and
again not in themselves the objective o f competitiveness. What is implicitly supposed
here is that faster productivity growth is associated with a rising share o f w o rld trade,
and that in this process the growth o f exports leads the growth o f imports. So net
exports rise until imports catch up, and this catching up is facilitated by the conse­
quent rise in the value o f the domestic currency and in domestic wage rates.
Now neo-Schumpeterian approaches to international competitiveness focus on
this kind o f process o f forging technological competitiveness, which for those whose
innovative efforts are most successful implies a sustainable increase in the share o f
world trade (or at the firm level, a sustainable increase in the share o f the relevant
 ______ ______ in n o v a t io n and c o m pe t it iv en e ss 547

world market). However, as has been discussed at length already, in the Schumpe­
terian perspective competition entails the positive sum game of establishing new
spheres of value creation, so innovations expand the overall magnitude of world
trade and the world market. Those that contribute most to this process of expansion
see their shares rise as they are responsible for more of the new element of value
creation, and not because o f a substitution effect within some fixed total level of
world trade or some fixed and given world market (or even within some steadily
exogenously growing world market). In this context, the neo-Schumpeterian analy­
sis of innovation and competitiveness is unlike equilibrium growth accounts, even
when those accounts incorporate an acknowledgement of research activity, if invest­
ment in innovation is treated as being inherently like investment in any other
economic activity, and if the only difference between activities is treated as lying in
their wider impact on other activities through externalities. Instead, in the neo-
Schumpeterian story the very nature and purpose of innovative activity is to disturb
and add to the existing circular flow of income generation, in an experimental and
non-equilibrium fashion.
Such neo-Schumpeterian models of innovation and growth might be specified in
at least two alternative ways. The first of these leans heavily on the distinction just
drawn between shorter-term price competitiveness and longer-term non-price
technological competitiveness. In Fagerbergs (1987,1988) technology gap formula­
tion of international competitiveness across countries, the impact on growth of
national rates of innovation and distance behind the technology leader are treated
primarily as additive elements, to be added on to the more traditional determinants
of economic growth in the form o f capital accumulation (the share of investment in
national output) and relative unit labor costs. The origins of viewing cross-country
growth in this kind o f additive framework can be traced to Abramowitz (1956),
Solow (1957), and Denison (1967), for whom technological improvements (and the
productivity growth to which they led) were an obvious means of accounting for the
substantial “residual” in variations in growth that remained after allowing for the
effect of the increase in factor inputs. So in this context capital accumulation proxies
for the extension o f the scale o f established activities, relative unit labor costs capture
cost-based “price” competitiveness, while the contribution of corporate research
and the capacity to catch up through imitating the achievements of a leader
represent “non-price” technological competitiveness.
Setting the p ro b le m u p in th is w a y is con ven ien t, as the em p irical evidence then
generally suggests th at te ch n o lo g ic a l com p etitiven ess is m ore im p o rtan t than the
more co m m o n ly c o n sid ered tra d itio n a l in flu en ces u p o n com petitiveness. Techno­
logical com p etitiven ess is ju d g e d to b e m o re sign ifican t than relative u n it lab o r costs,
although Krugm an (1994b) an d Y o u n g (1995) p o in t to the co n tin u in g im p o rtan ce o f
capital accu m u latio n w ith in th is k in d o f fram ew o rk . T h e evidence for three conn
tries— Japan , the U K , an d th e U S — o ver the p erio d 19 6 0 -7 3 illustrated in
Table 20.1. B ased o n th e e stim atio n o f his em p irical m o d el o f intern ational
 ТШ еЛІ The decomposition of the predicted growth in national market shares
from an estimated empirical model of cross-country competitiveness,
for 1 9 6 1 - 7 3 ^

Japan UK USA

G ro w th In te c h n o lo g ic a l c a p a b ilitie s 66.9 6.9 -0 J

Rise in re la tive u n it la b o u r costs “ 0.9 0,8 1.6
In itia l te c h n o lo g ic a l c a p a b ilitie s (c a tc h -u p ) 20.9 1Б.9 7.3
In v e s tm e n t as share o f GDP, and g ro w th o f w o rld 16.5 -3 9 .8 - 3 8 .2
dem and
T otal g ro w th in m a rke t share (predicted by 103.3 “ 16.2 -2 9 .8
m odel)

Source: Fagerberg (1988).

competitiveness, Fagerberg (1988) was able to decompose the m odel's predicted

change in each country's share o f world trade (which were reasonably good predic­
tions o f the actual changes in market shares) into four elements, as shown. What
emerges is that the traditional consideration o f relative wage costs contributed
rather little to overall competitiveness in any o f these countries (although it was
statistically significant in all the equations o f the model in which it appeared). In
contrast, the growth in indigenous technological capabilities in Japan, and the
diffusion o f foreign frontier technologies, account for a good deal o f the Japanese
competitive success o f that period. The loss o f world trade shares by the U K and the
US over the same period can be attributed mainly to weak capital accum ulation, and
Fagerberg explained this m ainly by the drain placed on national resources by the
high shares o f m ilitary spending in these two countries.
However, when capital accumulation contributes positively to a favorable growth
rate, at least some element o f it reflects the establishment o f new fields o f activity, and
is a response to the creation o f new innovative opportunities. Therefore, it is not
clear that the contribution o f the growth o f traditional factor inputs can really be
cleanly distinguished from that o f innovation, unlike in the logic o f a standard
production function approach. So to exclude capital accum ulation from the contri­
bution o f technological competitiveness provides only a conservative lower bound
estimate o f the significance o f the latter, and perhaps concedes too m uch to orthodox
skeptics o f the role o f innovation in growth and competitiveness. Fagerberg (1988)
was aware o f this issue, and so he included a separate equation in his simultaneous
system for capital accumulation as a function o f the growth o f output, which in turn
depended as we have seen upon the increase in technological capacity, so that he
 I N N O V A T IO N AND C O M P E T IT IV E N E S S 549

acknowledged indirectly the influence of technological competitiveness upon cap­

ital investment. This need to revise the traditional production function logic be­
comes especially relevant when trying to compare innovative or technological
“assimilationist” explanations of (East Asian) competitive success with those of
“accumulationists,” if using aggregate measurements in the context of substantial
structural change (Nelson and Pack 1999). A s we have already noted, neo-
Schumpeterian economists have particularly emphasized the connection between
structural change and growth through innovation.
Some evidence on what distinguished the East Asian growth experience from that
of other countries that sustained similarly high rates of capital accumulation over
the 1960-89 period is set out in Table 20.2. The table shows eleven countries that
enjoyed very high shares of investment in GDP, of over 20 per cent, as indicated in the
first column. The right-hand column shows the residuals of a regression across 101
countries of GDP per capita on the investment share as a proxy measure of the rate of
capital accumulation, and on three other control variables (a catching~up effect
proxied by the intial level o f GDP per capita in 1960, the growth of population to
capture the availability of labor supplies, and the proportion of the relevant cohort
of the population educated to at least secondary school standards). This was part of
the study of Nelson and Pack (1999). What emerges is that among high investment
countries, the East Asian tigers— Hong Kong, Korea, Singapore, and Taiwan— stand

Table 20.2 Actual growth rates achieved by countries, 1960-89, over and above
that predicted by ( in te r a lio ) their rates of capital accumulation

Investment/GDP (°/o) Actual minus predicted growth

rate o f GDP per capita

Hong Kong 27.3 0.031

Korea 24.9 0,032
Singapore 34.3 0.017
Taiwan ■25.0 0.047
Gabon 40.0 -0 .0 3 0
Algeria 35.0 -0 .0 2 6
Greece 24.2 0.008
Panama 24.0 0.002

Portugal 23.7 -0 .0 0 2

Jamaica 25,0 -0 .0 3 7

Ireland 22.2 0.011

550 JOHN C A N T W E L L

out as m anaging to achieve growth rates well in excess o f what m ight have been
predicted from their favorable rates o f capital accum ulation alone. What was
different in these economies was their greater ability to innovate» to upgrade and
restructure their indigenous industries» and to learn and absorb m ore effectively
from foreign technologies. Capital accum ulation can em body innovation to the
extent that it is linked to the transform ation o f the productive activities being
conducted.
So an alternative approach also in the Schumpeterian tradition is to treat techno­
logical accum ulation and capital accum ulation as sim ply aspects o f a common
process, rather than as independent (even if com plem entary) contributions to
growth. In this case innovation can be seen as driving up profitability and hence
lowering the share o f wages in output (even though wages are rising faster, and so
m ay be unit labor costs), which leads to a higher share o f investment in output, and
so higher capital accum ulation and growth as a result o f higher technological
accum ulation (Cantwell 1989, 1992). The basic idea here is that in fast-growing
countries just as an increase in imports tends to follow an increase in exports with
a lag, so wages tend to follow productivity increases with a lag, enabling innovation
to create a fresh source o f profitability and growth. Yet this also suggests that
technological competitiveness is in part cost-based. It should be noted, though,
that labor productivity is defined here sim ply as the value o f output per worker
employed, which implies that productivity growth is as m uch attributable to
product quality improvements (that raise the value or unit price o f output, as
stressed in Fagerbergs approach), as it is to the cost reductions associated with
process improvements. In this alternative neo-Schum peterian form ulation we need
w orry less about the distinction between embodied and disem bodied technological
change, or the distinction between improvements in product quality and delivery
times as opposed to improvements in processes that are reflected in costs and prices.
The renewed interest in international competitiveness and variations in growth
rates has spawned a substantial new literature on cross-country convergence or
catching-up versus divergence or falling-behind (see e.g. Baum ol, Nelson, and Wolff
1994). The evidence suggests that whether one observes convergence or divergence
depends upon the period studied and the countries selected. In any case, the overall
trend in cross-country variance at a world level m ay not be the m ost im portant issue.
Rather than trying to work out whether East Asian convergence statistically out­
weighed the effect o f African divergence in aggregate, the issue is m ore why and how
firms in East Asia had the capabilities to catch up in the period since i960» while
those in Africa did not. The concepts o f a techno-socio-econom ic paradigm (Free­
man and Perez 1988), or o f an evolution in the institutional characteristics of
capitalism (Lazonick 1991» 1992), can be useful in this respect as a means o f explain­
ing occasional shifts in technological leadership or longer term competitiveness, and
in the direction o f those shifts. Emphasizing again the role o f structural change in
growth, and in particular during periods o f paradigm change, when the prevailing
 I N N O V A T I О N A N D C OMP E T I T I V E N E S S 551

characteristics of innovation undergo transformation, helps to explain the exist ence

of windows of opportunity in which the catching-up of selected countries may be
especially dramatic. At these times leaders can have special difficulty in adjusting to
the new conditions since they have become most locked in to the types of innovation
favored under the earlier paradigm, while others that lie behind initially may find
that their rather different institutions and methods of social organization are in fact
quite well suited to adapting so as to promote just the kinds of structural change in
which lie now the greatest opportunities for fresh innovation.

20.3 C o m p e t it iv e n e s s at th e In d u st r y
Lev el: th e N exu s of R ela tio n sh ips
b e t w e en F ir m s and t h e i r E n v iro n m en t

When speaking of shifts in competitiveness between firms or between different

national groups of firms that constitute the major players in an international indus­
try, Mowery and Nelson (1999) prefer the term industrial leadership, so as to
emphasize that such leadership may be due as much to the national or regional
environment in which firms operate, or to institutions that are specific to an indus­
try, as to factors that are purely internal to the firms in question. From detailed
historical case studies of the evolution of national industries, they conclude that
competitiveness derives from the contributions of each of, and the interactions
between, firms, regions, and countries, and the sectoral support systems that
connect these different levels of analysis. Their account provides a clear justification
for a section that covers competitiveness at the international industry level, rather
than attempting to move directly from the country level to the firm level
In this framework the factors that are thought to influence competitiveness may be
grouped under the headings of resources or capabilities, institutions (notably for
higher education and science, and in financial systems), markets or demand condi­
tions, and inter-company networks. Models that are based at the level of the evolution
of particular products or technologies, or which are predicated on a notion of
competence-destroying innovation (when moving from one type of product or
technology to another that represents a radical departure from the past), may each
have some relevance. Yet Mowery and Nelson argue that this applicability is limited
when dealing with the performance of large multiproduct firms over very long periods
of time. Likewise, the policy debate surrounding the relationship between government
policy interventions and trends in competitiveness is overly polarized, when each side
552 JOHN C A N T WE L L

o f the debate has in mind a rather simple model o f that relationship (or lack o f a
relationship) which may apply at some places and at certain times, but cannot be
universally applied in the way that some advocates seem to imagine.
The relationships that exist between the development o f the technological cap-
abilities in firms that are responsible for competitiveness and the institutions o f the
wider society vary from one country to another, but in particular they tend to be
different in countries that belong to an already leading industrialized group and
those that are attempting to catch up with them (see also Fagerberg and Godinho,
Ch. 19 in this volum e). It is noticeable that there have been a greater num ber o f cases
in which governments in catching-up economies, partly through measures o f
domestic protection, have contributed m ore actively to the fostering o f capabilities
in local infant industries and in indigenous companies. This was true o f the US and
Germ any when they were catching up with Britain in the nineteenth century (Landes
1969), it was true o f Japan when it was catching up with the West during the
twentieth century (Ozawa 1974), and it was true o f Korea when it was catching
up after i960 (Enos and Park 1988). It is true that there are occasionally other cases
o f catching-up economies, like those o f Singapore or M exico in recent years, that
have taken advantage instead o f various aspects o f trade liberalization. However,
what is m ost noticeable in all these instances o f successful catching-up is that the
trade policies o f governments were merely part o f a much wider package o f support
for the longer term nourishment o f capabilities in indigenous firms. Since the
emergence o f science-based industries towards the end o f the nineteenth century
this has meant especially investing in science and higher education, in the training o f
engineers, and in the learning o f skills more widely (Freeman and Lou^a 2001).
Equally important, where there were measures o f trade protection local firms
accepted their part o f the bargain to invest very substantially in capability creation
in an outward-looking and export-oriented fashion, rather than sim ply remaining
an inefficient enclave as in so many other cases o f protectionism or so-called import-
substituting industrialization.
O f course, the institutional structures o f catching-up economies changed mark­
edly (and any protectionist measures were largely reversed) as their firm s caught up
and themselves sometimes forged ahead and became innovative leaders in their own
right. This is perhaps the most vivid illustration o f the general observation that the
development o f technological capabilities in firms and the character o f the insti­
tutions that support these competitive efforts in the wider society tend to coevolve
with one another (Nelson 1995), through a process of continual interaction. Another
perspective on these interrelated national systems for the construction o f competi­
tiveness is offered by Porter (1990), as represented through the four com ers o f a
diam ond o f factor conditions; demand conditions; related and supporting indus­
tries; and firm strategy, structure, and rivalry, In Porter s view the capacity o f firms to
innovate depends ciitically on having sufficient domestic rivalry in their own home
country o f origin, but also on the presence o f spillovers between firms associated
 I N N O V ATION AND C O M P E T I T I V E N E S S 55^

with localized clusters (to which issue we return below). In other words, innovation
requires an appropriate mix o f interfirm rivalry and cooperation or exchange
(Richardson 1972)- Lazonick (1993) has argued that when confronted with a major
new competitive challenge from some new source of innovation from outside,
domestic industries may need to shift this balance away from rivalry and towards
cooperation in order to respond effectively. To express this another way in the light
of the trend towards globalization mentioned earlier, it may be that firms in some
national industry may need to increasingly collectively focus their efforts in what
they do locally (as opposed to activities they may locate abroad) to be better
mutually aligned with whatever may be the fields of particular local excellence or
of specialization in innovation. This would have the effect of tending to reinforce
national patterns of comparative advantage in innovation.
As has been mentioned already, with the emergence of science-based industries
over a century ago, the need for an infrastructure that suitably supports relevant
education, skill formation, and training became critical to the competitiveness of
industries, and is widely believed to have become more important still in the modem
techno-socio-economic paradigm associated with computerization and informa­
tion processing. For firms to be able to create capabilities requires costly and difficult
internal learning processes, but these in their turn depend upon having suitable
organizational and technical skills in the management and workforce on which they
rely. The composition o f skills in the workforce of the home base of firms is therefore
critical to the success or failure of countries that are trying to catch up, but it also
becomes a central influence upon the fields in which any national group of firms has
its specific pattern of comparative advantage in innovation and capability creation.
Of course, this is not just a one-way street, since the types of investments and
commitments to training that are made by firms themselves in the course of
learning, the professional associations they help to form, and the pressures they
place upon governments and others imply again a process of coevolution between
firms and their environment in this respect too.
Table 20.3 helps to illustrate the significance o f education and skills in the
catching-up of the four East Asian tiger economies (see also Fagerberg and Godinho,
Ch. 19 in this volume). Korea stands out as having surpassed even the commitment
of the traditional industrialized countries to higher education in the natural sciences
and mathematics. Yet a key to the success of these countries as a group lies more in
the investments they have undertaken in support of engineering graduates while
Hong Kong lies behind the industrialized group (and this may help to account for
wby iis local learning and upgrading has been more limited than in the other three,
as discussed by Tail 2001), Singapore is above the industrialized country average, and
Korea and Taiwan are way ahead o f that average for tertiary level engineering
enrolments as a proportion of the population. Other developing countries are
generally well behind the achievements in engineering education of the tiger econ-
omies, although the Philippines, Argentina, and Mexico have at least matched the
554 J O H N C A NT WE L L

Table 20 3 Educational enrollments in technical subjects at tertiary level a s a % of

the total population in selected countries, in 1 9 9 5 or closest year
available

Natural science, maths E ngineering

andcomputing
Japan 0.07 0.39
France 0.53 0 .0 9
Germany 0.39 0 .4 9
UK 0.31 0.38
USA 0.39 0.31

H ong Kang 0.20 0.2S

Singapore 0.10 0.47
Korea 0 .5 6 0 .9 8
Taiw an 0 .2 4 0 .8 6

Indonesia 0.02 0.11

Malaysia 0.07 0.07
P h ilipp in es 0.22 0.33
Thailand 0.14 0.19
C hina 0.03 0 .1 0
India 0.10 0.02
Argentina 0.21 0.29
Brazil 0.09 0 .1 0
Mexico 0.06 0.27

Source: tail (2001).

position o f Hong Kong. Considering the enorm ous size o f its population, it is also
clear that China has been catching up fast in this area.
T h e Japan ese and G e rm an system s are k n o w n to fo cu s o n a b ro a d an d deep skill
base by em p h asizin g the acq u isitio n o f general en gin eerin g skills a n d the good
stand ard o f basic edu cation o f the p o p u latio n as a w h ole, w h ile the U S and U K
system s tend to be m o re elitist an d fo cu s o n the d evelo p m en t o f h igh er-en d skills
o ver a n arro w er range o f p eople (Prais 1995; L azo n ick and O ’S u llivan 1997; Lazonick
1998). I his helps to explain w hy the Jap an ese an d G e rm a n fields o f co m parative
advantage in in n o vatio n (and th u s, o f co m p etitive ad van tage) in clu d e m o to r
vehicles and engineerin g that rely (and have in creasin g ly co m e to rely) o n a broad
 innovation and competitiveness 555

skill base, while the US and UK advantages include aerospace, software, pharma­
ceuticals, biotechnology, and medical equipment that rely on very highly skilled
individuals and an intensive R&D effort.
When examining the extent of path-dependency in the specific technological
traditions of national groups of large firms, and in their patterns of technologi­
cal specialization as a measure of their relative contributions to each of the major
international industries (the cross-sectoral pattern of their technological competi­
tiveness), some cross-border interdependencies appear to emerge. This raises again
the issue of whether or not there have been any elements of convergence across
countries, but in this context in the specific mix of strengths and weaknesses in
international industries, rather than in aggregate productivity or performance.
Examining patterns of technological specialization among national groups of the
largest firms from six countries (the US, Germany, the UK, France, Switzerland, and
Sweden) based on their patterns of corporate patenting, it has been observed that
these profiles are path-dependent and tend to persist to some extent even over
periods of sixty years, from the interwar period to the present day (Cantwell
2000). This may be taken to imply that the positive effect on the continuity of
collective technological trajectories of intercompany technological cooperation and
spillovers within national groups has tended to outweigh the negative effect of
mobility in cross-company distributions of activity. There is some evidence that
through the evolution in these patterns of technological competence that has
occurred, certain national groups have come somewhat closer to one another than
they were in the past, or they have changed in similar ways. Indeed, it might be
argued that the six national groups examined can now be divided into three clusters
of two countries each.
The first cluster comprises the largest US and UK firms, in which the profile of
technological competence can be characterized as being resource-based, oil-related,
and defence-related. It is increasingly also health-related. As shown in Table 20.4, in
the US case since the interwar period a continuing comparative advantage in
innovative activity in the largest industrial firms has been sustained in the oil,
food products, rubber products, aerospace (defence and larger-scale transport
systems) and building materials industries. The greatest continuing strengths for
the largest British companies over the same historical period has been in textiles,
other transport (defense) and oil since the 1930s. Thus, it can be argued that there has
been some convergence in the profiles of the US and UK innovation systems
(Vertova 1998). UK firms have also seen a post-war shift into technological compe­
tence in the pharmaceutical industry, although it can be claimed that this too
represents the revival of a much earlier nineteenth-century tradition in biological
&wdmedical technologies. In any event, consistent with the overall UK or US pattern
of technological development, the British pharmaceutical industry had links with
the food industry, unlike in Germany where it derived purely from the chemicals
industry (Cantwell and Bachmann 1998). In the US there has been a related post-war
556 JOHN CANTWELL

continuation and strengthening o f the medical instrum ent industry (grouped under
professional and scientific instruments in Table 20.4), and a m ore recent m ove into
biotechnology although this has not yet been reflected in a com parative advantage
in the pharmaceutical industry as a whole.
The second cluster is that o f the Germ an-and Swiss-owned corporate grou p s, in
which technological development since the end o f the nineteenth century has been
largely science-based, and revolved around the dom inance o f the chemicals indus­
try In the post-war period this has been increasingly com plem ented b y engineering
excellence, although some recent comm entators have seen this direction o f change
(as opposed to a move Into the other science-based area o f electronics) as a weakness
o f the m odern Germ an innovation system (Albach 1996; Audretsch 1996). The
leading German firms have held a consistent focus on developm ent in the chemicals
and metal product industries, with some recent shift towards industries m ore reliant
on engineering-based technologies, linked in part to the emergence o f a wider range
o f innovative smaller specialist supplier companies. The Swiss concentration histor-

Table 20.4 The industries in which the largest nationally owned firms have
persistently held comparative advantage in innovation, 1 9 2 0 *3 9 and
1 9 7 8 *9 5

US-owned UK-owned
Food a n d d rin k T e x tile s
O ffic e e q u ip m e n t a n d com puting O th e r tra n s p o rt e q u ip m e n t
O th e r tra n s p o rt e q u ip m e n t C oal a n d p e tro le u m p ro d u cts
(o th e r th a n m o to r ve h ic le s )

R u b b e r a n d p la s tic pro d u cts

N on-m etal lie m in e ra l p ro d u cts

Coal a n d p e tro le u m p ro d u cts (o il)
P ro fessio n al a n d s c ie n tific in s tru m e n ts

German-owned Sw iss-ow ned

C h e m ic a ls C h e m ic a ls
P h a rm a c e u tic a ls P h a rm a c e u tic a ls
M e ta l p ro d u cts M e c h a n ic a l e n g in e e rin g
M o to r vehicles

French-owned Swedish-owned
M e ta l p ro d u cts
M e c h a n ic a l e n g in e e rin g
R u b b er an d p la stic pro d u cts

N o n -m e ta llic m in e ra l pro d u cts

Source: Cantwell (2000).

 INNOVATION AND COMPETITIVENESS 557

ically on chemicals and pharmaceuticals makes it a microcosm of (part of) the

German innovation system, which has also been shifting in the direction o f engin­
eering excellence.
The third cluster may be more a matter of coi ncidence than due to any historical,
geographical, or cultural ties, involving as it does the French and Swedish national
groups of companies. This grouping has emphasized infrastructural types of tech­
nology, spanning engineering, construction, transport and communications
systems, and some recent moves into health care. In the French case comparative
advantage in large firm innovation has been sustained since the interwar years in
metal products, rubber products, and building materials, while some earlier
strengths in electrical communications technologies have been subsequently con­
solidated. This infrastructural orientation is less reliant upon large-scale private
corporate R8cD than the German system has been, but is not as resource-oriented as
the US or UK company systems of technological development. Swedish techno­
logical excellence has also been engineering-based (and has become increasingly so)
around the metals and vehicles industries, but it has been shifting more closely
towards the French pattern with the recent rise of development in the areas of
telecommunications and pharmaceuticals.
The apparent convergence o f certain national systems of large-firm innovation
with continuing differentiation between these clusters of groupings may be an aspect
of the rise in technological interrelatedness and interlinked systems of technologies,
which have eroded the more highly specialized national systems of the past. Hence,
the significance o f technological lock-in and path-dependency in each respective
system has still been accompanied by some selected convergences between particular
national groups.

20.4 C o m p e t i t i v e n e s s a t t h e R e g i o n a l
a n d F irm L ev els

The significance o f the “ regional dimension’*of an innovation system has emerged as

another aspect of an interactive model (Kline and Rosenberg 1986) that emphasizes
the relationships o f local companies with knowledge sources external to the firm
see also Asheim and Gertler (Ch. 11 in this volum e). Such relationships between
firms and the science infrastructure, between producers and users of innovations at
an interfirm level, between firm s and the wider institutional environment are
strongly influenced by spatial proxim ity mechanisms that favor processes o f polar­
ization and cumulativeness (Lundvail 1988; von Hippel 1989). Furthermore, the
55B JOHN C A N T WE L L

employment of informal channels for knowledge diffusion (of so-called tacit or

uncodified knowledge) provides another argument for the tendency of innovation
to be geographically confined (Hagerstrand 1967; Lundvall 1992) *The lack of existent
capabilities in weaker regions hampers the potential for inter-regional technology
diffusion (Fagerberg, Verspagen, and Сапіёк 1997)-
Som e evidence on the extent o f the locational concentration o f the corporate
capabilities for innovation that underpin competitiveness in Europe is set out in
Table 20.5. It shows that except in the case o f Germ any in which there are four major
regions that are each responsible for 13 per cent or m ore o f the innovative capacity o f
the largest companies, o f which the leading region accounts for 27 per cent, in all the
other European economies the equivalent share o f the biggest region is around 50 per
cent or higher. This represents a quite remarkable geographical concentration o f
innovative capacity, far m ore than the extent o f concentration o f population or the
total value o f output. Given therefore that the interactions between the development
o f technological capabilities by firm s and the supporting institutions found in their
environment takes place m ainly in such regionally bounded areas, this section
begins by discussing the relationship between regional concentrations o f activity
or “ industrial districts" and the competitiveness o f individual firm s (Malmberg,
Solvell, and Zander 1996; Porter and Solvell 1998; Enright 1998; Scott 1998).
One particular aspect o f regional systems that is underlined here is their interplay
with the international dispersion o f the creation o f new technology and the new
innovatory strategies o f multinational corporations (M N C s), which have been
associated with a restructuring o f M N C technological operations at a subnational

Table 20.5 The shares of patenting of the largest industrial firms attributable to
research facilities located in the biggest single region of selected
European countries, in 1 9 6 9 -9 5

Country/Region Percentage
Belgium(Flanders-Brussels) 78.6
France(liede France) 57.9
G erm any (NordrheinW estfalen) 27.0
Italy(Lombardia) 52.3
Netherlands (South Netherlands) 62.7
Sweden(S to ckh olm -O stra Mellansverige) 49.5
Switzerland (Basel) 57.5
UK(SouthEast England) 46.9

Source. Cantwell and lammarino (2001), and (for Germany) Cantwell, lam m arino, and Noonan (2001).
 ___________ i n n o v a t i o n a n d competitiveness 559

level On the one hand, as seen above, there are general external economies and
spillover effects which attract all kinds of economic activities in certain regions and
determine, in the case of corporate integration, the localization of new research units.
These centripetal forces strengthen the interborder intrafirm integration and the
feedback of knowledge, expertise, and information which occurs within networks of
affiliates. On the other hand, sector-specific localization economies intensify intra­
border sectoral integration, implying local external networks between affiliates,
indigenous firms and local non-market institutions. By tapping into local knowledge
and expertise, foreign affiliates gain a competitive advantage which can not only be
exploited locally but may also be transferred back to the parent company, enhancing
its global technological competence. Thus, Narula and Zanfei (Ch. 12 in this volume)
refer to the recent shift away from asset-exploiting and towards asset-augmenting
investments by MNCs, which is typically associated with a greater dispersion of
innovative activity in the international network of an MNC. However, the nature
of the relationship between MNC international innovation systems and local
systems varies across regions (Cantwell and Iammarino 2000, 2001). This entails
different types of regional strategy for technological competitiveness.
Evidence has now emerged that the choice of foreign location for technological
development in support of what is done in the home base of the MNC depends upon
whether host regions within countries are either major centers for innovation or not
(termed “higher-order” or “ lower-order” regions by Cantwell and Iammarino
2000). Whereas most regions are not major centers and tend to be highly specialized
in their profile of technological development, and hence attract foreign-owned
activity in the same narrow range of fields, in the major centers much of the locally
sited innovation of foreign-owned MNCs does not match very well the specific fields
of local specialization, but is rather geared towards the development of general
purpose technologies (GPTs) that are core to cross-industry innovation today
(notably information and communication technologies, ICT) or in the past
(notably mechanical technologies). The need to develop such GPTs is shared by
the firms of all industries, and the knowledge spillovers between MNCs and local
firms in this case may be inter-industry in character. Thus, ICT development in
centres of excellence is not the prerequisite of firms of the ICT industries, but instead
involves the efforts of the MNCs o f other industries in these common locations.
Turning to competitiveness at the level of an individual firm, the determinants of
cross-company growth summarized in Table 20.6 derive from a cross-sectional
regression analysis o f 143 o f the world's largest firms between 1972 and 1982 (Cantwell
and Sanna-Randaccio 1993). As has been remarked earlier, a key aspect of innovation
and growth in the firm has to do with the largely industry-specific environment that
firms have in common and which regulates their individual behavior and partially
reflects their mutual interactions (Levin, Cohen, and Mowery 1985). Thus, the
growth of demand and o f technological opportunities in their own industry are
key influences on corporate performance.
5бО JOHN C A NT WE L L

Table 20.6 The statistically significant determinants of comfjarative grewvth

among the world's largest industrial firms, 1 9 7 2 -- 8 2 : 7

Regressor Signof coiefficient

own-industrydemand
G ro w th o f +
Growthofo w n -in d u s try technological +
opportunities
Firm size —

Firm-specific technological competitiveness

Degree o f market power +
Relativemultinationalityw ith in own-industry +
Increaseinmultinationalityoverperiod +

Source: Cantwell and Sarma-Randaceio (1993).

It is curious that although the w ork o f Penrose and Downie m entioned at the start
o f this chapter emphasized issues o f the creation o f firm -specific capabilities and
intra-industry competitive rivalry, until quite recently even that m inority o f econo­
mists that did work on firm growth paid relatively little attention to these issues.
However, now the notion o f corporate competence has m oved center stage in the
strategic management literature (see Lam, Ch. 5 in this volum e). Penrose had argued
that the competitive advantages o f a firm derive essentially from the cumulative and
incremental learning experience o f its management team, which experience differ­
entiates it from other firms. The distinctiveness o f the firm's accum ulated experience
and knowledge determines the set o f opportunities for growth which it perceives
ahead o f its rivals when screening the external environm ent (the growth o f demand
and technological opportunities in its industry). Corporate technological competi­
tiveness is the principal advantage o f this kind associated with differentiated learn­
ing (Cantwell 1989; Teece, Pisano, and Shuen 1997). A higher technological capability
lowers the unit costs and raises the demand curve o f the firm at a given rate of
growth, and it facilitates new entry Into related product lines. A m ore technologic­
ally competent firm is able to utilize its existing experience to lower the costs of
expanding its managerial and technical team in related areas. Technological com­
petitiveness comes out as one o f the most statistically significant influences upon
firm growth o f the variables listed in Table 20.6. In that large firm study, techno­
logical competitiveness was measured by the firm ’s share o f patenting in its industry
relative to its market share (o f industry output), in terms o f the relevant world
industry
 I N N O V ATION AND COM PETITIVENESS 561

20.5 C o n c l u s i o n s

To return to where this chapter began, competitiveness derives from the creation of
the locally differentiated capabilities needed to sustain growth in an internationally
competitive selection environment. Such capabilities are created through innov­
ation, and because capabilities are varied and differentiated, and since the creative
learning processes for generating capabilities are open-ended and generally allow for
multiple potential avenues to success, a range of different actors may improve their
competitiveness together. Innovation is a positive sum game that consists of the
efforts often of many to develop new fields of value creation, in which on average the
complementarities or spillovers between innovators tend to outweigh negative
feedback or substitution effects, even if there are generally at least some actors that
lose ground or fail. The basic conclusion is that efforts to promote competitiveness
through innovation can rarely be understood in isolation from what others are
achieving at the same time. This applies whether we are speaking of countries, of
national groups of firms in an industry, of subnational regions, or of individual
companies. Indeed, it is worth emphasizing that the degree of interaction between
innovators in search o f competitiveness has tended to rise substantially historically,
and has attained new heights in recent years.
Firms are less independent than they were, and they nowall float in a much deeper
sea of background knowledge, which Nelson (1992) refers to as the “public” element
of technology. There are at least four aspects to this: intercompany knowledge flows
have increased, there is a growing role for governments and other non-corporate
institutions in knowledge development and transfer, the importance of science for
technology has risen and diversified in its impact, and there has been a tendency
towards more rapid codification and the formation and spreading of professional
and scientific communities. We can now think of firms and the individuals aboard
them like ships floating in a sea o f public knowledge which connects them, or more
accurately potentially public knowledge since the extent that they can draw upon it
depends upon their own absorptive capacity and on their membership of the
appropriate clubs (whether intercompany alliances or professional associations
and the like). Over time, especially since 1945, firms have been designed to float
deeper down in the water, but they still always leave a critical part comprising their
own tacit capabilities above the surface, which does not sink down or fall into the
general mass. Indeed, holding stronger capabilities above the surface is positively
related to the depth to which one can reach below the surface, both for the absorptive
capacity to extract complementary knowledge and for the extent to which one
contributes oneself to the public knowledge pool. Universities and governments
have increasingly contributed to the sea of public knowledge as well. Additionally,
am°og firms that deliberately cooperate through technology-based alliances per­
sonnel can be exchanged so as to coordinate learning efforts.
562 JOHN C A N T W E L L

The sharing o f knowledge between firms implies not just that tech n o lo g y must be
d evelo p ed th ro u g h an interactive so cial an d cultural e v o lu tio n rather than through a
biological evolutionary process involving competition between g en etically inde­
pendent entities, but also that followers and innovative adapters may stand to make
greater gains than the original leaders in some new field o f technological endeavor.
For example, knowledge developed in on e co n text may ultimately p ro ve to h ave a
b ig g er im p a ct in another, w h ich was n o t foreseen b y the o rig in a to r o r even perh ap s
in itia lly b y the m o st in n o vatively su ccessfu l recipien ts. Firms also n o w d evo te m uch
greater effort to attem p ts to u n d erstan d th eir o w n tech n o lo g ica l p ra ctice a n d that o f
others. C o d ific a tio n o f k n o w led ge is the o u tco m e o f a co n sc io u s effo rt, sh iftin g back
the d iv id in g lin e betw een w h at is p o ten tially p u b lic an d w h at is tacit (Cohendet and
S tein m u eller 2000; C o w a n , D a v id , a n d F o ra y 2 0 0 0 ). So firm s th at b e c o m e especially
adept at co d ificatio n m a y fin d that this is a so u rce o f co m p e titive ad van tage since
th ey can then m o re rea d ily d raw o n the p u b lic p o o l.
To engage in this in terco m p an y in teractio n fru itfu lly, firm s m u st m ain tain an
adequate d iversificatio n o f th eir in -h o u se tech n o lo g ical effo rts, since the clo ser that
k n o w led ge is to the p ro p rie ta ry interests o f a firm the m o re lik ely th at it w ill o n ly be
sh ared in return fo r so m eth in g else that is p ro b ab ly tech n o lo g ica lly co m p lem en tary,
w h ich is w h at each firm needs to jo in the relevant co rp o ra te clu b (C an tw ell and
B arrera 1998). T h e en trep ren eu rial fu n ctio n has n o t been e lim in a ted b u t it is more
in stitu tio n a lly em b ed ded in an ab ility to n etw o rk an d m ak e n e w co n n ectio n s (see
P ow ell an d G ro d a l, C h . 3 in this vo lu m e ).
T h e in teractio n effects betw een in n o va to rs h as b een fu rth er c o m p o u n d e d b y the
role o f IC T as a m eans o f co m b in in g fields o f k n o w led g e creatio n that w ere p revi­
o u sly kept largely ap art (o r w h at K o d a m a 1992, term s te c h n o lo g y fu sio n , an d has led
to the creation o f new fields such as bioinformatics). IC T thus broadens the field for
potential innovation by linking formerly separate areas o f innovative activity, ICT
p o ten tially co m b in es the v a rie ty o f tech n o logical fields th em selves and so increases
the sco p e fo r w id e r in n o vatio n .
So, in the light o f these recent changes in em p h asis in the co n text fo r in n o vation
and competitiveness, how can we reevaluate some o f the earlier literature in the field
that has been su m m arized here? W h ich lines o f research h ave p erh ap s run o u t o f
steam fo r the tim e being, an d w h ich offer the m o st p ro m isin g n ew avenues o r
o p p o rtu n ities fo r fu tu re research? T h e revival o f the classical issues o f co m p etitive-
ness at a n ation al level (Sectio n 20.2) has been u sefu l in that it represents a return o f
interest into the m a jo r q u estion s o f the w ealth o f n ation s, w h ich are o f p rim ary
im p o rtan ce fro m a social v iew p o in t. H ow ever, this n ew literatu re h as also exposed
the lim itatio n s o f tryin g to tackle w h at are really issues o f stru ctu ral ch an ge at an
aggregate kveL It might perhap s be ju d g ed that w e h ave g o n e a b o u t as fa r as w e can
for n o w w ith pu rely cro ss-co u n try m odels. It may w ell be fo r th is reaso n that the
fo cu s o f the latest research has tended to shift to the in d u stry an d firm levels,
allo w in g fo r (indeed n o w em p h asizin g m o re) the sco p e fo r tech nological
interactions betw een firm s an d in d u stries, so as not to lose sight o f the w ider context
or the aggregate effects.
When recasting the an alysis at an in d u stry level (Section 20.3), we now know a
good deal about h isto rical sh ifts in p attern s o f in d u strial leadership betw een cou n ­
tries, about the role o f e d u c a tio n an d skills in catch in g up econom ies, and how
particular kinds o f skills h elp to a cco u n t fo r in ter-in d u stry discrepancies in in n o va­
tive potential A rg u a b ly w e still k n o w too little ab out the interaction between
governments, n o n -b u sin ess in stitu tio n s, an d firm s (especially large firm s) in the
process o f establishing co m p etitiven ess. In particu lar, w e w o u ld like to kn o w m ore
about u n iv e rsity -in d u stry (sc ie n c e -te c h n o lo g y ) in teraction over a w id er range o f
countries, beyond the rela tiv ely clear p ictu re that w e have fo r the U S (see M ow ery
and Sampat, C h. 8 in th is v o lu m e ). T h e con text here is w hat seem s to be the grow ing
significance o f a local scien ce b ase fo r the co n stru ctio n o f corporate capabilities and
hence com petitiveness, in c lu d in g an d p erh ap s especially in latecom er econom ies.
Note that this n ew ly e m e rg in g v ie w reverses the “ trad itio n al” perspective that
developing cou n tries sh o u ld co n cen trate o n (o rgan izatio n al in n ovation in) low er-
skill activities, and leave scien ce to the largest m o st developed econom ies.
Coming to com p etitiven ess at the firm o r clu ster level (Section 20.4), the latest
research has also h ad a ren ew ed fo cu s o n the role o f in tercom pan y interaction in
knowledge creation an d in n o v a tio n , esp ecially in su b n atio n al regional areas, and
through alliances o r c o o p e ra tiv e agreem en ts. H ere w e n o w kn o w m ore o f the details
of the localized ch aracter o f in n o v a tio n , an d o f the steady grow th in technology-
based alliances as a m ean s o f facilititatin g com p etitiven ess through know ledge
exchange and spillovers. W o rk o n firm size an d in n o vatio n or grow th seems to
have rather run o u t o f steam fo r n o w , at least in so far as it had regarded in dividual
firms as quite in d ep en d en t en tities. W e need to k n o w m o re ab o u t the specificities o f
knowledge flow s b etw een reg io n s an d b etw een firm s, o f h o w and w here tech no­
logical knowledge is so u rce d b y firm s, an d then h o w such know ledge is effectively
combined in n etw o rks o f in terrelated in n o v a tio n w ith in and betw een firm s. This is
surely an exciting agen d a fo r fu rth e r research.

Re fe r e n c e s

Abramowitz, M. (1956), “ Resources and Output Trends in the United States since 1870,”
46(1): 5-23.
American Econom ic R e v ie w
Albach, H. (1996), “ Global Competitive Strategies for Scienceware Products, in G. Koop-
mann and H.-E. Scharrer (eds.), The Economics of High-Technology Competition and
Cooperation in Global Markets, Baden-Baden: Momos Verlagsgesellschaft, 203-17.
Audretsch, D. B. (1996), “ International Diffusion of Technological Knowledge, in
G- Koopmann and H.-E. Scharrer (eds.), The Economics of High-Technology Competition
and Cooperation in Global Markets, Baden-Baden: Momos Verlagsgesellschaft, 107-35.
564 JOHN C A NT WE L L

B a u m o l , W. J., N e l s o n , R. R ., and W o l f f , E. N . (eds.) (i994)> C o n v e rg e n c e o f P r o d u c tiv ity :

C ro s s -N o tio n a l S tu d ie s a n d H is to ric a l E v id e n c e , O xfo rd an d N ew York: O xfo rd U niversity
Press.
* C a n t w e l l , J, A. (1989), T ech n o lo g ica l In n o v a tio n a n d M u lt in a t io n a l C o rp o ra tio n s , Oxford:
Basil Blackwell.
-------(1992), “ Ja p a n s Indu strial C o m petitiven ess an d the T echnological C apab ilities o f the
leading Japanese F ir m s ” in T. S. A rriso n , C. F. Bergsten, E. M . G rah a m and M . C. H arris
(eds.), Ja p a n 's G r o w in g T ech n o lo g ica l C a p a b ilit y : Im p lic a tio n s f o r th e U S E c o n o m y W ash­
ington, D C : N ation al A cad em y Press, 16 5-88.
-------(2000), “ T echnological L o ck-in o f Large Firm s since the In terw ar Period,” E u ro p e a n
R e v ie w o f E c o n o m ic H is to r y 4(2): 14 7-74 .
-------and B a c h m a n n , A . (1998), “ C h an gin g Patterns o f Techn ological Leadership: Evidence
fro m the Pharm aceutical Industry,” In te r n a tio n a l J o u r n a l o f In n o v a t io n M a n a g e m e n t 2(1);
4 5 —7 7 *
-------an d B a r r e r a , M . P. (1998), “ The Localisation o f C o rp o rate T echn ological Trajectories
in the In terw ar C artels: C o o p erative Learnin g versus an Exch an ge o f K now ledge,” E co ­
n o m ic s o f In n o v a tio n a n d N e w T ech n o lo g y 6 (2 -3 ): 257-90.
-------an d I a m m a r in o , S. (2000), “ M u ltin atio n al C o rp o ratio n s and the L o cation o f Techno­
logical In n o vatio n in the UK Regions,” R e g io n a l S tu d ie s 34(4): 317-22.
— -----------(20 0 1), “ E U R egions and M u ltin ation al C o rp o ratio n s: C h an ge, Stability and
Strengthening o f Technological C o m p arative A dvan tages,” I n d u s t r ia l a n d C o rp o ra te
C h a n g e 10 (4): 10 0 7 -37 .
------------ and N oonan, C. A . (2001), “ Sticky Places in Slippery Space—the Location o f
In n o vatio n b y M N C s in the European R egions,” in N . Pain (ed.), I n w a r d In v e s tm e n t ,
T ech n o lo g ica l C h a n g e a n d G ro w th : T h e Im p a c t o f M N C s o n th e U K E c o n o m y ; O xford:
P ergam on, 210 -39 .
* — — and S a n n a - R a n d a c c i o , F. (1993), “ M u ltin atio n ality an d firm grow th,” W eltw irtsch af-
tlich es A r c h iv 129(2): 275-99.
C o h e n d e t , R , and S t e i n m u e l l e r , W. M . (2000), “ T h e C o d ific a tio n o f Knowledge:
A C o n cep tu al and E m p irical E xp lo ratio n ,” I n d u s t r ia l a n d C o r p o r a te C h a n g e 9(2): 195-209.
C ow an , R ., D a v id , P. A ., and F o r a y , D. (2000), “ The Explicit Economics of Knowledge
Codification and Tacitness,” In d u s t r ia l a n d C o rp o ra te C h a n g e 9(2): 211-53.
D e n i s o n , E. E (1967), W h y G ro w th R a te s D iffe r : P o s t-W a r E x p e r ie n c e in N in e W estern
C o u n trie s, W ashington, D C : B ro o kin gs Institute.
D o s i, G ., P a v i t t , K. L. R., and S o e t e , L. L. G . (1990), T h e E c o n o m ic s o f T ech n ica l C h a n g e a n d
In te r n a tio n a l T ra d e , Lon don : H arvester W heatsheaf.
-------and S o e t e , L. L. G. (1988), “ Technical C h an ge and In tern ation al Trade,” in G. Dosi,
C. Freem an, R. R. N elson, G . Silverberg, and L. L. G . Soete (eds.), T ech n ica l C h a n g e a n d
E c o n o m ic T h e o r y Lon don : Pinter.
D o w n i e , J. (1958), T h e C o m p e titiv e Process, L o n d o n : D u ckw orth .
E nos, J. L., and P ark , W. H . (1988), T h e A d o p tio n a n d D iffu s io n o f Im p o r t e d T ech n o logy: T h e
C a s e o f K o re a , Lon don: G ro o m H elm .
E n r i g h t , M . J. (1998), “ R egional Clusters and Firm Strategy,” in A . D. C handler, R Hag-
strom , and 6 . Solved (eds.), T h e D y n a m ic F ir m : T h e R o le o f T e c h n o lo g y S t r a te g y O rg a n iz a ­
tion a n d R e g io n s , O xfo rd and N ew York: O xfo rd U n iversity Press, 3 15 -4 2 .

* Asterisked items are suggestions for further reading.

 __________^ n o vat i on and c o m p e t i t i v e n e s s 565

Fagerberg , J. (1987), “A Technology Gap Approach to Why Growth Rates Differ,” Research
Policy 16(1): 87-99*
*(1988), “International Competitiveness/5Econom ic Jo u rn a l 98: 355-74.
*(2002), Technology, G row th a n d C om petitiveness: Selected Essays, Cheltenham: Edward
Elgar.
■---- Verspagen, B., and C aniels , M. (1997)» “Technology, Growth and Unemployment
across European Regions/5 R eg io n al Studies 31(5): 457-66.
Freeman, C , and LougA, F. (2001), As T im e Goes B y: From the Industrial Revolutions to the
Information R evolution , O xford and New York: Oxford University Press.
---- and Perez, C (1988), “ Structural Crises of Adjustment, Business Cycles and Investment
Behaviour/5in G. Dosi, C. Freeman, R. R. Nelson, G. Silverberg, and L. L. G. Soete (eds.),
Technical Change a n d E con om ic Theory ; London: Pinter 38-66.
Hagerstrand, T. (1967), In n ovatio n D iffusion as a Sp atial Process , Chicago: University of
Chicago Press.
Kline, G . a n d R osenberg , N. (1986), “An Overview of Innovation/5in R. Landau and N.
Rosenberg (eds.), The Positive Su m Strategy: H arnessing Technology fo r Econom ic G row th ,
Washington, DC: National Academy Press, 275-306.
Kodama, F. (1992), “ Technology Fusion and the New R&D,” H arva rd Business R eview
(July-August): 70-8.
*Krugman, P. R. (і 994«)> “ Competitiveness: A Dangerous Obsession/5Foreign Affairs 73(2):
28-44.
---- (1994b), “The Myth of Asia’s Miracle/5Foreign A ffairs 73(6): 62-78.
---- (1996), “Making Sense of the Competitiveness Debate/5 O xford R eview o f Economic
Policy 12(3): 17-25,
*Lall , S. (2001), C om petitiveness, Technology a n d Skills , Cheltenham: Edward Elgar,
Landes , D. S. (1969), The U n b o u n d Prom etheus: Technological a n d Industrial D evelopm ent in
Western Europe fro m i j s o to the Present , Cambridge and New York: Cambridge University
Press.
Lazonick ,W . (1991), Business O rgan ization a n d the M yth o f the M arket Economy, Cambridge
and New York: Cambridge University Press,
*(1992), “Business Organization and Competitive Advantage: Capitalist Tranformations in
the Twentieth Century/5in G. Dosi, R. Giannetti, and P. A, Toninelli (eds.), Technology and
Enterprise in a H istorical Perspective , Oxford and New York: Oxford University Press,
119-63,
~~— (1993)) “Industry Clusters versus Global Webs: Organizational Capabilities in the
American Economy/5In d u stria l a n d C orporate C hange 2(1): 1-24.
■"— (1998), “Organizational Learning and International Competition/5 in J. Michie and
I* Grieve-Smith (eds.), G lo ba lizatio n , G row th a n d G overnance , Oxford and New York.
Oxford University Press, 204-38.
---- and O’Suleivan, M. (1997), “ Big Business and Skill Formation in the Wealthiest
Nations: The Organizational Revolution in the Twentieth Century, in A. D, Chandler,
E Amatori, and T. Hikino (eds.), B ig Business a n d the W ealth o f Nations, Cambridge and
New York: Cambridge University Press, 497-521.
Levin, R. C , C ohen, W. M., and M owery , D. C. (1985), “ R&D Appropriability, Opportunity
and Market Structure: New Evidence on Some Schumpeterian Hypotheses/5 Am erican
Economic R ev iew 75: 20-4,
$66 JOHN C A NT WE L L

L u n d v a l l , B. A. (1988), “ In n o vatio n as an Interactive Process: F ro m U se r-P ro d u c e r Inter»

action to the N ation al System o f In n o vatio n ,” in G . D o si, C. Freem an, R. R. N elson,
G . Silverberg, an d L. L G. Soete (eds.), T ech n ica l C h a n g e a n d E c o n o m ic T h eo ry, London:
Pinter, 349-69,
— — (ed.), N a t io n a l S ystem s o f In n o v a tio n , L o n d o n : Pinter,
M a l m b e r g , A ., S o l v e l l , 6 ., and Z a n d e r , I. (1996)1 “ Spatial C lu sterin g, L o cal A ccu m u lation
o f K n ow led ge and Firm C om petitiven ess,” G eo g ra fisk a A n n a le r j S в (2): 85-97.
M a r r i s , R. (1964), T h e E c o n o m ic T h e o r y o f M a n a g e r ia l C a p ita lis m , L o n d o n : M acm illan .
M o w e r y , D, C ., an d N e l s o n , R. R. (1999), S o u rces o f I n d u s t r ia l L e a d e r s h ip : S tu d ie s o f S e v e n
In d u strie s , C am b rid ge: C am b rid ge U n iversity Press.
N e l s o n , R. R. (1992), “ W hat is ‘ C o m m ercial’ and W hat is 'P u b lic 5 ab o u t Technology, and
W hat Sh o u ld Be?” in N . R osenberg, R. Lan dau, and D, C . M o w e ry (ed s.), T ech n o lo g y a n d
th e W ealth o f N a tio n s , Stanford: Stan fo rd U n iversity Press, 57 -7 1.
-------(i995)> “ C o -evo lu tio n o f In d u stry Structure, T echn ology an d S u p p o rtin g Institutions,
and the M ak in g o f C o m p arative A dvantage,” In te r n a tio n a l J o u r n a l o f th e E c o n o m ic s o f
B u s in es s 2 :17 1- 8 4 .
* -------and P a c k , H. (1999), “ The A sian M iracle an d M o d ern G ro w th T heory,” E co n o m ic
Jo u r n a l 109: 4 16 -36 .
O z a w a , T. (1974), Ja p a n 's T ech n o lo g ica l C h a lle n g e to th e W est, 1 9 5 0 - 1 9 7 4 : M o tiv a t io n a n d
A c c o m p lis h m e n t, C am b rid ge, M ass,: M IT Press.
P e n r o s e , E. T. (1959), T h e T h e o ry o f the G ro w th o f th e F ir m , O xfo rd : B asil Blackw ell,
* P o r t e r , M . E. (1990), T h e C o m p e titiv e A d v a n t a g e o f N a tio n s , N ew York: Free Press,
-------an d S 5l v e l l , 6 . (1998), “ The Role o f G eo grap h y in the Process o f In n o vatio n and the
Sustainable C o m p etitive A dvantage o f Firm s,” in A . D. C handler, P. H agstrom , and
6 , Solvell (eds.), T h e D y n a m ic F ir m : T h e R o le o f T ech n o logy, S tra te g y ; O rg a n iz a tio n a n d
R e g io n s , O xfo rd and N ew York: O xfo rd U n iversity Press, 44 0 -57.
P o s n e r , M . V. (19 6 1), “ International Trade and Technical C h ange,” O x fo r d E c o n o m ic P a p ers
13(3): 3 ^3 - 4 1 .
P r a i s , S. J. (1995), P ro d u c tiv ity , E d u c a tio n a n d T ra in in g : F a c ts a n d P o lic ie s in In te rn a tio n a l
P ersp ective, Cambridge and New York: Cambridge University Press.
R i c h a r d s o n , G . B. (1972), “ The O rgan ization o f Industry,” E c o n o m ic J o u r n a l 82: 883-96.
S c o t t , A . J. (1998), “ The G eo grap h ic F o u n d atio n s o f In d u strial P erform ance,55 in A. D.
C handler, P. H agstrom , and 6 . Solvell (eds.), T h e D y n a m ic F ir m : T h e R o le o f Technology,
Strategy, O rg a n iz a tio n a n d R eg io n s, O xfo rd and N ew York: O xfo rd U n iversity Press,
384 -40 1,
S o e t e , L. L. G. (1981), “A G eneral Test o f Technological G ap Trade T h eo ry,” W eltw irtsch a f-
tlich es A r c h iv 117(4 ): 638-60.
So l o w , R. (1957), “ Technical C hange and the A ggregate P ro d u ctio n Fu n ction ,” R e v ie w o f
E co n o m ic s a n d Statistics 39: 312 -2 0 .
* T e e c e , D. J., P i s a n o , G ,, and S h u e n , A , (1997), “ D yn am ic C ap ab ilities and Strategic
M a n a g e m e n t” S tra te g ic M a n a g e m e n t Jo u r n a l 18(7): 537-56.
T y s o n , L. D A, (1992), W h o s B a s h in g W h o m : T ra d e C o n flic t in H ig h T ech n o lo g y In d u stries ,
Washington, DC: Institute for International Economics.
V e r n o n R. (1966), “ International Investm ent and In tern ation al Trade in the P rod u ct Cycle ”
,

Q u a rte rly Jo u r n a l o f E co n o m ic s 80(2): 19 0 -20 7.

(! 9 7 4 )> T he Location o l E co n o m ic A c tiv ity ” in J. H. D u n n in g (ed.), Economic Analysis
a n d the M u lt in a t io n a l E n terp ris e, Lon don : Allen an d U n w in , 8 9 -114 .
 CHAPTER 21

INNOVATION AND
EMPLOYMENT

MARIO PI A NTA

21 . 1 I n t r o d u c t i o n

T h e relatio n sh ip betw een in n o va tio n an d e m p lo y m e n t is a c o m p le x o n e an d has

lo n g been a top ical issue in eco n o m ic th e o ry .1 M o v in g fro m the classical question
“ d oes tech n o lo g y create o r destroy jo b s?” recent research h as in vestigated the im pact
o f d ifferen t types o f in n o vatio n and the stru ctu ral an d in stitu tio n a l factors affecting
the q u a n tity o f e m p lo ym en t change. Q u a lity aspects h ave received in creasin g atten­
tio n , w ith q u estion s o f “ w h at type o f jo b s are created o r d estroyed b y in n o vatio n ?”
T h is lin e o f research has asked, “ h o w does the c o m p o sitio n o f skills ch an ge” and
“ h o w does the w age stru ctu re change,” lead in g to a large literatu re o n skill biased
tech n ical change an d o n w age p o larizatio n .
T h is ch ap ter exam in es the en o rm o u s b o d y o f sch o la rly research o n this topic
w ith in the ad van ced econ om ies. First, the perspectives, scope, an d types o f in n o v ­
atio n s are co n sidered, id en tifyin g the d ifferen t e m p lo y m e n t effects th ey m ay have.
Secon d, the effects on the qu an tity o f em p lo y m e n t are review ed at the firm , industry,
and m a cro e co n o m ic level. T h ird , ch an ges in the q u a lity o f e m p lo y m e n t are e xam ­
ined, co n sid erin g the effects on skills an d w ages, an d the im p a ct o f organ ization al
in n o vatio n , again at d ifferent levels o f analysis. A su m m a ry o f stylized facts c o n ­
clud es the chapter w ith a d iscu ssio n o f fu tu re research issues.
 _______________ ___ IN N O VATIO N AND EMPLOYMENT 569

21.2 P e r sp e c t iv e s and Scope

o f In n o v a tio n s

The literature on innovation and employment has addressed different research

questions rooted in several streams o f research. Table 21л summarizes the main
perspectives in terms o f assum ptions as to the economic system, methodologies, and
levels of empirical analysis (for a review o f theoretical approaches, see Petit 1995),
The mainstream approach looks at innovation as a change in technology across
the economy, leading to econom ic growth and employment. The lack o f an explicit
conceptualization o f technology, either in the decisions o f private agents, or in the
tools of public policy, and its exogenous nature remain, however, a limitation o f
Keynesian growth theories, as well as o f Solow’s growth model (see Ch. 18 by
Verspagen in this volum e). M odern theories have approached the question o f the
employment impact o f technology from two perspectives. On the one hand, growth
theories have moved from an exogenous view o f technological change to efforts, in
the new g ro w th theory.>to conceptualize innovation— proxied by technology, learn­
ing, and educational variables— as an engine o f endogenous growth. Labor econo­
mists, on the other hand, have explained changes in employment (and wages) with
reference to the dem ography o f jobs, m acroeconom ic factors, wage costs, bargaining
modes, and the flexibility o f labor markets, m oving later to consider competitiveness
and technology factors (see Table 21.1).
However, the fundam entally disequilibrating nature o f technological change is
usually treated in a context that still assumes a general (or partial) equilibrium o f
markets, that is, all output finds its demand, and all workers ready to accept the
current wage find em ploym ent. Technological change is often reduced to new
production processes (and new production functions), with the models rarely
envisaging the emergence o f product innovation. When employment losses appear
in such studies, they rarely lead to perm anent (or structural) unemployment; rather,
they lead to downward adjustm ents in wages so that the additional jobless are
returned to work. I f this state o f affairs cannot be found in the real world, then the
responsibility is attached to the lack o f flexibility o f labor markets, with excessive
union power or institutional rigidities, such as the m inim um wage.
A more convincing approach to the study o f innovation and its consequences is
one that addresses from the start the d is e q u ilib r iu m nature o f economic change. This
Hew has been developed by neo-Schum peterian perspectives, by Kaldorian, Struc­
tural, and Evolutionary approaches (see Table 21.1).
Neo-Schumpeterians have argued that advanced economies are witnessing the
emergence o f a new techno-econom ic paradigm based on Information and Com-
munication Technologies (ICTs) (see Ch. 1 by Fagerberg in this volume). Such
powerful technological changes do create and destroy a large amount o f jobs. The
Table 2 1 Л A summary of approaches to innovation and employment
Main research questions
Does technology create or destroy jo b s ?
Equilibrium of product and labor markets
What is the amount of jobs created/lost
What is the skill composition
What is the structure of wages
What are the returns from innovation
What is the innovation input to growth
Disequilibrium perspectives
What is the type of innovation
What is the amount/nature of unempl.
What are the structural factors
What are the demand factors
What are the distribution effects
What are the institutions
Genera! approach Major streams of literature and key findings Key assumptions and methodology Main level of
analysis
Labour economics Job demography and flexibility of labor Product and labor markets are in Firms, industries,
markets equilibrium macroeconomy
Technical change favors more skilled The absolute level of job lost/
workers, replaces the unskilled, exacerbates gained is irrelevant
inequality
Supply of educated workers shapes technical Complementarity between ICTs
change and high skills
Technological unemployment is irrelevant
Growth Technology, productivity, growth, Standard production function, Industries, macro
employment: innovation may raise the focus on process innovation
natural rate of unemployment
New growth theory Endogenous innovation, growth and Innovating and non innovating Macroeconomy
employment: unemployment may happen firms, spillovers, focus on process
innovation
Evolutionary Technological opportunities, variety, Innovation brings disequilibrium Firms, industries
regimes: firms’ strategies and industries' in markets
outcomes are different New product markets emerge
Neo-Schumpeterian Techno-economic paradigms and long Radical innovations, Industries, macro
waves: mismatches can lead to pervasiveness, diffusion of new
unemployment technology systems and ICTs
Structural Sectoral composition of economies: Innovation is differentiated: Industries
specificity of innovation and demand, constrasting effects of new
different job results products and processes
Regulationist Macro models for testing indirect effects of Industries are different, demand is Macroeconomy
innovation: compensation mechanisms may important
not work Countries are different,
institutions are important
 inno vation and employme NT 571

question as to where and to what extent jobs are created or destroyed depends upon
the highly dynamic process that shapes the content o f specific technological innov­
ations and the speed o f their adoption, with an often blurry distinction between the
two (see below). Moreover, jobs lost and new jobs offered may take place in different
areas or require different skills, leading to mismatches. The speed o f adjustment is
therefore crucial and makes the difference between frictional unemployment (easily
absorbed by well-functioning labor markets) and technological unemployment.
Sustained and sustainable growth can be expected only once the mismatches
between the new technologies and the old economic and social structures and
institutions are overcome, with a two-way adjustment. Innovation has to be adapted
to social needs and econom ic demands; economic and social structures evolve under
pressure from new technologies. New technologies need to be matched by organiza­
tional changes, new institutions and rules, learning processes, the emergence o f new
industries and markets, and the expansion o f new demand. Several studies on the
emergence o f technological paradigm s and key technologies in the past have pointed
to the long time required before the impact o f these elements (positive or negative)
on economic growth and em ploym ent become evident (Freeman, Clark, and Soete
1982; Freeman and Soete 1987, 1994; Freeman and Louya 2001). While extremely
powerful in its explanation o f long term economic changes and historical evolution,
this approach has yet to be “ operationalized” with more specific questions on the
type of innovation and on the interaction with economic and employment variables.

Box 21.1 Technology and unemployment: a classical debate

Since the emergence of the Industrial Revolution, the extensive substitution of labor by
tnachinery incorporating the new technology of the time has led economists and policy
makers to debate the economic and social consequences. At the end of the eighteenth
century, James Steuart drew attention to the difficulty of reabsorbing the unemploy­
ment caused by sudden mechanization, in spite of the positive effects from the
construction o f new machines and price reductions, and already envisioned a role
for the government. Adam Smith linked the invention of machines to the division of
labor and emphasized its labor-saving effects. Jean-Baptiste Say had less doubts about
the ability of markets to adjust, while Thomas Malthas emphasized the positive effects
resulting from the strong demand dynamics experienced by England at the time. The
optimism of classical economists in the early nineteenth century contrasted with the
dramatic impoverishment of the English working classes—industrial workers, small
artisans, and displaced peasants— who had started to organize trade unions and to
s launch Luddite straggles against the job losses and deskilling brought about by
, :fccbamzation. David Ricardo was convinced that the economy could compensate
the negative employment effects, but in a famous passage in the chapter On machin­
e r y added in the third edition of his Principles of Political Economy and Tamttm%
572 MARIO P I A N T A

argued that “ The opinion, entertained by the labouring class, that the employment of
machinery is frequently detrimental to their interests, is not founded on prejudice and
error, but is conformable to the correct principles of political economy* (Sicardo
1951:392), '
The most articulate criticism of compensation theory was developed by Karl M as,
who emphasized the losses for workers in terms o f jobs, skills, wages, and control over
their work resulting from the way mechanization was proceeding at the time. Arguing
that unemployment grows as technical change displaces labor more rapidly that the
accumulation o f capital demands new workers, M a s developed important insights on
the functioning o f capitalism. The drive to capital accumulation leads to a constant
search for new production techniques and new products (a key starting point in Joseph
Schumpeter’s theory o f innovation). High unemployment assures lower wages and
greater control over workers, but capital accumulation ultimately encounters the
problems o f finding adequate markets and demand, and making adequate profits
(for a reconstruction o f the debate see Heertje 1973; Vivarelli 1995).

21.2.1 New Products, Processes and Organizations

Sch u m p eter (1934) d efin ed p ro d u ct innovation as “ the in tro d u c tio n o f a new
g o o d . . . o r a n ew q u a lity o f a good ,” an d process inn ovation as “ the in tro d u ctio n
o f a n ew m eth o d o f p ro d u ct i o n . . . o r a new w a y o f h an d lin g a c o m m o d ity co m m er­
cially.” 2 T h e d evelo p m en t (o r the ad o p tio n ) o f process in n o va tio n s leads to greater
efficien cy o f p ro d u ctio n , w ith savin gs in lab o r an d /o r cap ital, an d w ith a potential
fo r price red uctions. T h e u sual o u tco m e is h igh er p ro d u c tiv ity and loss o f em p lo y­
m ent; to the extent that process in n o vatio n s increase p ro d u c t q u a lity o r reduce
prices, a rise in d em an d (w h en elasticity is high) m ay result in m o re jobs.
N ew p ro d u cts (o r services) can be rad ical in n o va tio n s (n ew to the w orld),
in crem en tal im p ro vem en ts o n p re vio u s in n o vatio n s, o r im ita tio n o f g o o d s already
p ro d u ced in o th er co u n tries o r firm s. G enerally, p ro d u ct in n o va tio n s increase the
q u a lity and v ariety o f go od s an d m ay open up new m arkets, lead in g (w h en elasticity
is high) to greater p ro d u ctio n an d em p lo ym en t. B u t n ew p ro d u cts can sim p ly
replace old ones, w ith lim ited eco n o m ic effects, o r be designed in o rd er to reduce
costs, w ith an im pact sim ilar to process in n o vatio n s (P ian ta 2 0 0 1).
N ew go od s enter the e co n o m y as c o n su m p tio n go o d s, in term ed iate go od s, or
investm ent goods, fo llo w in g the dem and s o f co n su m ers, firm s, an d investors.
In n o vative investm ent go od s have a d u al nature: they start as n ew p ro d u cts in the
in d u stries p ro d u cin g them , but b ecom e process in n o va tio n s in the industries
acq u irin g them . T h eir em p lo ym en t consequences are likely to be p o sitive in the
m ach in ery p ro d u cin g sectors, an d negative (w hen d em an d offsets are in su fficien t, as
noted above) in the indu stries m ak in g new investm ents (E d q u ist, H o m m en , and
M cK elvey 2001).
 ______ _ _ INN° v л т Iо N A N D EMPL OYME NT 573

The distinction between process and product innovations should not be carried
too fan Most innovative firm s introduce both at the same time, but in most firms and
industries (see Ch. 14 by M alerba in this volume) it is possible to identify the
dominant orientation o f innovative efforts, associated with strategies o f either
p ric e competitiveness (and m ainly process innovations) or technological competi­
tiveness (and m ainly product innovations). In addition to product and process
innovation, organizational innovation also can affect the quantity and quality o f
employment, and is usually closely linked to the introduction o f new technologies
(Caroli 2001). Sections 21.22 and 21.3 below will consider the impact on job creation
and loss o f each o f these types o f innovation.

21*2.2 Innovation, Imitation, Adoption, and Use

The employment consequences o f new technologies obviously require the applica­
tion of these technologies in production activities. Firms innovate when they first
market a new product or introduce a new process; followers in the same industry
may imitate them (perhaps with incremental improvements). Firms may adopt new
processes or use new products generated in other industries, leading to the diffusion
of innovations— and to their em ploym ent impact— throughout the economy (see
Ch. 17 by Hall in this volum e).
For empirical analyses, this distinction can be captured either by studies on the
emergence o f particular technologies, following their evolution through all the
previous steps, or b y studies on firms (or industries) based on surveys that identify
product innovations novel to the market (original innovators) and those new to the
firm only (imitators), and the introduction o f new processes (adopters). Traditional
indicators such as R & D and patents fail to capture a large part o f the latter two
modes of technological development.

ІШ 21,2 Evidence on innovation and employment

Conceptual and empirical difficulties have long led economists to adopt a rather
homogeneous view of innovation, described either by R&D expenditure (one of its
inputs), or proxied by patenting activity (one of its output). In the last decade, the
Spread of innovation surveys in Europe and of surveys on panels of firms in the US has
provided important new evidence on the variety of innovative activities.
The key results for Europe in the period 1994-6 are that innovation is present in 51 per
centof manufacturing firms and in 40 per cent of services firms (see Chapters 6 and 16 by
Smith and M ies in this volume). Close to 40 per cent of manufectoring firms
Aange their production processes, while a fifth develop minor product improvements
 Industry cades

01 Food products
02 Textiles
03 Apparel
Percentage change in employment, 1994-2000

0 4 Leather and footwear

05 Wood
06 Pulp and paper
07 Printing and publishing
08 Petroleum products
09 Chemicals
10 Rubber and plastic
11 Non metallic minerals
12 Basic metals
13 Fabricated metal products
14 Machinery and equipment
15 Office, computing machinery
16 Electrical machinery
17 Communication equipment
18 Precision instruments
19 Motor vehicles
20 Other transport equipment

Countries
FR France
IT Italy
NL the Netherlands
UK United Kingdom
Share of new products in sales,1994-1996

Fig. 21-1 Share o f new products in sales and employment change

Source: CIS2-SIEPI Innovation Database, University of Urbino and OECD StAN data.
 INNOVATION AND EMPLOYMENT
575

yaffil another fifth introduce products new to their market, whose impact on sales is
limited to 12 per cent. In all indicators, important differences are found across
countries, associated with the nature of their industrial structure and national innov­
ation system (European Commission -Eurostat 2 0 0 1 ).
The latter variable offers the most accurate description o f the economic relevance o f
innovations, and is related, in Figure 2 1 . 1 , to the employment performance (average
annual rate of change) of 20 manufacturing industries in four EU countries (France,
Italy, the Netherlands, and the UK). Looking at industry data for the share of new
products in sales (drawn from the CIS 2 -SIEPI database, which provides data at the
two-digit industry level for major countries), a very strong variability is evident, from
more than 20 per cent in office computing and telecommunications, to close to zero in
more traditional industries (again, with important cross-country differences due to
national specializations). Looking at overall employment change, we first see that,
between 1994 and 2 0 0 0 , the majority o f industries lost jobs. The relationship between
product innovation and employment experienced in European industries looks like a
positively oriented curve, but a closer look at the distribution of cases is important.
ІСГ industries (computing, telecommunications, precision instruments, and other
transport, including aerospace) are generally in the top right quadrant, where new
products in sales have a share higher than the average value for the EU and where no
dramatic job losses are found. As expected, industries characterized by the new technol­
ogy (and by a high concentration of product innovations) show the highest impact of
new products in their turnover and better employment performances. In a few cases,
however, moderate job losses are found even in this group, as strong international
competition may lead to the decline of some ICT industries in some countries.
Traditional industries (textiles, wearing apparel, leather, and a few others) tend to
concentrate in the bottom left quadrant, where a below average innovativeness (and a
strong dominance o f process innovations) is matched by dramatic job losses.
The remaining sectors, in the top left quadrant, combine a low or intermediate impact
of innovation with modest job losses or substantial gains, showing again a positive
association. While several factors alongside innovation affect employment change
(macroecomic conditions, competitiveness, etc., see Pianta 2000), the distribution of
Figure 21Л highlights, on the one hand, the generally positive link we expect between
product innovation and jobs; on the other hand, it shows the presence of winners and
losets maU industries, reflecting the importance o f national specializations, of economic
АпШіге$, and o f the intensity o f international competition in open economies.

21.3 T h e E f f e c t s o n t h e Q u a n t i t y
of Employment

The relationship between innovation and jo b s is investigated in this section, w hich

looks at the im p act on the quantity o f employment, defined in terms o f the number
of existing jobs, or, m ore accurately, in terms o f the total hours o f work. T h is link can
57& MARIO P I A N T A

be exam in ed at d ifferen t levels o f analysis: firm s, in d u stries, an d the aggregate

econ o m y. Table 21.2 su m m arizes the m o st relevant e m p iric a l evid en ce em ergin g
fro m the literature.

21.3.1 Direct Effects at the Firm Level

F irm s are w h ere in n o va tio n s are in tro d u ced an d w h ere th ey sh o w th eir im m ed iate
direct effects on em p lo ym en t. A g ro w in g literatu re has e xp lo red the issue w ith a
v ariety o f m o d els, n ation al studies, an d panels o f firm s (fo r review s, see Petit 1995;
C h en n ells an d V an R eenen 1999; Spiezia an d V iv are lli 2002). E m p iric a l w o rk in this
field has gen erally u sed an n u al su rveys o f firm s in pan el data; h ow ever, panels are
u su ally n o t representative o f the w h o le m a n u fa ctu rin g in d u stry, an d in m o st cases
leave o u t services altogeth er; an d therefore it is d ifficu lt to generalize th eir co n clu ­
sions.
T h e evidence o n the overall em p lo ym en t im p act o f in n o v a tio n at the level o f firm s
tends to be p o sitive: firm s that in n o vate in p ro d u cts, an d also in processes, grow
faster and are m o re likely to exp an d th eir e m p lo ym en t th an n o n -in n o v a tiv e ones,
regardless o f in du stry, size, o r o th er ch aracteristics.3 T h e v a rie ty o f in n ovative
strategies, jo b creation, and d estru ctio n pattern s have been h ig h lig h ted in such
studies, together w ith the firm s’ ch aracteristics (stru ctu ral facto rs, flexibility, co m ­
petences, etc.) that ten d to be associated w ith better p erfo rm an ces.
A reverse relatio n sh ip can also be considered. In the lo n g -ru n d evelo p m en t o f
firm s, phases o f rap id grow th o f em p lo ym en t m ay be seen as d eterm in an ts o f
in n o va tio n s as firm s in n o vate in o rd er to cope w ith the rig id ity o f p ro d u ctio n
processes and in creasin g w ages, w hile try in g to cap tu re, th ro u g h greater p ro d u c tiv ­
ity and quality, the o p p o rtu n ities o f exp an d in g m ark ets.4
Som e studies also suggest that the p o sitive em p lo ym en t effects o f tech n ological
in n o va tio n are lin k ed to o rgan izatio n al changes. A stu d y o n a large an d rep resen ta­
tive sam p le o f French firm s fo u n d that firm s that ad o p ted ad van ced m an u factu rin g
system s in the p e rio d 19 8 8-9 3 an d in tro d u ced in p arallel o rg a n iz a tio n a l change had
greater em p lo ym en t grow th than others, regardless o f size o r sector, an d that this
p o sitive effect w as greater than in firm s that in tro d u ced o rg an izatio n al in n o vatio n
o n ly (G reen an 2003; see S ectio n 21.3.3 b elow ).
H ow ever, firm -level studies on the in n o v a tio n -e m p lo y m e n t lin k are unable to
p o in t out w hether the o u tp u t and jo b gains o f in n o va tin g firm s are achieved at the
expen se o f co m p etito rs, o r w h eth er there is a net effect o n aggregate in d u stry
e m p lo y m e n t.5 It is often d ifficu lt to generalize b eyo n d the g ro u p s o f firm s in vesti­
gated and to co m p are results across countries. W hen pan els are used, a large part o f
 Table 2 1 .2 E ffe c ts o f in n o v a tio n o n th e q u a n tity o f e m p lo ym e n ts Selected em p irical stu d ie s

Sp ey Countries Years Level of analysts Innovation data sources Results on employment

F irm -le v e l studies

M a e b in and W a d h w a n i UK 1984 Cross firm , m anuf. British w orkplace industrial Positive
,1 'ІІГ ■ relations survey
Irouwer, Kldnknecht, N etherlands 1983-8 Cross firm , m anuf. D utch survey N egative
and Reljnen 1993 Positive w ith product innovation
Ш§Ьіг, Ryan, and Van UK 1976-82 Panel o f firm s, m anuf. SPRU innovation database Positive more fle x ib ility
Keener* 1996 and patents
Van fteeneo 1997 UK 1976-82 Panel o f m anuf, firm s Survey on UK firm s Positive
Smolny 1998 G erm any 1980-92 Panel o f m an u f. firms Survey on G erm an firm s Positive
"Greahanand G uellec France 1986-90 Cross firm , m anuf. In n o vatio n survey Positive a t th e firm level
2000 Cross sector N egative a t th e industry level fo r

I NNOVATI ON AND EMPL OYMENT

process innovations

in d u s try -le v e l studies

teyer-Kramer 1 9 9 2 G erm any 1980s In p u t o u tp u t model ail Industry data N egative, d iffe re n tia te d by
econom y sector
ѴІѵагеШ, Evangelista, and Ita ly 1985 Cross sector Innovation survey N egative o f process innovation
.Plante1 9 9 6 ' 3 0 m anuf. industries Positive o f product innovation
Pianta 2000* 2001 SEU 1989-93 Cross sector Innovation survey O verall negative
countries 21 m anuf. industries Positive o f product innovation
A ntnnueci and Pianta 8 EU 1994-9 Cross sector Innovation survey Overall negative
■ ,2002/ countries 10 m anuf. industries Positive o f product innovation
Evangelista and Savona Ita ly 1993-5 Cross sector service Innovation survey Overall negative, d ifferen tiated
■2002t 2003 industries by service industries and size

ЧЛ
VI
VI
 Vl
Table 21.2 (cont.)

MARIO PIA N T A
S tu d y C o u n trie s Y ears Level o f an alysis In n o v a tio n d a ta sources Results o n employment

M a c r o e c o n o m ic -le v e l studies
Layard an d Nickel! 1 9 8 5 UK 1954-83 M a c ro model Labor p ro d u c tiv ity Neutral
Viva relit 1995 US an d Ita ly 1966-86 M a c ro m odel REtD lin ked to p ro d u c t and Differentiated by compensation
process in n o v a tio n s m e c h a n is m and country
S im o n e tti, Taylo r, an d US, Italy, 1965-93 M a c ro m o d e l REtD lin ked to p ro d u c t and Differentiated by compensation
Vivarelli 2 0 0 0 France, Jap an process in n o v a tio n s mechanism
S im o n e tti a n d T a n c io n i UK an d Ita ly 1970-98 M a c ro m o d e l q u a rte rly REtD lin ked to p ro d u c t an d Differentiated by compensation
2002 data process in n o v a tio n s mechanism

Simulation studies
Leontiefand Duchin 1986 US 1980-2000 In p u t o u tp u t m o d e l ail A s s u m p tio n s on Negative
econom y p e rfo rm a n c e
K a lm b a c h an d K urz 1 9 9 0 G e rm a n y 2000 In p u t o u tp u t m o d e l all A s s u m p tio n s Negative
econom y
IPT5~E5T0 2001 Europe 2000-2020 Gen, e q u il. m o d e l all Assumptions on Positive, differentiated by
econom y p ro d u c tiv ity g ro w th innovation policy
 IN N O V AT IO N AND EM PL OYM EN T 579

thejobs created or lost may be accounted for by the entry or exit of firms left outside
thepanel. In order to address these issues, we need to turn to industry-level studies

21.3.2 The Effects at the Industry Level

Analyses at the industry level address not only the direct employment effects of
innovation within firms, but also any indirect effects that operate within the indus­
try. These indirect effects include the competitive redistribution of output and jobs
from low to high innovation-intensive firms, and the evolution of demand (and
therefore output and jobs) resulting from the lower prices due to innovation, given
the price elasticities of the industry s goods. The industry level may be the most
satisfactory level of analysis, as it is able, on the one hand, to differentiate between the
variety of technological regimes and strategies and, on the other hand, to bring in the
demand dynamics of specific sectors, taking into account country differences in
economic structures.
Studies on industries (see Table 21.2) show that the employment impact of
innovation is positive in industries (both in manufacturing and services) character­
ized by high demand growth and an orientation toward product (or service)
innovation, while process innovation leads to job losses. The overall effect of
innovative effort depends on the countries and periods considered, but, in general,
the higher the demand growth, the greater the importance of innovative industries
(both in manufacturing and services) and the greater the orientation of firms within
an industry toward product innovation, the more positive are the employment
effects of innovation.
Demand factors are important because an industry’s demand is constrained by
the composition and dynamics of domestic and foreign demand. High demand
growth leaves room for a variety of firm strategies and for better employment
outcomes, while stagnant demand deepens the selection process among firms and
emphasizes the role o f technological competition (Pianta 2001).6
The empirical investigations of the employment impact of innovation at the
industry level include studies using R8cD or patenting as innovation proxies,7
input-output models,8 and more recent works based on innovation surveys (unfor­
tunately, available only for Europe). Set in the context of the European debate on
jobless growth” in the 1990s, the evidence points to an extensive process of
restructuring in many manufacturing sectors where the growth of value added is
not matched by increases in jobs (see Pianta, Evangelista, and Peram 1996; Fager-
berg, Guerrieri, and Verspagen 1999)*
Studies using innovation survey data9 show that, in Europe, employment change
(in most cases a decrease) is positively affected by the dynamics of demand and by
the relevance of product innovations, while a higher intensity of innovative
SBC MARIO PIANTA

e xp en d itu re p e r $e has a negative im p act o n jo b s, su ggestin g a p re va ilin g p attern o f

lab o r-rep lacin g tech n o lo g ical change.
S im ila r fin d in g s co m e fro m a m o d el in w h ich e m p lo y m e n t is affected b y d em an d
d yn am ics, lab o r costs, and in n o va tio n variab les o n p ro d u c t o r p ro cess in n o vatio n s;
in the co n text o f the m o d est aggregate g ro w th o f the 1990s, E u ro p ea n in d u stries (in
eigh t cou n tries) w ere d o m in a ted b y process in n o va tio n s, w ith g en erally negative
effects o n jo b s (A n to n u cci and P ian ta 2002).
T h e fin d in gs fo r service in d u stries d o n o t d iffe r su b stan tially fro m those o f
m a n u fa ctu rin g . Studies on Italy h ave fo u n d an o verall n egative effect, concen trated
a m o n g the largest firm s, o n lo w -sk illed w o rk ers, o n cap ital-in ten sive an d finance-
related sectors, an d w here the im p act o f IC T s has been m o st w id esp read . Sm aller
firm s an d tech n o lo g y -o rien ted activities show , on the o th er h an d , net em p lo ym en t
gain s (E van gelista an d S av o n a 2002, 2003; see also C h . 16 b y M iles in this v o lu m e ).
T h e evidence at the sectoral level, especially fo r E u ro p ea n co u n tries, suggests a less
o p tim istic v ie w on the e m p lo ym en t im p act o f in n o va tio n . T h e slo w aggregate
grow th o f the 1990s has con strain ed d em an d , lim itin g the p o ten tial em p lo ym en t
ben efits o f tech n o lo gical change. W h ile p ro d u c t in n o va tio n has h ad p o sitiv e effects
on o u tp u t an d job s, increased in tern atio n al co m p e titio n has p u sh ed firm s tow ard
restru ctu rin g, and process in n o va tio n s have d o m in a ted several in d u stries in m any
E u ro p ean cou n tries, lead in g to a prevalen ce o f la b o r-sav in g effects. T h is contrasts
w ith the exp erien ce o f the U S in the 1990s, m arked b y h igh d em an d an d em p lo ym en t
grow th .

21.3.3 Direct and Indirect Compensation Effects at

the Macroeconomic Level
T h e m o st com plete v ie w o f the e m p lo ym en t im p a ct o f in n o v a tio n is p ro v id e d by a
m a cro e co n o m ic perspective that can integrate all the in d irect effects th ro u gh w h ich
tech n o logical change affects em p lo ym en t . T h is is the a p p ro ach typ ical o f the debate
on “ co m p en satio n m ech an ism s” since the tim es o f R ica rd o an d M a rx , p ittin g those
w h o argu e that the eco n o m ic system has b u ilt-in m ech an ism s w h ich assure the
reco very o f jo b s lost due to in n o vatio n , again st critics w h o p o in t o u t the lim itatio n s
o f th eir effects and the p o ssib ility o f tech n o logical u n e m p lo y m en t. M o st o f this
debate conceptualizes tech n o logical change as the in tro d u c tio n o f n ew capital
go od s, typical o f n in eteen th -cen tu ry m ech an izatio n , th u s co n cep tu alizin g “ in n o v ­
a tio n ” as “ process in n o vatio n .” Secon d, the debate typ ically assu m es e q u ilib riu m in
p ro d u ct m arkets, i.e. no d em an d co n strain t is co n sid ered , fo llo w in g S a y ’s law.
Recent research in this d irection has relied on K a ld o rian ap p ro ach es, o n the w ork
by Pasinetti (1981) and B o yer (1988a an d b). A detailed treatm en t o f this issue is in
 __ I NNOVATI ON AND EMPL OYMENT 581

Vivarelli (i 995)> has sum m arized the compensation mechanisms and the way
they may (or may not) operate in the economy.
The compensation mechanism v ia d e c re a s e in p r ic e s is one o f the most important
ones.* new technologies m ay make lower prices possible, increasing international
competitiveness and output, offsetting job losses due to the original innovation
This outcome, however, is contingent on the lack o f demand constraints, on the
decision o f firms to transfer in lower prices the productivity gains due to the
innovation, and on the lack o f oligopolistic power in the relevant markets (Sylos
Labini 1969).
The compensation m echanism v ia n e w m a c h in e s may create jobs in the industries
in which the new means o f production are made, responding to the increased
demand for equipment by users. However, at the aggregate as well as at the firm
level, the rationale for mechanization is by definition saving on the overall use o f
labor, putting a limit on the relevance o f this mechanism.
The compensation m echanism v ia n e w in v e s tm e n t argues that the temporary
extra profits available to the innovator may be turned into new investment if profit
expectations are favorable (and assum ing that Say's law operates); this, however,
may expand production capacity and jobs, or m ay introduce additional labor saving
effects.
The compensation m echanism v ia d e c re a s e in w a g es is typical o f the neoclassical
view of the labor market . As technological unemployment appears, wages should fall
and firms should hire m ore workers. This mechanism, however, is based on strong
assumptions as to the feasibility o f any com bination o f labor and capital, competi­
tive markets, flexibility o f wages, and labor markets.
The compensation m echanism v ia in c re a s e in in c o m es operates in the opposite
way, through the increased dem and associated with the distribution o f part o f the
gains from innovation through higher wages, as has happened in large, oligopolistic
firms in mass production industries. However, any wage increases can hardly be
large enough to sustain additional aggregate demand.
Finally n e w p r o d u c t s m ay lead— as discussed above— to new economic activities
and new markets ( w e lfa r e effects) or, on the other hand, they may simply replace
existing goods ( s u b s titu tio n e ffe c t , see Katsoulacos 1986),
Aggregate empirical patterns for the US are examined in a descriptive way by
Baumol and W olff (1998). Considering five innovation indicators for the whole
economy and their link to the structure and changes in unemployment in the US in
the 1950-95 period, they conclude that faster innovation leads to a higher natural
rate of unem ploym ent" and to longer spells o f frictional unemployment.10 Layard
and Nickell (1985), on the other hand, have argued that the working o f various
compensation mechanism s ruled out the possibility o f technological unemploy­
ment in the UK.
Building on Boyer (1988a and b) and on the Regulationist approach, Vivarelli
(1995) has developed a sim ultaneous equations model for testing the compensation
5І2 MARIO PIANTA

mechanisms in the US an d Italy, He has fo u n d th at the m e ch a n ism v ia decrease in

prices is the m ost effective o n e, an d that the po sitive effects o f n ew p ro d u c ts an d lab o r
m arkets o p erate in the U S (w h ere n ew jo b s are created ), b u t n o t in Italy, w h ere net
job losses have been fo u n d . T h is a p p ro a ch h as been fu rth er d evelo p ed b y Sim o n etti,
Taylor, an d V iv are lli (20 0 0 ), w h o have co n sid ered fo u r co u n tries, an d b y S im o n etti
an d T an cio n i (20 0 2), w h o have d evelo p ed a m o d el fo r an o p e n e c o n o m y in the case
o f the U K an d Italy, A ll have fo u n d a d ifferen tiated im p a ct o f co m p e n sa tio n
m ech an ism s in d ifferen t n atio n a l econ om ies.
W h ile this a p p ro ach is the m o st co m p reh en sive an d sa tisfa c to ry fo r e xp la in in g the
o verall im p a ct o f tech n o lo gical ch an ge o n em p lo ym en t, the c o m p le x ity o f the
co n stru ctio n o f the m o d el, the p ro b lem s in sp ecify in g all relevan t relation sh ips,
an d the lack o f adequ ate data lim it the feasib ility o f this ap p ro ach . T h e overall
fin d in gs o f these studies p o in t to a d ifferen tiated im p act o f in n o v a tio n d ep en d in g o n
co u n tries5 m a cro e co n o m ic co n d itio n s an d in stitu tio n a l factors. T h e em p lo ym en t
im p acts o f in n o va tio n gen erally are m o re p o sitive in eco n o m ies in w h ich new -
p ro d u c t gen eratio n an d in vestm en t in n ew e co n o m ic activities are h igh er, and in
w h ich the d em a n d -in crea sin g effects o f price red u ctio n s are greater.

21.3.4 Simulation Studies

T h e e m p lo ym en t im p act o f in n o va tio n has also been stu d ied th ro u g h the use o f a
sim u latio n ap p ro ach . L e o n tie f an d D u ch in (1986) h ave estim ated th at the d iffu sio n
o f co m p u te r tech n o lo g y an d a u to m atio n in the U S e c o n o m y w o u ld h ave negative
em p lo ym en t effects u sin g an in p u t-o u tp u t m o d e l in c o rp o ra tin g stro n g a ssu m p ­
tio n s o n the p ro d u ctivity -e n h a n cin g effects o f p ro cess in n o va tio n , b u t no d em an d
d y n a m ic s.11 Likew ise, a stu d y on the im p act o f m icro electro n ics in fo rty sectors o f
the U K e co n o m y in the 1980s suggested that either net jo b gain s o r losses were
possible, d ep en d in g o n the assu m p tio n s on the speed o f d iffu sio n an d u sers’ dem an d
o f m icro electron ics (W h itley an d W ilson 1982). A d ifferen t a p p ro a ch — a general
e q u ilib riu m m o d e l w ith a sectoral stru ctu re, w h ich assu m es fu ll e m p lo y m e n t— has
been used fo r sim u latin g the e m p lo ym en t im p act o f d ifferen t scen ario s o f tech n o l­
o g y-b ased p ro d u c tiv ity grow th and o f the co m p o sitio n o f c o n su m p tio n , in a recent
stu d y b y IP T S -E S T O (2001) on the E u ro p ean U n io n . T h e results sh o w an overall
p o sitive im p act on jo b s, d ifferen tiated acco rd in g to the altern ative sectoral d istrib u ­
tio n s o f R & D an d in n o vatio n efforts; the best o u tco m es result fro m the co n cen ­
tration o f efforts in high tech n o lo g y indu stries. W h ile they are in terestin g as
e xp lo ratio n s o f alternative fu tures, the results o f such sim u la tio n s are w eaken ed b y
the m o d e ls’ in ab ility to iden tify either tech n o logical u n e m p lo y m en t (w hen general
e q u ilib riu m is used) o r m ost co m p en satio n effects (w h en in p u t-o u tp u t m o d els are
used), and on the arbi trarin ess o f the assu m p tio n s on the diffusion and produ ctivity
of new technologies.
Summarizing the results o f this section , b oth sectoral and aggregate studies
generally point o u t the p o ssib ility o f tech n o lo gical u n em ploym en t, w hich em erges
when industries o r co u n tries see the p revalen ce o f process in novations in contexts o f
weak demand, F irm s in n o v a tin g in b o th p ro d u cts an d processes m ay b e successful in
expanding o u tp u t an d jo b s regardless o f the eco n o m ic context, but often do so at the
expense o f n o n -in n o v a tin g firm s. T h e specificities o f industries, countries, and
macroeconomic co n d itio n s are cru cial d eterm in an ts o f the results obtained in
empirical studies.
All the analyses o f this section refer to national economies, sometimes even
assuming a closed economy. W hen we consider an open economy, the picture
becomes more complex, as, on the one hand, innovation may lead to competitive­
ness and exports, weakening the demand constraint while, on the other hand,
domestic demand may increase im ports when foreign competitors are more innova­
tive in terms o f price or quality. No empirical analysis has so far addressed the
innovation-employment question in a truly global dimension. In manufacturing
sectors that are highly internationalized, the introduction and diffusion o f innov­
ation leads to job creation and losses in a large number o f different countries, making
the distribution o f the benefits and costs o f technological change very complex, The
case of developing countries in this context is particularly interesting, as new
technologies are at the center o f the structural changes and dynamic learning
economies typical o f catching-up countries which nevertheless face greater difficul­
ties in capturing the em ploym ent benefits o f technological change (see Karaomer-
lioglu and Ansal 2000 and Ch. 19 by Fagerberg and Godinho in this volume).

2 1 .4 T h e E ffe c t s on th e Q u ality
of E m p lo ym en t

Most approaches that assum e equilibrium in labor markets (and therefore no

technological unem ploym ent, see Table 21.1) have disregarded the effects of innov­
ation on the quantity o f jobs, focusing instead on change in the skill mix o f
employment and on wage polarization. A large (mainly) US literature on skill-biased
technical change (reviewed in Acem oglu 2002: 7) argues that technical change is
biased towards skilled workers as it replaces unskilled labor and increases wage
inequality. In fact, a strong com plem entarity between technology and skills has
5§4 MARIO P I A N T A

ch aracterized m o st o f the tw entieth century, w h en in n o v a tio n h as p ro b a b ly alw ays

been skill biased , in co n trast to the u n skilled b ias typ ical o f the n in eteen th century,
w h en m ech an izatio n led to the d eskillin g o f artisan s (see B ra ve rm an 19 7 4 ),12
T h is section co n sid ers the stream s o f research that h ave fo u n d evid en ce o n the
co m p le m e n ta rity betw een IC T s and h igh skills, as w ell as th o se h ig h lig h tin g the
d eskillin g process associated w ith the greater co n tro l o ver p ro d u c tio n m ad e po ssib le
b y IC T s. A related fin d in g in m u ch o f this research em ph asizes the ten d e n cy fo r IC T
a d o p tio n to increase p o larizatio n in the w age stru ctu re as a resu lt o f ch an ges in the
skill stru ctu re o f em p lo ym en t. A n o th e r stream o f this research exam in es the effects
o f o rgan izatio n al in n o v a tio n — alon e o r co m b in ed w ith tech n o lo g ical in n o v a tio n —
o n skills, w ages, an d e m p lo ym en t (fo r review s see C h en n ells an d V an R een en 1999;
San ders an d ter W eel 2000). Table 21.3 su m m arizes the m o st relevan t literatu re.

21,4л Skill-biased Technical Change

M a n y studies— m ain ly o n U S firm s an d indu stries— argue that, in the last two
decades, a long-establish ed trend o f increase in the skill-in ten sity o f em plo ym en t
has been accelerated b y the in tro d u ction o f in fo rm atio n tech n o lo g y an d com puters.
T h e issue has generally been investigated u sing a factor su b stitu tio n fram ew ork,
sh ow in g that direct o r indirect m easures o f tech n o lo gy are im p o rtan t exp lan ato ry
factors fo r the relative increase o f skilled lab o r (see B erm an , B o u n d , and G rilich es 1994;
A utor, Katz, an d K ru eger 1998). O ne stream o f w o rk com pares the effects o f tech n ol­
o g y w ith those o f increased international trade, fin d in g that the fo rm er accoun ted for
m ost o f the fall in d em an d fo r less-skilled w orkers (see B erm an , B o u n d , and M ach in
1998). O ther studies have fo u n d that new technologies are ad o pted m o re extensively in
plants w ith m ore skilled w orkers, b u t do not increase the d em an d fo r skills (D orns,
D u n n e, and Trotske 1997. See the review in Chennells and V an R eenen 1999).
W h en m o re refin ed m easures o f skill are used, how ever, the evidence on skills bias
is less clear. M o v in g fro m the sim p le m easures o f blue o r w h ite co llar jo b s, o r o f years
o f sch o o lin g, to m o re refined in d icators o f skills, in clu d in g co g n itive (typ ical o f
tech nical staff), interactive (typ ical o f su p e rv iso ry staff), an d m o to r com petences
(typ ical o f m an u al w orkers) in U S in d u stries betw een 1970 an d 1985, H ow ell and
W olff' (1992) fo u n d that exp en d itu res on co m p u ters an d n ew in vestm en t were
associated w ith raisin g d em an d fo r h igh co gn itive skill w o rk ers, alth o u gh w ith
differences across o ccu p atio n s an d indu stries.
H o well (1996) rejects the idea o f a lin k betw een co m p u terizatio n , u p sk illin g, and
w age inequality, and finds that m a jo r shifts in skill stru ctu re o ccu rred betw een 1973
and 1983, w ith little change tak in g place afterw ard s, w hen the d iffu sio n o f IC Ts
accelerated and co m p u ter-related investm ent p er em plo yee in creased d ra m a tic ­
a l l y 13 In recent years, the em p lo ym en t shares o f h ig h -sk illed blue collars and lo w
І Ш ё 1 1 ,3 ^ f e c t s o f m flo y a tio n o n the q u a lity o f e m p lo ym e n t: skills, w ages, an d o rg a n iz a tio n s: Selected em pirical stu d ie s
Study
S ttl tosed technical change
F irm -level studies
Machin 1996
Doms, Dunne, and Troske
1997
Bresnahan, Brynjolfsson,
and Hitt 2002
Industry-level studies
Howell and Wolff 1992
Berman, Bound, and
Griliehes 1994
Wolff 1996
Autor, Katz, and Krueger
1998
Machin and Van Reenen
1998
Countries Years Level of analysis Innovation data sources Results
UK 1984-90 Cross firm, manuf. cross British workpl. ind. rel. surv. Positive effect on high skill jobs
sectors R EfcD, innov., computer use Negative on least skilled ones
US 1988, 1993 Panel and cross firm Use of 5 manuf. Higher skill where technologies
technologies are used; in panel: only for
computers
Higher wages where technologies
are used; in panel no effects
US 1987-94 Panel and cross firm IT stock and use Positive effect on skill demand
combined with organ, change
INNOVATION AND EMPLOYMENT
US 1970-85 Cross sector all private Exp. on computers, new Positive on cognitive skill jobs
economy investment Negative on interactive and
motor skills
US 1979-89 Cross sector 4 digit R6tD and computer invest Higher skill where technology is
manuf. industr. higher
us 1970-90 Cross sector all private Exp. on computers, REtD Positive effect on complexity of
economy tasks
Negative on motor skills
US 1960-95 Cross sector all Computer use, REtD, TFP Faster upskUltng in high tech
economy industries, and after 1970
7 OECD 1970-89 Cross sector 2 digit REtD intensity Higher skill where REtD is higher
countr. manuf. industr. In all countries
 ЧЛ
Table 2 1 .3 (cont.) 00
o

M A R I O PI A N T A
S tu d y C o u n t r ie s Y e a rs L e v e l o f a n a ly s is I n n o v a t io n d a t a s o u r c e s R esults

T e c h n o lo g y a n d w a g e d is p e r s io n

F ir m - le v e l s tu d ie s
C asavo la, G avosto , a n d I t a ly 1986-90 C ro s s f i r m s °/o o f in t a n g i b l e c a p it a l P o s it iv e e f f e c t o n s k ills
Sestito 1998 IM P S w a g e d a t a L im it e d w a g e d is p e r s io n

Van Reenen, 1996 UK 1976-82 P a n e l a n d c ro s s fir m In n o v , c o u n t s a n d p a t e n t s P o s it iv e e f f e c t s o n w a g e s

Black and Lynch 2000 US 1993-96 P a n e l a n d c ro s s fir m C o m p u te r use, o rg a n , H ig h e r w a g e s a n d p ro d u c tiv . w it h

m a n u f, changes te c h , u s e a n d o rg a n , c h a n g e
i n v o lv in g e m p lo y e e s

i n d u s t r y l e v e l s t u d ie s

Bartel and Lichtenberg US I960, 1970 C ro s s s e c t o r m a n u f . C o m p u te r in v e s t, R & D W a g e p r e m ia w i t h n e w e r

1991 1980 te c h n o lo g ie s
Bartel a n d S ic h e rm a n US 1979-1993 C ro s s s e c t o r m a n u f . C o m p u te r in v e s t , TFP, e tc W a g e p re m ia in h ig h te c h
1999 N a t’L Long. S u rv e y o f Y o u t h i n d u s t r ie s

O r g a n iz a tio n a l in n o v a t io n

F i r m - l e v e l s t u d ie s

C a r o li a n d V a n R e e n e n UK a n d 1984-90 P a n e l a n d c ro s s fir m °/o o f w o r k e r s a f f e c t e d P o sitive e ffe c ts on skills in th e U K

2001 F ra n c e 1992-96 C o m p u te r u se W e a k in F r a n c e
O r g a n , in n o v . h a s s t r o n g e r
n e g a tiv e e ffe c ts o n lo w skill
Piva and Vivarelli 20 02 Ita ly 1991-97 P a n e l a n d c ro s s fir m R & D , o rg a n , c h a n g e P o sitive e ffe c ts o f o rg a n , in n . o n
s k ills , n o e f f e c t o f t e c h n o l o g y

G re e n a n 2 0 0 3 F ra n c e 1988-93 P a n e l a n d c ro s s fir m T ech , in v es t, o rg a n , c h a n g e P o s it iv e e ffe c ts o f o r g a n , in n . o n

s k ills , le s s e f f e c t o f t e c h n o l o g y
 I NNOVATI ON AND EMPL OYME NT 587

skilled white collars have declined more rapidly, and little change in the skill
structure o f services em ploym ent is apparent. ICT investment since the mid-1980s
therefore appears to have negatively affected the low skilled white collars (mainly
female) much more than the lowest blue collar skills.

214.2 Wage Polarization and the Labor Market

Although many studies do not separate the effects o f technological change on skills
and wages, some have investigated the evolution o f wage differentials in three
contexts: across industries with varying R & D and capital intensities; among workers
and firms with different levels o f use o f computer technology and ICTs; and among
workers or social groups with different educational levels (see Sanders and ter Weel
2000; Acemoglu 2002). These studies find weak evidence o f polarization, although
the relationship o f polarization to technological change is confounded by the
tendency for computers to be used by more competent workers who already earn
higher wages (Chennells and Van Reenen 1999)* On the other hand, the technology-
wage polarization link has been questioned by studies pointing out the lack o f an
acceleration in these effects in recent years, and suggesting instead that sectoral shifts
in employment and grow ing international trade have exercised more influence on
the structure o f wages (Addison and Teixeira 2001).
In comparing US and European results, a paradox emerges. Empirical patterns
show that, in the last two decades, low skill and low wage jobs have grown more
rapidly in the US than in Europe, a trend associated with the faster growth o f the
supply o f labor, and greater polarization in the wage structure. Skill-biased technical
change thus may be m ore a European than an American phenomenon. In this view, a
stagnant labor supply and m ore educated workers, as well as slower aggregate growth
and greater competitive pressure, have led European firms and industries to adopt
technological and organizational innovations during the past two decades that have
favored relatively skilled workers. The effects o f this skills bias in European employ­
ment on wage polarization have been attenuated by European labor market insti­
tutions (e.g. stronger labor unions). But in the United States, faster growth o f new
jobs at the top and bottom end o f the skill structure, combined with weaker labor
unions, have produced considerable polarization in wages,14

214*3 The Effects of Organizational Innovation on

Employment and Skills
In an extensive su rvey o f o rg a n iz a tio n a l change in U S m an u factu rin g and service
fitms, A ppelbaum an d B att (1994) fin d w id esp read ad o p tio n o f new m anagem ent
588 M A R I O PI А К Т А

p ractices in c o n ju n ctio n w ith the in tro d u c tio n o f n ew tech n o lo gies, b u t n o in fo r­

m a tio n is p ro v id e d o n a n y ch an ge in em p lo y m e n t levels. B o th te c h n o lo g y ad o p tio n
(such as the n u m b e r o f n o n -m a n a g ers u sin g co m p u ters) an d n ew w o rk p la ce p ra c ­
tices are closely associated w ith h igh er p ro d u c tiv ity an d w ages (B lack an d Lynch
20 0 0 ). Im p ro ved p erfo rm an ces an d u p sk illin g o f the w o rk fo rce are also fo u n d in U S
firm s w h en IC T in tro d u ctio n is m atched b y n ew o rg a n iz a tio n s d elegatin g au th o rity
to w o rk ers an d team s (B ry n jo lfsso n an d H itt 2000; B resn ah an et al. 2002).
Several E u ro p ea n studies (C a ro li an d V an R een en 2 0 0 1 o n F ran ce an d B ritain ;
G reen an 2003 on France; P iva an d V iv are lli 2002 o n Italy) h ave sh o w n that o rg an iza­
tio n al in n o va tio n is m o re im p o rta n t th an tech n o lo g ical in n o v a tio n in sh aping
chan ges in o ccu p a tio n a l stru ctu re an d skills. T h is is gen erally n o t associated w ith
an increase in the n u m b er o f em ployees, w ith the excep tio n o f m an ag em en t o ccu ­
pation s.
O rgan izatio n al an d tech n o lo gical changes in services, o n the o th er h an d , have
reflected the o p p o rtu n ities offered b y IC T to o verco m e tim e an d space co n strain ts in
the p ro v isio n o f services, lead in g to m a jo r flow s o f jo b creatio n an d d estru ctio n , and
to rap id ly ch an gin g skill requ irem en ts. A v a rie ty o f strategies o f restru ctu rin g,
em ergence o f n etw orks, su b co n tractin g, an d o u tso u rc in g h as resu lted, lead in g to
p o larizatio n effects in skills and w ages (Petit and Soete 2001b ; F rey 1997).
T h e rath er fragm en ted evidence so far available o n o rg a n iz a tio n a l in n o vatio n
suggests that it plays a cru cial role alon gsid e tech n o lo g ical in n o v a tio n in sh aping
p ro d u c tiv ity an d e m p lo ym en t o u tco m es. T h e tw o can h ave a co m p lem en tary
relatio n sh ip (especially w h en a v irtu o u s circle o f gro w th is in p lace) lead in g to a
co m b in ed effect on p erfo rm an ce an d u p sk illin g that can be greater th an their sum .
O n the oth er h an d, changes in o rg an izatio n s o r in tech n o lo gies m a y be pu rsu ed as
altern ative road s in contexts o f restru ctu rin g an d jo b losses.

21.4.4 A Broader View

A lth o u g h there is little d o u b t that, in the lo n g term , tech n o lo g ica l change is
asso ciated w ith im p ro vem en ts in the skills an d w ages o f so m e w o rk ers, the specific
effects o f in n o va tio n o n e m p lo ym en t in p a rticu la r co u n tries an d p e rio d s reflect the
o p e ra tio n o f m an y other factors, in clu d in g e co n o m ic stru ctu res, the strategies o f
firm s, the o p e ra tio n o f lab o r m arkets rules an d in stitu tio n s, an d n atio n a l econ o m ic
policies. T h e m a c ro e co n o m ic d im en sio n is m issin g fro m m u c h o f the current
literatu re on tech n ological in n o va tio n and em p lo ym en t. Little research has
add ressed the b ro ad er d istrib u tio n effects o f tech n o lo g ical ch an ge a m o n g profits,
rents, w ages, sh orter w o rk h ou rs, and lo w er prices.
T h is literature also has restricted its fo cu s to the d em an d fo r labor, despite
ab u n d an t eviden ce that new tech nologies interact w ith b ro a d e r so cial changes that
 I N N O V A TI ON and e m p l o y m e n t 589

affect the supply o f labor (see Acem oglu 2002). Slower population growth, higher
immigration flows, an ageing population and the forms o f women s presence in the
labor force all influence labor supply. In addition, the quality o f labor suppiy~~and
the potential for innovation— are affected by education and training, learning
processes, and the accum ulation o f competences by workers and firms.

21.5 C o n c l u s i o n : S t y l i z e d F a c t s and
Di r e c t i o n s for F u t u r e Re s ea rc h

This chapter has examined the relationship between innovation and employment,
reviewing a large em pirical literature. The com plexity o f the issues is such that no
single approach can account for the direct and indirect consequences o f techno­
logical change, or for its effects on the quantity and quality o f labor employed in the
economy.15 Theories and em pirical research have to proceed in parallel, with a close
interaction between concepts and measures, hypotheses and tests, building links to
related research areas. This final section summarizes some key stylized facts
emerging from the em pirical evidence and highlights possible directions for future
research.

21.5.1 A Few Stylized Facts

The never-ending race o f innovation and em ploym ent The evolution o f (most)
economies shows that, when growth, structural change, and demand dynamics
take place together, in the long run the jobs lost to technological change are found
elsewhere in the economy. I f no innovation took place, economic activities facing
competitive pressure would cut costs, wages, and eventually, jobs. The key question
is flie rate at which technological innovation and diffusion eliminate jobs, versus the
pace at which new econom ic activities create new jobs.
Technological unemployment can happen. On the basis o f the available evidence,
current technological change m ay be a cause o f unemployment. There is no auto­
matic mechanism ensuring that national or regional economies fully compensate for
innovation-related job losses. Indeed, Europe in the 1990s appears to have been a
region in which technological innovation and diffusion, in combination with a long
list of supporting influences, eroded employment.
590 MARIO PI A N T A

The type o f innovation is important T h e evid en ce sh ow s that it is essential to

d iscrim in a te betw een product in n o v a tio n (n o vel o r im itativ e) that has a gen erally
p o sitiv e em p lo y m e n t im p act, an d process in n o v a tio n (a d o p tio n a n d use o f new
tech n ologies) u su a lly w ith negative effects; these fin d in g s em erge regardless o f the
theoretical a p p ro a ch used.
Organizational innovation is closely linked to technological change. O rg an izatio n al
in n o va tio n is freq u en tly an in d isp en sab le co m p lem en t to the a d o p tio n o f n ew
tech n ologies, especially IC T (see C h . 5 b y L a m in th is v o lu m e ). T h ese co m p le­
m e n ta ry o rgan izatio n al chan ges critically affect the p ro d u c tiv ity an d em p lo ym en t
con sequen ces o f tech n o lo gical in n o v a tio n an d d iffu sio n , esp ecially in IC T .
One important consequence of technological innovation and adoption is change in
the “skill bias'" o f employment U n skilled jo b s have lo n g b een d eclin in g in absolute
term s (in E u rop e) o r g ro w in g slo w ly (in the U S ), w h ile jo b s fo r m o re educated
w o rk ers have been created at a faster pace in m o st co u n tries an d are associated w ith
greater in n o vative efforts. N evertheless, co n sid erab le u n ce rta in ty rem ain s o ver the
sp ecific effects o f IC T -related in n o va tio n since the 1980s o n the skill req u irem en ts o f
n ew jo b s, as o p p o se d to co in cid en tal chan ges in the su p p ly o f m o re ed u cated labor.
Wage polarization has been significant although the specific effects of new technologies
on this trend are difficult to identify. Since the 1980s, m ost countries have experienced a
grow in g divid e w ithin w ages (and m ore generally in incom es). T h is is the result o f
changes in econ om ic structures, in firm s’ strategies, an d in govern m en t policies.
Technological change has alm ost certainly co n trib u ted to this trend, b u t the specific
link betw een in n ovation and w age differentials is unclear. Trends in w age dispersion
and inequality reflect a n u m ber o f broader, coincidental trends in the evolution o f
lab or m arkets, em ploym en t form s, social relations, and n ation al policies.
Aggregate demand and macroeconomic conditions are important A lth o u g h the
role o f d em an d has gen erally been d o w n p layed in the in n o v a tio n literature, it m ay
help exp lain the negative em p lo ym en t im p act o f in n o v a tio n fo u n d in E u ro p e in the
lo w -gro w th decade o f the 1990s (see C h . 18 b y V erspagen in this vo lu m e ).
Innovation interacts with trade. In o p en eco n o m ies, trad e is an im p o rtan t
facto r— alon gsid e in n o vatio n — affectin g em p lo y m e n t an d w ages. T h e evidence
available (m a in ly fo r the U S) suggests that tech n o lo g ical ch an ge has affected jo b
losses, p a rtic u la rly a m o n g un skilled w o rk ers, an d lo w er w ages m o re su bstantially
than has increased in tern atio n al trade. B u t the in teractio n s b etw een in n o va tio n and
trad e sh ape co m petitiven ess (see C h . 20 b y C an tw ell in this v o lu m e ), the d irection o f
tech n ological change, the evo lvin g d iv isio n o f labor, an d the e m p lo y m e n t outcom es.
The national innovation system is a critical mediating influence in the effects of
technological innovation and diffusion on employment A c o u n try ’s tech n ological
o p p o rtu n ities and ab ility to d evelop lo n g -term learn in g an d in n o v a tin g capabilities
are roo ted in the nature and ch aracteristics o f its n atio n al in n o v a tio n system . Its
strengths, o rien tatio n , and p rio rities are lik ely to be reflected in the em p lo ym en t
effects o f a c o u n try ’s in n o vative efforts (see C h . 7 b y E d q u ist in this v o lu m e ).
 ____________________________________________ I^ J N O V ^ T I O N A N D E M P L O Y M E N T 591

Labor m a rk e t c o n d itio n s a n d in s titu tio n s m a tter. The employment outcomes o f

technological change depend on the operation o f job markets, the operation o f
institutions that set wages, and the institutions influencing worker learning, flex­
ibility, and welfare protection. Labor market institutions also influence the supply o f
labor, which can affect the m atch between labor supply and the skill and competence
requirements, as well as the employment effects, o f new technologies.

21,5.2 Directions for Future Research

The following issues emerge from this chapter as promising directions for future
studies.
Both theories and em pirical studies should strive to develop explanations and
predictions that are consistent at the various levels o f analysis o f the innovation-
employment relationship— firm s, industries, and the aggregate economy. All too
often, the findings at one level o f analysis contradict the patterns at a more aggre­
gated level. Achieving this goal requires that firm-level data be representative o f
whole sectors, and further requires that industry coverage includes much if not all o f
the economy (including services). Innovation survey data, when accessible, could
make this effort possible.
Studies on fir m s should address the innovation-em ploym ent question in the
context o f the evolution o f firms and market structures through the processes o f
diversity generation (through innovation) and selection (in the marketplace, Nelson
and Winter 1982). This avenue for research recognizes that the potential for job
creation depends on m echanism s o f firm growth and new firm generation.
An important challenge to innovation studies is to extend empirical research to
cover the s e rv ic e secto r ; rather than confining such work to manufacturing indus­
tries. From a theoretical perspective, an analysis linking structural and technological
(and organizational) change is especially important for understanding the expan­
sion of services in advanced economies. The increasing availability o f industry level
data also for services (such as in European innovation surveys) makes this extension
possible. Since m ost new jobs are created in services, this extension also is essential
for policy.
The need to devote greater attention to o rg a n iz a tio n a l in n o v a tio n s and to link
them to technological ones has already been pointed out; this appears as a major
direction for future studies.
Bringing the analysis o f the innovation—employment link into the m a cro ec o n o m ic
context o f a n o p en e c o n o m y is also im portant, in order to highlight the interactions
between technological developm ents on the supply side, and growth potential on the
demand side, as well as the interactions between innovation and trade in shaping
gmwth and employment outcomes.
592 MARIO P I A N T A

As international production, especially in manufacturing industries, becomes a

key aspect of industrial structures, an additional level of analysis, looking at individ­
ual industries at the global level, may emerge. Data constraints are serious, but
substantial evidence is now available on the international production networks of
multinational corporations, and the link to employment variables across the rele­
vant countries may become possible.
The appropriate labor m arket arrangements for favoring a virtuous circle between
innovation and a high quantity and quality of jobs and wages, through greater
learning, competence building and improved working conditions, is an additional
area for future research.
Finally, as already suggested, research could address the distributional effects of
innovation throughout the economy. So far, innovation has mainly benefited firms
and consumers, in the form of higher profits and lower prices. When we consider
also its effect on wage polarization, the result has been an increasingly uneven
distribution of incomes, an issue with major policy relevance.

N o t e s

1. Fo r their com m en ts and suggestions, I w ish to th an k the editors, C h arles E dquist, Bengt-
о
A ke Lu n d vall, M arco V ivarelli, and several con trib u to rs to this vo lu m e,
2. T he em ergence o f n ew fo rm s o f o rganization , the o p en in g o f n ew m arkets and o f new
sources o f m aterials w ere other types o f in n o vatio n con sidered by Schum peter.
3. G en erally po sitive effects o f in dicators o f in n o vatio n o n the n u m b er o f jo b s (once the
characteristics o f firm s have been controlled for) have been fo u n d in the studies o f U K
firm s by Van Reenen (1997), w ho related a large panel o f m an u factu rin g firm s to the SP R U
database o f B ritish in n ovation s, and b y M ach in an d W adhw ani (1991), u sin g the British
w orkplace in dustrial relations su rvey on the ad o p tio n o f ICTs. In a pan el o f G erm an firm s,
Sm o ln y (1998) has fo u n d a positive effect o f p ro d u ct in n o vatio n and no effect o f process
in n ovation s. D utch firm s investigated b y B rouw er, K leinknecht and R eijnen (1993)
show ed an overall negative link betw een in n o vatio n and job s, but w here p ro d u ct in n o v ­
ations were d o m in an t, better em ploym ent outcom es have been fo u n d .
4. T his view is pro po sed b y A n ton elli (2 0 0 1:17 3 ) in a stu dy on the Italian car m aker Fiat over
the tim e p erio d 19 0 0 -19 7 0 , w hich show s that em ploym en t grow th G ran ger-cau sed patent
grow th w ith different tim e lags. In turn, patent grow th led to p ro d u ctivity grow th,
5. In France, G reen an and G uellec (2000) fo u n d a po sitive relation ship betw een both
p rod u ct and process in n o vatio n and em ploym ent at the firm level, but at the in dustry
level, o n ly the fo rm er played a positive role. U sing data fro m the French h o u seh old survey,
Entorf, G ollac, and K ram arz (1999) fo u n d that co m p u ter use reduces the risk o f u n ­
em ploym ent in the short term , but not in the lon g term .
6. H igh d em an d is a necessary, but not sufficient, con d itio n fo r em p loym en t grow th. In
order to expand jobs, dem and and outpu t have to gro w faster than prod u ctivity, a
con d ition that has been often fou nd in the U S, but not so m u ch in Europe.
__ _________________________________________ 1 N N O V A T I O N a n d E M P L O Y M E N T 593

7 - Typical of the former approach is the OECD Jobs Study (OECD i994), which invest!
gated the high and persisting unemployment in advanced countries and downplayed the
role of technological change. The report emphasized the positive role o f new technolo­
gies associated to structural change and showed employment decline in low-technolaav
sectors and also in some R&D intensive industries.
8. Meyer Kramer (1992) has used a model of the whole German economy to assess the
impact o f ICTs, proxied by direct and indirect R&D in 51 sectors. The findings suggest a
generally negative em ploym ent impact, with some positive effects in higher technology
industries.
9. An early use o f these data for Italian manufacturing industry is in Vivarelli, Evangelista,
and Pianta (1996), who found a generally negative employment impact of technological
change and the expected contrasting consequences o f product and process innovations.
Pianta (2000,2001) investigated five European countries (Denmark, Germany, Italy, the
Netherlands, and Norway) in 1989-93 across twenty-one manufacturing industries,
finding a positive employm ent effect o f changes in demand and product innovations,
and an overall negative impact o f innovation intensity.
10. They argue that “ The evidence supports the conclusion that an increase in the pace of
innovation (all else equal) will raise both the natural rate o f unemployment and the
average length o f time during which an unemployed worker is 'between jo b s'" (p.10).
11. Their assumptions on productivity growth were based on the improvements in the
engineering performance o f robots. The long time-lags in their adoption and their poor
initial performance have meant that their actual economic effect on productivity has
been much lower.
12. Petit (1995) suggests that techniques and human labor had a complementary relation­
ship in pre-industrial times, following strong social norms. Industrialization was char­
acterized by the substitution o f machines for labor on the basis o f profit seeking; and the
post-war rapid growth was associated with technologies designed to overcome labor
shortages and increase productivity, while the current emergence o f a new technol ogical
system based on ICTs is marked by uncertainties.
13. Computer-related investment per employee rose from 150 $US in 1982 to about 1000 $US
in 1992 for both m anufacturing and services (Howell 1996: 292). The share o f low-skilled
blue collar US workers in manufacturing declined from 45*1 per cent in 1978 to 39*7 per
cent in 1982 and has since then remained stable, ending in 1990 with a 41 per cent share. In
the late 1980s, when ICTs became important, the most serious reduction in the employ­
ment shares concerned the high-skilled blue collars (which since 1978, had a stable share
around 21.7 per cent, a share which declined after 1985 to 19.7 per cent in 1990) and the
low-skilled white collars (a share which stayed stable at around 12 per cent until 1985 and
then declined to 10.6 per cent in 1990). At the opposite end, the share o f high-skilled
white collars increased in the 1978—83 period (from 19.5 to 24.1 per cent), then was stable
until a rise in 1989-1990 to 25.8 per cent (Howell 1996: 299, tables 1 and 2).
H* This may amount to a shift in “ wage norm s": “ in the face o f mounting competition,
employers reduced unit labour costs and increased flexibility in the production process
by following the Tow road'— lower wages, little training, and fewer permanent employ
ees" (Howell 1996:301), Com bined with a large use o f part-time and temporary workers,
anti-union practices, relocation to low-wage sites and inflows o f low-wage foreign
workers, these developments have led to an increase in the supply o f labor competing
594 MARIO PIANTA

for low-skill jobs, leading to a major fall in their wages, accompanied by a 25 per cent
reduction in the real minimum wage over the 1980s (ibid. 292).
15. There is also a “ national specificity” to most o f the different approaches. The “ skills”
research is largely US-focused, while most industry-level studies focus on Europe. There
is little English-language empirical work on the em ploym ent-innovation relationship in
Japan or other Asian high-growth economies.

R efe r e n c e s*

* A c e m o g l u , D. ( 2 0 0 2 ), “ Technical Change, Inequality and the Labor Market,” Journal of

Economic Literature, 40(1): 7-72.
* A d d i s o n , J., and T e i x e i r a , P. (2001), “ Technology, Employment and Wages,” Labour 15(2):
191-219,
A d l e r , P. (ed.) (1992), Technology and the Future o f Work, New York: Oxford University
Press.
A n t o n e l l i , C. (2001), The Microeconomics o f Technological Systems, Oxford: Oxford Univer­
sity Press.
A n t o n u c c i , T., and P i a n t a , M. (2002), “ The Employment Effects o f Product and Process
Innovations in Europe,” International Review o f Applied Economics 16(3): 295-308.
A p p e l b a u m , E., and B a t e , R. (1994), The New American Workplace, Ithaca: ILR Press.
A u t o r , D ., K a t z , L., and K r u e g e r , A . (19 9 8 ), “ Com puting Inequality: Have Computers
Changed the Labor Market?” Quarterly Journal o f Economics 1 1 3 : 1 1 6 9 - 2 1 4 .
B a r t e l , A. R, and L i c h t e n b e r g , F. R. (1991), “ The Age o f Technology and its Impact on
Employee Wages,” Economics o f Innovation and New Technology 1: 215-31.
------ and S i c h e r m a n , N. (1999), “ Technological Change and Wages: An Interindustry
Analysis,” Journal o f Political Economy 107: 285-325.
B a u m o l , W ., and W o l f f , E. (19 9 8 ), Side Effects o f Progress: How Technological Change
Increases the Duration o f Unemployment, Jerome Levy Economics Institute o f Bard College,
Public Policy Brief 4 1.
^ B e r m a n E., B o u n d , J., and G r i l i c h e s , Z. (1994), “ Changes in the Demand for Skilled Labor
Within US Manufacturing Industries: Evidence from the Annual Survey o f Manufactures,”
Quarterly Journal o f Economics 109: 367-98.
------------- and M a c h i n , S. (1998), “ Implications o f Skill Biased Technological Change:
International Evidence,” Quarterly Journal o f Economics 113:1245-79.
B l a c k , S., and L y n c h , L. (2000), “ What’s Driving the New Economy: The Benefits o f
Workplace Innovation,” N BER working paper 7479.
B o y e r , R. (1988a), “ Technical Change and the Theory o f Regulation,” in Dosi et a l 1988:
67-94.
—— (1988b), “ Formalizing Growth Regimes,” in Dosi et al. 1988: 608-35.
B r a v e r m a n , H. (1974) >Labour and Monopoly Capital, New York: Monthly Review Press.

* Asterisked items are suggestions for further reading.

___________________ i n n o v a t i o n a n d e m p l o y m e n t 595

Bresnahan , T. F., B rynjolfsson , E., and H itt , L. M. (2002), "Information Technology

Workplace Organization and the Demand for Skilled Labor: Firm-Level Evidence;1 Q uar­
terly Journal o f Econom ics 117: 339-76.
Brouwer, E., K leinknecht , A., and R eijn en , J. O. N. (1993), "Employment Growth and
Innovation at the Firm Level: An Empirical Study” Jo u rn al o f Evolutionary Economics v.
153~9 *
Brynjolfsson, E., and H itt , L. M. (2000), "Beyond Computation: Information Technol­
ogy, Organisational Transformation and Business Performance ” Jou rn al o f Economic
Perspectives 14: 23-48.
Caroli, E. (2001), "N ew Technologies, Organizational Change and the Skill Bias: What do we
Know?” in Petit and Soete 2001: 259-92.
---- and Van R eenen , J. (2001), "Skill Biased Organizational Change? Evidence From a
Panel of British and French Establishments,” Q uarterly Jou rn al o f Economics 116(4):
1149-92.
Casavola, R, Gavosto , A., and S estito , P. (1996), “ Technical Progress and Wage Disper­
sion in Italy: Evidence from Firm s5Data,” A nn ales d'Econom ie et de Statistique (Jan-June):
387-412.
Chennells, L., and V an R eenen , J. (1999), "H as Technology Hurt Less Skilled Workers? An
Econometric Survey o f the Effects o f Technical Change on the Structure o f Pay and Jobs ”
London: Institute for Fiscal Studies working paper 27,
Cyert, R. M., and M owery , D. C (eds) (1988), "The Impact o f Technological Change on
Employment and Econom ic Growth,” Cambridge, Mass,: Ballinger.
Doms, M., Dunne , T„ and T rotske , K. (1997), "Workers, Wages, and Technology,” Quarterly
Journal o f Econom ics 112: 253-89.
Dosi, G., Freeman , C , N elson , R., S ilverberg , G., and Soete, L. (eds), (1988), Technical
Change and Econom ic T heory ; London: Pinter.
Edquist, C., H ommen , L., and M c K elvey , M. (2001), Innovation a n d Em ploym ent: Product
versus Process Innovation, Cheltenham: Edward Elgar.
Entorf, H,, G ollac, M ., and K ram arz , F. (1999), "N ew Technologies, Wages and Worker
Selection,” Jo u rn a l o f L a b o r Econom ics 17(3): 464-91,
European C ommission -E urostat (2001), Statistics on Innovation in Europe . Data
1996-1997, Luxembourg: European Commission.
European Foundation for the I mprovement of Living and Working C onditions
(2001), Third E uropean S u rve y on W orking C onditions 2000. Dublin: European Foundation
for the Improvement o f Living and W orking Conditions.
Evangelista, R., and Savona , M. (2002), "T h e Impact o f Innovation on Employment and
Skill in Services. Evidence from Italy,” In tern a tio n al R ev iew o f A p p lied Economics.
------(2003), “ Innovation, Em ploym ent and Skills in Services. Firm and Sectoral
Evidence,” Structural C h an g e a n d E conom ic D yn am ics 14: 449-74.
Eagerberg, J., G u errieri , P., and V erspagen , B. (eds) (i 9 9 9 )> The Econom ic Challenge fo r
Europe: A dap tin g to In n ovatio n B a sed G ro w th , Northampton: Elgar.
F reeman, C , and LoucA, E (2001), A s T im e Goes By. From the Industrial Revolution to the
Information R evolu tio n , Oxford: O xford University Press.
-and Soete, L. (1994), Work fo r A ll or M ass U nem ploym ent? London: Pinter
— *(eds.) (1987), Technical C h an ge a n d F u ll E m ploym ent >Oxford: Basil Blackwell.
----- Clark, J., and S o e t e , L. (1982), U n em ploym en t a n d Technical Innovation, London.
Pinter,
596 MARIO PIAN TA

F r e y , L. (1997), “ II lavoro nei servizi verso il secolo XXI,” Quaderni diEconom ia delLavoro 57,
M ilan : A ngeli.
G r e b n a n , N . (2003), “ O rgan isation al C hange, Technology, E m p lo ym en t an d Skills: A n
E m p irical Stu d y o f French Manufacturing,” C a m b r id g e J o u r n a l o f E c o n o m ic s 27: 287-316.
-------an d G u e l l e c , D. (2000), “ Technological In n o vatio n an d E m p lo ym en t R eallocation,”
L a b o u r 14 (4 ): 54 7-90.
* G r e e n a n , N ., UH o r t y , and M a ir e s s e , J. (eds.) (2002), P r o d u c tiv ity , In e q u a lit y a n d the
D ig it a l E c o n o m y ; C am b rid ge: M ass.: M IT Press.
H a n s e n , A. H, (1964), B u s in es s C ycles a n d N a t io n a l In c o m e , L o n d o n : A llen and U nw in
(expanded edn.).
H e e r t ie , A. (1973), E c o n o m ic s a n d T ech n ica l C h a n g e >L o n d o n : W eidenfeld an d N icolson .
H o w ell , D. (1996), “ In fo rm atio n Technology, Skill M ism atch an d the W age Collapse:
A Perspective o n the U S Experience,” in O E C D 1996: 29 1-30 6 .
-------and W o lff , E. (1992), “ Technical C h an ge and the D em an d fo r Skills b y U S Industries,”
C a m b r id g e J o u r n a l o f E c o n o m ic s 16 :12 8 -4 6 .
IP T S -E S T O (Institute fo r Prospective T echn ology Studies, E u ro p ean Science an d Technology
O b servatory) (20 0 1), Im p a c t o f T ech n o lo g ica l a n d S t r u c tu ra l C h a n g e o n E m p lo y m e n t .
P ro s p e c tiv e A n a ly s is 2 0 2 0 , S y n th e s is R e p o r t a n d A n a ly t ic a l R e p o r t , Seville: E u ropean C o m ­
m ission Jo in t Research Centre.
K a l m b a c h , P., and K u r z , H. D. (1990), “ M icro electro n ics an d E m p lo ym en t: A D ynam ic
In p u t-O u tp u t Stu d y o f the West G erm an Econom y,” S tr u c tu ra l C h a n g e a n d E co n o m ic
D y n a m ic s 1: 317-8 6 .
K a r a o m e r lio g lu , A ., an d A n s a l , T. (2000), “ C o m p en satio n M ech an ism s in D eveloping
C ou n tries,” in V ivarelli and Pianta 2 0 0 0 :16 5 -8 1.
K a tso u la co s , Y. S. (1986), T h e E m p lo y m e n t E ffe c t o f T e c h n ic a l C h a n g e , B righ to n : W heat-
sheaf.
* K r u e g e r , A . (1993), “ H o w C o m p u ters have C h an ged the W age Structure: Evidence from
M icro D ata 1984-1989,” Q u a r te r ly Jo u r n a l o f E c o n o m ic s 108: 3 3-6 0 .
L a y a r d , R ., an d N ic k e l l , S. (1985), “ The C auses o f B ritish U n em p loym en t,” N a tio n a l
In s titu te E c o n o m ic R e v ie w 111: 62-85.
L e o n t ie f ,W., and D u c h in , E (1986), T h e F u t u r e Im p a c t o f A u t o m a tio n on W orkers , O xford:
O xfo rd U n iversity Press.
о

L u n d v a l l , B. A. (1987), “ Technological U n em p loym en t in a Sm all O pen Econom y,” in

R. Lund, P. Pedersen, and J. Sch m id t-Soren sen (eds), S tu d ie s in U n e m p lo y m e n t, C o p en ­
hagen: N ew Social Science M o n ograp h s, Institute o f O rgan isation and Indu strial So ci­
ology, 21-4 8 .
M a c h in , S. (1996), “ C hanges in the Relative D em an d fo r S k i l l s i n A . B o o th , and D. Snow er
(eds), A c q u ir in g S k ills , C am b ridge: C am b rid ge U n iversity Press, 12 9 -4 6 ,
-------and V an R e e n e n , J. (1998), “ Technology and C hanges in Skill Structure: Evidence from
Seven O E C D C o u n tries,” Q u a rte rly Jo u r n a l o f E co n o m ic s 113 :12 15 - 4 4 .
------ A n d W a d h w a n i , S. (1991), “ The Effects o f U nions on O rgan isation al C hange and
E m p lo ym en t: Evidence from W IR S,” E c o n o m ic Jo u r n a l 10 1: 324 -30 .
M e g h ir , C ., R y a n , A ., and V an R e e n e n , J. (1996), “ Jo b C reatio n , T echnological Innovation
and A d justm en t C osts: Evidence fro m a Panel o f B ritish Firm s,” A n n a le s d ’ec o n o m ie et
sta tistiq u e 4 1-2 : 256-73.
M e v e r -K r a m e r , F. (1992), T he Effects o f N ew Technologies o n E m p lo ym en t,” E co n o m ics of
In n o v a tio n a n d N e w T ech n o lo g y 2:131-4 9 .
 INNOVATION AND EMPLOYMENT
597

Nelson , R*, and P h elps , E, (1966), Investment in Humans, Technological Diffusion and
Economic Growth,” A L A Papers a n d Proceedings 56: 69-75,
----- and Winter , S. (1982), A n E vo lu tio n a ry Theory o f Econom ic Change, Cambridge, Mass,:
Belknap Press.
Noble, D. (1984)» Forces o f P roduction : A Social H istory o f Industrial Autom ation , New York:
Knopf.
OECD (1994 )* The O E C D Jo b S tu d y E vid en ce a n d Explanations. Part f Labour M arket Trends
and Underlying Forces o f C han ge , Paris: OECD.
----- (i99h)» E m ploym en t a n d G row th in the K now ledge-B ased Econom y Paris: OECD.
Padalino, S., and V ivarelli , M. (1997), "The Employment Intensity of Economic Growth
in the G-7 Countries,” In tern a tio n al L a b o u r R ev iew 136:191-213.
Pasinetti, L. (1981), Structu ral C han ge a n d Econom ic G ro w th Cambridge: Cambridge
University Press.
Perez, C. (1983)» “ Structural Change and the Assimilation o f New Technologies in the
Economic and Social Systems,” Futures 15(5): 357-75.
*Petit, P. (1995) у “ Em ploym ent and Technological Change,” in P. Stoneman (ed), Handbook
of the Economics o f In n ovation a n d Technological Change, Amsterdam: North Holland,
366-408.
*----- and Soete , L. (eds) (2001л), Technology a n d the Future o f European Employment,
Cheltenham: Edward Elgar.
----------- (2001b), "Technical Change and Employment Growth in Services: Analytical and
Policy Challenges,” in Petit and Soete 2001a: 166-203.
Pianta, M. (2000), "T h e Em ploym ent Impact o f Product and Process Innovation” in
Vivarelli and Pianta 2000: 77-95.
----- (2001), "Innovation, Dem and and Employment,” in Petit and Soete 2001a: 142-65.
----- Evangelista , R., and Perani, G. (1996), "The Dynamics o f Innovation and Employ­
ment: An International Com parison,” Science , Technology Industry R eview 18: 67-93.
Рші, P. (1995), "Econom ic Growth, Technological Change and Employment: Empirical
Evidence for a Cum ulative Growth Model with External Causation for Nine OECD
Countries, 1960-1990,” Stru ctu ral C han ge a n d Econom ic D ynam ics 6: 185-213.
----- (1996), "An Integrated Cum ulative Growth Model: Empirical Evidence for Nine OECD
Countries, 1960-1990,” L a b o u r 10: 93-150.
Piva, M., and V ivarelli , M. (2002), "T h e Skill Bias: Comparative Evidence and an Econo­
metric Test,” In tern a tio n al R e v ie w o f A p p lie d Econom ics 16(3): 347-58.
Ricardo, D. (1951), P rin ciples o f P olitical E conom y a n d Taxation, in P. Sraffa (ed.), The Works
and Correspondence o f D a v id R icardo , vol. 1, Cambridge: Cambridge University Press (3rd
edn. 1821).
Sanders, M., and ter W eel , B. (2000), "Skill Biased Technical Change: Theoretical Con­
cepts, Empirical Problems and a Survey o f the Evidence,” Druid working paper, Copen­
hagen Business School and Aalborg University.
Schumpeter, J. A. (1934), T h eory o f E conom ic D evelopm ent , Cambridge, Mass.: Harvard
University Press (1st edn. 1911).
Shaiken , H. (1984), Work Transform ed: A utom ation a n d L abor in the C om puter Age, New
York: Holt, Rinehart and Winston.
Simonetti, R., T aylor , K., and V ivarelli , M. (2000), "M odelling the Employment Impact
of Innovation: Do Com pensation Mechanisms Work?” in Vivarelli and Pianta 2000.
598 M A R I O PI A N T A

S im o n etti , R. and T a n c io n i , M . (2002), “A M acro eco n o m etric M o d el fo r the A n alysis o f

the Im p act o f Technological C h an ge and Trade o n E m p lo y m e n t” Journal of Interdisciplin­
ary Economics 13 :18 5 -2 2 1.
S m o ln y , W. (1998), “ In n o vatio n , Prices and E m p lo ym en t: A T h eo retical M o d el and an
A p p licatio n fo r W est G erm a n M an u factu rin g Firm s,” Journal of Industrial Economics 46:
359-81.
S p ie z ia , V., and V iv a r e l l i , M . (2002), "In n o v a tio n an d E m p lo ym en t: A C ritical S u rv e y ” in
G reenan, U H orty, and M airesse 2 0 0 2 :10 1- 3 1.
S ylos L a b in i , R (1969), Oligopoly and Technical Progress, C am b rid ge, M ass.: H arvard
U n iversity Press (1st edn. 1956).
V a n R e e n e n , J. (1996), "T h e C reatio n and C ap tu re o f E co n o m ic Rents: W ages an d In n o v ­
ation in a Panel o f U K C om pan ies,” Quarterly Journal of Economics 111( 1) : 19 5-226 .
-------(i997)> “ E m p lo ym en t and Technological In n o vatio n : Eviden ce fro m U .K . M an u factu r­
ing Firm s,” Journal of Labor Economics 15: 255-84.
— — E v a n g e l is t a , R ., and P ia n t a , M . (1996), “ In n o vatio n an d E m p lo ym en t in the Italian
M an u factu rin g In d u stry ” Research Policy 2 5 :10 13 - 2 6 .
* V iv a r e lli , M . (1995), The Economics of Technology and Employment: Theory and Empirical
EvidenceyA ldershot: E d w ard Elgar.
*a n d P ia n t a , M . (eds.) (2000), The Employment Impact of Innovation: Evidence and Policy
L o n d o n : Routledge.
W h it l e y , J. D ., an d W ilso n , R. A . (1982), “ Q u an tifyin g the E m p lo ym en t Effects o f M icro -
Elect ronics,” Futures 14(6): 486-95.
W o lfe , E. (1996), “ T echn ology and the D em an d fo r Skills,” Science Technology Industry 18:
9 5-124 .
 C H A P T E R 22

SCIENCE,
TECHNOLOGY,
AND INNOVATION
PO LICY

BENGT-AKE LUNDVALL
SU SAN A BORRAS

22.1 In tr o d u c tio n

This chapter is about what governments have done and could do to

production, diffusion, and use o f scientific and technical knowledge m order
realize national objectives. , +
We begin the chapter w ith "story-telling” based on sketchy is о
aim o f the two stories is to illustrate that innovation policy covers a Wl *mDortant
that have been on the agenda far back in history while still P ^
today. We move on to sketch the history o f innovation policy, sp i 1 g
three ideal types: science, technology, and innovation po icy. e and highlight
men,, , „ d X r soorces to do so. Finally wo point to fttture challenges, and htgHtgh.
research opportunities.
боо BE N G T-A K E LU N D V A LL AND SU SA N A BORRAS

2 2 Л Л From G u n s. . .
O n e im p o rta n t early exam p le o f successful tech n o lo g y p o lic y w as th e in itiative by
H e n ry V I II to d evelop the co m p etitive p ro d u c tio n o f can n o n s m ad e o u t o f iro n in
E n g la n d in the first h a lf o f the sixteenth century. T h e reaso n fo r estab lish in g this
tech n o lo g y p o lic y p ro g ra m w as that E n g la n d d esp erately n eed ed m o re can n o n s to
w in the o n g o in g w a r w ith France, and that the b ron ze ca n n o n s at use w ere to o scarce
an d to o exp en sive (Y aku sh iji 1986).
O ne facto r m a k in g the success o f the p ro g ra m p o ssib le w a s th at E n g la n d had
access to iro n ore as w ell as forests fo r fuel. A n o th e r im p o rta n t facto r w as that skilled
p eo p le w h o w ere exp ert forgers h ad m o ved to E n g la n d fro m the C o n tin e n t, esp e­
cially fro m France. S o m e cam e because th ey h ad been exp elled becau se o f their
religio u s faith, w h ile oth ers w ere b ro u g h t to E n g la n d becau se o f th eir u n iq u e skills.
T h e p ro g ra m w as realized th ro u g h estab lish in g a c o n so rtiu m u n d er the lead ersh ip o f
W illiam L eavitt, h ead o f the royal iro n w o rk s in N ew b rid g e, and b rin g in g together
exp ertise co m in g fro m d ifferen t co u n tries, especially fo u n d ers w ith th eir ro o ts in
France.
M a n y o f the in stru m en ts o f tech n o lo g y an d in n o v a tio n p o lic y that w ere used at
that tim e are still in use today. T ech n ological d iffu sio n th ro u g h im m ig ra tio n was
im p o rtan t. C o m p e titio n p o lic y an d p ro p e rty rig h t regim es w ere m an ip u lated , and
p u b lic p ro cu re m e n t used. A fter a p erio d w h en several co m p e tin g co m p an ies w ere
allow ed to exp o rt can n o n s (pirates w ere a m o n g the m o re ad van ced users!) m o n o p ­
olist righ ts w ere gu aran teed to one p ro d u c e r (R alp h H o g g e). B a rrie rs to the m o v e ­
m en t o f skilled people to the C o n tin en t and e xp o rt licen sin g w ere established to
m ake sure that o n ly P rotestant co u n tries got access to the tech n o logy. Public
p ro cu rem en t w as certain ly a p o lic y in stru m en t that p ro m o te d the refin em en t o f
the tech n o logy, in this case w ith a n ation al secu rity objective.

2 2 Л .2 . . . to Butter
In the m id d le o f the nineteenth ce n tu ry the D an ish econ o m y w as h ig h ly dependent
o n the exp o rt o f co rn to E n glan d . D an ish rye served as the “ fu e l” d riv in g the English
h orse-b ased tran sp o rt system . In the second h a lf o f the ce n tu ry an d especially in the
1870s a d ram atic red u ction in overseas tran sp o rt costs gave R u ssia an d the U S easier
access to this an d oth er m arkets. Prices w ere b ro u g h t d o w n to a level w h ere D anish
fa rm in g co u ld not com pete an ym ore. A m a jo r ag ricu ltu ral crisis b ro k e out.
A t that tim e there w as already som e e xp o rt o f an im al p ro d u cts in clu d in g butter
esp ecially from the m a n o r houses. B u t the m o st im p o rta n t b a rrier w as the uneven
q u ality o f the p ro d u cts. Stan d ard ization w as a key elem ent in the tra n sfo rm a tio n o f
D an ish p ro d u ctio n and exp o rts aw ay fro m co rn to w ard b u tter an d oth er anim al
 S C I E N C E , T E C H N O L O G Y , AND I N N O V A T I O N POLICY 6Q1

products. This transform ation involved a combination o f technical and social

innovation and a com bination o f state and “ non-state intervention.” The public
policies involved were “ diffusion oriented” rather than “ mission oriented.” The
episode made deep im prints on the Danish innovation system that are still possible
to identify.
The most im portant technology brought into use was related to dairy processing.
To separate the fat from the m ilk the “ separator” was developed in Sweden, however
it was first produced in Denm ark and first brought into use in Danish dairies. But the
single most important innovation driving the transformation was social rather than
technical and it came neither from the state nor from the market. Grundtvig—
priest, philosopher, and nationalist— was a key person who founded a social move­
ment that changed the m ode o f production in Danish agriculture. In his writings
and in the lectures he held all over the country, he emphasized the need for farmers to
get educated and take on responsibility for their own fate. Over a short timespan
local “ folk high schools” spread quickly in the countryside.
Inspired b y this id e o lo g y an d fo rced b y eco n o m ic necessity farm ers got engaged in
cooperative o w n ersh ip a ro u n d n ew lo cal d a iry plants. In the 1880s, the num ber o f
cooperatively o w n ed d airies g rew fro m three to 700. Betw een 1850 and 1900 the share
of butter in D an ish e x p o rts to the U K w en t fro m о per cent to 60 per cent. The
cooperative fo rm o f o w n e rsh ip created a fram ew o rk su p p o rtive for rapid diffusion
of dairy techniques an d fo r stan d ard izatio n .
State policies su p p o rte d these develo p m en ts. T h e A gricu ltu ral U niversity in
Copenhagen w as estab lish ed as e arly as 1856 (W agner 1998). D a iry consultants
played an im p o rtan t ro le in the p ro cess o f d iffu sin g good practice and on this
basis the d airy o rien ted A g ric u ltu ra l R esearch L a b o rato ry was established in 1883.
This example might be o f special im portance for developing countries with “ soft
states.” In these countries ordinary state-led science, technology and innovation
policy might not be sufficient to overcom e the obstacles for economic development.
A broader social m obilization m ight be necessary in order to overcome barriers to
socio-economic development.

22.1.3 Modern Innovation Policy between Guns and Butter

If we go to the leading econom y in the world in terms o f technology the US we
find elements in the innovation system that remind us o f both these stories. 1 he
successful development o f the atom ic bom b at Los Alamos has many characteristics
in common with the “ gun-story.” It was a crash-mission project aimed at winning a
war and it brought together skilled people from different parts o f the world. It also
set the agenda for m uch o f what followed in terms o f technology policy not only in
the US but all over the world.
боа B E N G T - АКБ L U N D V A L L A N D S U S A N A B O R E A S

B u t the U S in n o v a tio n system does n o t have ch aracteristics in c o m m o n so lely

w ith the gu n -sto ry. O n e o f the m ost su ccessfu l exam p les o f in n o v a tio n p o lic y in the
U S h as been the u p g ra d in g o f ag ricu ltu ral activities in term s o f p ro d u c ts and
p ro d u ctivity. T h e estab lish m en t o f lan d u n iversities an d e xte n sio n services w ere
cru cial fo r train in g, an d also fo r research an d d evelo p m en t o f n e w tech n o lo gies and
p ro d u cts. It w as especially im p o rta n t fo r the rap id d iffu sio n o f n e w id eas am o n g
farm ers. In spite o f this success there h ave been few attem p ts to in tro d u c e sim ilar
d iffu sio n o rien ted po licies in relatio n to m a n u fa c tu rin g an d services. T h e U S n o w
has a ‘"M an u factu rin g E xten sio n P artn ersh ip ” created in the late 1980s, w h ich seem s
likely to su rvive p o litica lly alth o u gh its effectiveness is u n certain . W h en attem pts
w ere m ad e to ap p ly the m o d el to co n stru ctio n in d u stries in the U S , the attem pts
failed (N elso n 1982).

22.2 Scien ce Po lic y, T ech n o lo gy P o lic y,

an d In n o v a t io n P o licy

A s can be seen fro m the oth er chapters in this b o o k — an d fro m the tw o stories told
a b o v e — in n o v a tio n p o lic y covers a w id e range o f in itiatives an d it is n ecessary to give
so m e stru ctu re to the co m p le x reality. O ne w ay to d o so is to in tro d u ce “ ideal types”
that b rin g o u t in a m o re d istin ct fo rm classes o f p h e n o m en a that are m u d d led and
m ixed in the real w o rld .
T h e d istin ctio n to be used here is betw een science p o licy, te c h n o lo g y policy, and
in n o vatio n policy. F o r each w e w ill d iscu ss h o w it b ecam e an exp licit p o lic y field,
w hat the m a jo r issues at stake are, w hat p art o f the in n o v a tio n system is in focus,
w h ich are the m ajo r actors in volved an d w h at are the in stru m en ts used. We w ill
p o in t to the critical sets o f data su p p o rtin g the p o licies. A n d since the O E C D has
played a u n iq u e role a m o n g in tern atio n al o rg an izatio n s in the d iffu sio n o f ideas
ab o u t in n o va tio n p o licy w e w ill use O E C D d o cu m en ts to illu strate so m e o f the m ore
im p o rta n t shifts in the p o st-w a r debate.
It w o u ld be m islead in g to argu e that w e pass fro m science p o lic y to tech n o lo gy
p o lic y an d then to in n o vatio n p o licy as w e pass fro m on e h isto rical stage to another.
For instance som e o f the classical issues in science p o lic y are very h ig h o n the current
p o licy agen da (Pavitt 1995; M artin an d Salter 1996). C u rren t d evelo p m en ts in
b io tech n o lo g y and p h arm aceu ticals have m ad e the d istan ce fro m b asic research to
co m m ercial ap p licatio n m u ch sh orter and therefore the o rg a n iz a tio n o f un i versities
has becom e a m a jo r issue (see C h . 8 b y M o w e ry an d S am p at). G en etic en gin eerin g
 S C I E N C E , T E C H N O L O G Y , A N D I N N O V A T I O N P O L I CY
603

Box 22.1 OECD and th e e v o lv in g d isc o u rse around science, te c h n o lo g y

a n d in n o v a t io n p o lic y

The OECD has played a key role in the evolution of the understanding of the policy
fields discussed in this chapter. It is certainly one o f the best sources for internationally
comparable data on science technology and innovation. Data are accessible through
regular publications in the form o f periodical policy reviews and through data bases
that are regularly updated. But it is also interesting to follow the policy discourse
organized at the OECD secretariat. What has been said at OECD meetings and
recommended by its expert groups might not always be transformed into practical
use in member countries but it reflects the new ideas. Therefore a brief summary of the
most important reports (in fact they all tend to arrive at the beginning of a new decade)
will be made here.
OECD 1963: Rationalizing science policy and linking it to economic growth. In an
OECD document from the beginning o f the sixties (OECD 1963a) а shift in attention
toward economic objectives was signaled for science policy. This document with
Christopher Freeman, Raym ond Poignant, and Ingvar Svennilsson as major contribu­
tors is quite remarkable in the emphasis it gives to national and rational planning. It
came the same year as the Frascati meeting on a new manual for gathering R&D
statistics took place and it argues for a strong link between better data on R&D and
more systematic policy. This report obviously aimed at giving science policy legitimacy
outside the narrow circles o f ministries o f education and science.
OECD 1970: Bringing in human and social considerations on technology policy.
This report (the “ Brooks Report” ) introduced a broader social and ecological perspec­
tive to science and technology policy. It also gave strong emphasis to the need to
involve citizens in assessing the consequences o f developing and using new technolo­
gies. The new focus reflected a combination o f growth satiation and dissatisfaction
with the social consequences o f technical change. Therefore it was also assumed that
mom should be given for wider concerns and the uncritical optimism was challenged.
OECD 1980; Innovation policy as an aspect of economic policy. The OECD expert
report “ Technical Change and Econom ic Policy” (OECD 1980) redefined the agenda
for innovation policy in the light “ o f a new economic and social context/ Its message
was that the slow-down o f growth and the increase in unemployment could not be seen
as something that m acroeconom ic expansionary policies could solve. Neither did the
report restrict itself to recommending an increase in investments in science and R&D.
As compared to earlier reports on science and technology, the focus was moved toward
die capacity o f society to absorb new technology. The experts included, among others,
Freeman, Nelson, and Pavitt. It is interesting to note that this report was published
simultaneously with a more traditional document on innovation policy (1980). This
was obviously a period when innovation policy began to be regarded as a legitimate
i k y field.
OECD 1990: Innovation defined as an interactive process. The TEP project, edited
id given its flavor by Francois Chesnais and with Luc Soete as one of the leading
Jemal experts brought together in one coherent framework a broad set of new
search results on innovation and used those to point in new policy directions
604 BE N G T-A K E L U N D V A L L AND S U S A N A BORRAS

(O ECD 1990), The report took as its starting point in n o v a tio n as m in t o c t n ^

process” and gave a prominent role to national innovation systems as an organizing
concept There was a stronger emphasis on network formation, new form s o f organiza­
tion and industrial dynamics than in earlier O ECD contributions to innovation policy.
The demand side and strengthening the absorptive capacity o f firms as well as the
feedback from users to the supply side were given strong emphasis. This report gave the
systemic version o f innovation policy an analytical foundation,
O ECD 2001: The new economy beyond the hype. In the middle o f the 1990s the
idea o f a new economy based upon IC T and vibrant entrepreneurship began to be
diffused in the U S with Federal Reserve President Alan Greenspan as an important
source. At the end o f the millennium O ECD began to analyze the phenomenon and its
first reports came in 2001. These reports were interesting because they were coordin­
ated by the Economics Department at O ECD that so far had treated innovation as a
secondary phenomenon and had been a spokesman for a pure market innovation
policy. The catalog o f recommendations with regard to innovation policy is more
extensive than what O ECD economists had ascribed to so far— see for instance the Jobs
Study (OECD 1995). The new economy episode is interesting because it is the first time
that innovation becomes widely accepted among economists as a fundamental factor
that needs to be analyzed and understood. At the same time, it is clear that the basic
hypothesis was too simple and the policy recommendations remained colored by the
traditional pro-market and anti-state philosophy o f the O ECD economists.

b rin gs in fu n d am en tal issues ab o u t ethics that w e n o rm a lly w o u ld co n n ect w ith

science policy. In the real w o rld the fo rm s o verlap an d m ix.

22.2.1 From Science Policy...

T h e tw o h isto rical exam p les given above b elo n g to the realm o f te c h n o lo g y p o licy
rath er th an science policy. Science p o lic y is a co n cep t that b elo n g s to the p o st-w ar
era. B efo re the w ar, region al an d federal go vern m en ts w ere fu n d in g u n iversity
research an d the tra in in g o f scientists. B u t th ey d id so p rim a rily fo r h isto rical and
cu ltu ral reasons and, b efore W orld W ar II, the idea o f science as a p ro d u ctive force
w as taken up m a in ly in the p lan n ed econ o m ies.
A cco rd in g to C h risto p h e r Freem an science p o lic y w as reco gn ized as a p o licy area
th ro u g h the p io n eerin g w o rk b y B ern al (1939)- B e rn a l w as a p io n e e r in m easu rin g the
R & D effort at the n ation al level in E n g lan d an d he stro n g ly reco m m en d ed a
d ram atic increase in the effo rt since he w as co n vin ced that it w o u ld stim ulate
e co n o m ic grow th and w elfare. In the U S , the V an n evar B u sh rep o rt fro m 1945,
“ Science: T h e Endless Frontier,” has a sp ecific status in d efin in g an agen da fo r the US
p o st-w a r science (and tech n o logy) policy. It d efin ed the task fo r science p o licy as
 S C I E N C E , T E C H N OLOGY, AND INNOVATION POLICY 6 05

contributing to national security, health and economic growth. Like Bernal, the
Bush report gave strong emphasis to the potential economic impact of investments
in science.
The real reason for the breakthrough o f science policy and the increased public
investment in research was probably the way World War II ended and the Cold War
started. The success o f the Los Alam os project made plausible the idea that a massive
investment in science (especially physics but also chemistry and biology), applied
science and technological development could produce solutions to almost any
difficult problem, and underscored the importance o f science and technology for
national security. This pressure to invest in the promotion o f science was reinforced
in the arms and the space races between the US and Soviet Union. In 1957 the launch
of Sputnik put extra pressure on the West, and especially the US, to invest in defense
and space-related research.
The major issues in science policy are about allocating sufficient resources to
science, to distribute them wisely between activities, to make sure that resources are
used efficiently and contribute to social welfare. Therefore, the quantity and quality
of students and researchers receives special attention. The objectives for science
policy that are actually pursued by governments are mixed and include national
prestige and cultural values besides social, national security, and economic
objectives.
The elements o f the innovation system that are focused upon are universities,
research institutions, technological institutes, and R&D laboratories. Science policy
is both about the internal regulation o f these parts o f the innovation system and
about how they link up to the environment— not least to government and industry.
However, strengthening this linkage becomes even more crucial in technology and
innovation policy.
There are two m ore or less standing debates within the science policy community.
The first debate has to do with how far scientific progress is identical with progress in
general. Critical scholars would point to how science is abused in the control o f
people and nature— including genetic manipulation and undermining ecological
sustainability. Those m ore positive to science would respond that none o f this can be
seen as emanating from science: rather, it should be seen as the result o f unwise use of
science. Scholars from hum anities and sociology might more frequently join the first
camp while scientists and technologists have a tendency to join the second.
The second debate is about to what degree science should be made the obedient
servant o f the state and/or capital and to what degree it should be autonomous.
Sociologists o f knowledge have devoted considerable attention to this question, also
in relation to different national research policy styles— with more or less govern­
mental steering and with different levels o f aggregation— large/small research activ­
ities (Rip and van der M eulen 1997; Jasanoff 1997). Others have argued that there is a
shift in the mode o f research and knowledge production, with different effects upon
the autonomy o f science (Cozzens et al. 1990; Gibbons et al. 1994; Jasanoff 2002).
6o6 BENG T-A KE LU N D V A LL AND SU SA N A BORRAS

U n iv ersity sch olars tend to argu e that “ freed o m ” an d “ a u to n o m y ” o f acad em ic

research is im p o rta n t fo r at least tw o reason s. O n e is the lo n g -te rm valu e o f
seren d ip ity— o n ly w h en b asic research is allow ed to m o ve alo n g its o w n trajecto ries
w ill it p ro d u ce the u n su spected that can o p en u p n ew aven u es fo r a p p lie d research
an d tech nical so lu tio n s. T h e secon d reaso n is that critical science is an im p o rta n t
elem ent in m o d e rn d em o cracy, becau se scien tific k n o w le d g e fro m in d ep en d en t
sou rces is an im p o rta n t in p u t fo r o p en , tran sp aren t an d rep resen tative p o litical
d ecision m akin g. M o st sch olars o f in n o v a tio n w o u ld agree th at th is is an area w h ere
there is a trad e-off. W h ile the idea o f basic research as “ free” scien ce— sign alin g a
co m p lete absence o f d irection an d use— is an illu sio n , the m assive su b o rd in a tio n o f
scien ce u n d er p o litica l an d e co n o m ic interests w o u ld certain ly u n d erm in e its lo n g ­
term co n trib u tio n to so ciety and econom y.
T h e m ain p o lic y actors in the p u b lic sector are m in istries o f e d u ca tio n and
research and research co u n cils. B u t sector m in istries in ch arge o f health , defense,
energy, tran sp o rt, and en viro n m en t m a y also p la y a role since th ey o rgan ize their
ow n research co m m u n ities, an d in so m e in d u strial eco n o m ies acco u n t fo r the
m a jo rity o f p u b lic sp en d in g o n R & D . M in istrie s o f fin an ce p la y a role w h en it
co m es to decide the total b u d get fo r research. C iv il o rg a n iz a tio n s rep resen tin g
co n su m ers an d citizens m a y be in vok ed as co rrectio n s to a b ias in fa v o r o f c o m m e r­
cial interests.
T h e in stru m en ts used are b u d g e ta ry d ecisio n s o n allo ca tin g fu n d s to p u b lic
research o rgan izatio n s, such as u n iversities, an d su b sid ies o r tax re lie f fo r p rivate
firm s. F in d in g in stitu tio n al m ech an ism s that lin k u n iversities and p u b lic la b o ra to r­
ies to the users o f research is o f cou rse a fu n d am en tal issue. B u t it b eco m es even m ore
so w h en w e tu rn to tech n o lo g y and in n o va tio n p o lic y an d th erefo re w e w ill save the
d iscu ssio n o f such in stru m en ts until later. D esig n in g in tellectu al p ro p e rty rights fo r
u n iversities has recently b ecom e a m a jo r issue (see C h a p te rs 10 b y G ra n stra n d and 8
b y M o w e ry an d S am p at in this vo lu m e ).
T h e evalu atio n o f research is an im p o rta n t p o lic y to o l an d it can be seen b o th as
creatin g incentives fo r sch olars and in stitu tio n s to b eco m e m o re effective an d as a
m eans to allocate p u b lic m oney. A cad em ic life h as b u ilt eva lu a tio n p ro ced u res into
it. To m ake an acad em ic career exam s m u st b e passed an d b o o k s an d articles w ill be
peer review ed. A t w e ll-fu n ctio n in g d ep artm en ts w eekly sem in ars exp o se o n go in g
w o rk to criticism fro m colleagues. T h ese days such processes in crea sin g ly involve
in tern atio n al expertise.
T h e criteria used for these o n g o in g evalu atio n s are m a in ly in tern al to scientific
co m m u n itie s and they m igh t be regarded as m isd irected b y extern al au th orities.
E valu atio n b y peers organ ized accord in g to d iscip lin es m a y p ro m o te “ m id d le o f the
ro a d ” w o rk rather than new ideas co m in g fro m cro ssin g d isc ip lin a ry b o rd ers. W here
this is the case, it m igh t be h elp fu l to establish altern ative sou rces o f reg ard in g in ter­
d isc ip lin a ry efforts m o re kindly. T h e E u ro p ean fram ew o rk p ro g ra m s fo r R & D have
represented an o p en in g in this d irection fo r E u ro p ean researchers.
 S C I E N C E , T E C H N O L O G Y , AND IN NOV ATI ON POLICY 607

Policy makers m ay also see internal evaluation as being too slack. This seems to
have been the case in the U K where a very ambitious and detailed reporting system
has been imposed on scholars doing research and teaching. So far the effect seems to
be a lot o f time used to report on research and that in England university scholars
have become the professional group most unsatisfied with working conditions.
While this kind o f draconian reform linking quantitative indicators o f performance
to access to public m oney might be useful in shaking up conservative institutions for
a brief period they tend to become a nuisance for the innovation system as a whole if
not loosened up again.
One o f the most fundam ental questions in science policy is whether it is true that
“good research is always useful research” or, more demanding, “ the higher the
scientific quality the m ore useful is the research.” I f it were true it would be a strong
argument for leaving at least some o f the allocation to the academic communities.
However, the evidence on this question is contradictory. While Zucker and Darby
(1998) demonstrate that star scientists are important for the success o f biotechnol­
ogy firms, Gittelman and Kogut (2003) show that at least within biotech there is no
clear relationship between prestigious publishing and high impact innovations— if
anything the relationship is negative (M owery and Sampat 2001).
It might be expected that the requirements between doing outstanding research
and creating well-designed technology are even more distinct in other fields where
the distance from scientific discovery to innovation is longer. The truth may be that
also science has a lot to learn from interacting with its users and that the best system
is one with several layers and with career shifts— some going only for “ excellence”
within a scientific discipline, others focusing only on user-needs and some operating
in a mode between the two.
Since the early 1990s, basic science has been constantly pushed by politicians to
demonstrate its social and econom ic usefulness. This has been termed the new
social contract o f basic science” in the age o f budgetary retrenchments, particularly
in Europe and the U S (M artin and Salter 1996). Several authors argue that this “ value
for money” attitude disregards some essential aspects o f science policy, such as the
training o f scientists and technicians, and the development o f knowledge capabil­
ities in areas where uncertainty about exploitation is so high that private investors
lack incentives (Sharp 2003) and that lessons from the US experience should be
examined critically, and along with for instance the Scandinavian and Swiss experi­
ences (Pavitt 2001).

22*2,2 **. to Technology Policy

Technology policy refers to policies that focus on technologies and sectors. 1 he era
of technology policy is one where especially science-based technologies such
6o 8 b e n g t -Ak e e u n d v a l l a nd su sa n a b o r r As

n u clear pow er, space tech n o logy, co m p u te rs, d ru gs an d gen etic en g in eerin g are seen
as b ein g at the v e ry core o f eco n o m ic grow th . T h ese tech n o lo g ies get in to fo cu s fo r
several reasons. O n the o n e h an d th ey stim u late im a g in a tio n b ecau se th e y m ake it
p o ssib le to do su rp risin g things— th ey co m b in e science w ith fiction . O n the o th er
h an d they o p en up n ew co m m ercial o p p o rtu n ities. T h e y are ch aracterized b y a h igh
rate o f in n o va tio n an d they add ress ra p id ly g ro w in g m arkets.
T ech n o lo gy p o lic y m ean s d ifferen t th in gs fo r c a tch in g -u p co u n tries th an it does
fo r h ig h -in c o m e co u n tries an d it m ig h t also m ean d ifferen t th in gs fo r sm all an d b ig
co u n tries. In b ig h ig h -in c o m e co u n tries the fo cu s w ill be o n estab lish in g a ca p a city
in p ro d u cin g the m o st recent scien ce-b ased tech n o lo gies, as w ell as a p p ly in g these
in n o va tio n s. In sm aller co u n tries it m ig h t be a q u estio n ab o u t b ein g able to ab sorb
an d use these tech n o logies as th ey co m e o n the m arket. C a tc h in g -u p co u n tries m ay
m ake efforts to enter into sp ecific p ro m isin g estab lish ed in d u stries u sin g n ew
tech n o lo gies in the p ro cess o f d o in g so.
C o m m o n fo r these strategies is that th ey ten d to d efin e “ strategic tech n o lo gies''
an d so m etim es the sectors p ro d u c in g th em are also d efin ed as strategic sectors. T h e
idea o f strategic sectors m a y be related to P e rro u x and to H irsch m a n , b oth students
o f Sch um p eter. P e rro u x used concep ts such as “ in d u strializin g in d u strie s" and
grow th poles w h ile H irsch m an in tro d u ced u n b alan ced g ro w th as a p o ssib le strategy
fo r less-d evelo p ed co u n tries (P e rro u x 1969; H irsch m a n 1969).
In the lead cou n tries, go vern m en t in itiatives o f the te c h n o lo g y p o lic y k in d w ere
triggered w h en n ation al p o litical o r e co n o m ic interests w ere th reaten ed and the
threats co u ld be lin ked to the co m m a n d o f sp ecific tech nologies. S p u tn ik gave extra
im p etu s to a fo cu s o n space tech n o lo g y an d the C o ld W ar m o tivated the m ost
am b itio u s tech n o lo g y p o lic y effo rt ever in the U S. In E u ro p e, S ervan Schreiber's
b o o k L e Deft a m ericain (Servan Sch reib er 1967) gave a p ictu re o f a g ro w in g d o m in ­
ance o f the U S m u ltin atio n al firm s especially in h ig h te c h n o lo g y sectors. It gave the
b ig E u ro p ea n co u n tries such as France, the U K , an d G e rm a n y in cen tives to d evelop a
p o lic y o f p ro m o tin g n ation al ch am p io n s in sp ecific sectors. A sp ecific im p o rta n t
event triggerin g French and, later, E u ro p ean efforts w as the e x p o rt em b argo o f
co m p u te r tech n o lo g y that w as seen in France as b lo ck in g its p ro g ress in the d evel­
o p m e n t o f nu clear technology.
T h e m o tiv a tio n b eh in d the tech n o lo g y p o licy in Ja p a n — an d later on in co u n tries
such as Taiw an an d K orea— is different. It is d riven b y a n atio n a l strategy a im in g at
catch in g up and in the Jap an ese case it has roo ts b ack to the M eiji rev o lu tio n w hen
the first ideas o f m o d ern izatio n based on im itatin g the te c h n o lo g y o f the W est w ere
fo rm ed .
A t this p o in t w e need to be aw are o f a n u m b er o f fu n d am en tal q u estio n s regard in g
tech n o lo g y policies.

* Is it at all legitimate and effective for the state to intervene for commercial reasons
in promoting specific sectors or technologies? Or is the only legitimate technology
 S C I E N C E , T E C HNO LO GY, AND IN NOVATION POLICY 609

policy one where national societal issues are at stake, including establishing
national military power? It is a paradox that in the country having the most
massive public intervention in terms o f technology policy (the US), most o f the
policy has been motivated by non-com m ercial arguments and the discourse has
been anti-state. Japan is the country with the most explicitly commercially driven
technology policy with a recognition o f a role for the state, and here the interven­
tion has been much m ore modest, at least in terms o f the amount o f public money
involved.
t What technologies should be supported? Is it always the case that high-tech and
science-based sectors should be given first priority? Again the Japanese govern­
ment as well as governments in smaller countries has been more apt to think about
the modernization o f old industries than the US and the big European countries.
* At what stage should the support be given? Should it be given only to “ pre-
competitive” stages or should it also be helpful in bringing the new products to
the market? In the second case there might be a combination o f government
support o f new technology and more or less open protectionism.
* What limits should be set for public sector competence? Technology policy may be
pursued with competence where government operates as a major user but when it
comes to developing new technologies for the market, the role o f governments
must be more modest. To be m ore specific, there are several historical examples o f
how government am bitions to make technological choices that reduce diversity
have ended in failure, for example, the “ minitel” experience in France, and the
high definition T V policy in the EU, both in the early 1990s.
* How can prom oting a technology or a sector best be combined with competition?
The period in the 1980s o f prom oting single firms as national champions in the
bigger European countries was not a great success while the Japanese public
strategy to prom ote “ controlled competition” among a handful o f firms was
more successful.
The objectives o f technology policy are not very different from those o f science
policy but— at least to begin with— it represented a shift from broader philosophical
considerations to a m ore instrumental focus on national prestige and economic
objectives. Technology policies were developed in an era o f technology optimism.
But later on— in the wake o f the 1968 student revolt— more critical and broader
concerns relating to technology assessment and citizen participation came onto the
agenda (OECD 1971).
The elements o f the innovation system in focus remain universities, research
institutions, technological institutes, and R & D laboratories. But the attention
moves from universities toward engineering and from the internal organization o f
universities toward how they link to industry. Technology policy may go even
further and include the com m ercialization o f technologies, but then we approach
what we will call innovation policy.
6 lO B E N G T -A K E L U N D V A L L AND S U S A N A BORRAS

In so m e co u n tries su ch as the U S , the m ain te c h n o lo g y p o lic y acto rs in the p u b lic

sector are sector m in istries p ro m o tin g and so m etim es p ro c u rin g te c h n o lo g y fo r
p u rp o se s o f teleco m m u n icatio n s, defense, health , tra n sp o rt, e n erg y etc. w h ile in
o th ers, such as Ja p an , th ey are m in istries in ch arge o f in d u stry an d trad e. M in istries
o f e d u catio n an d research are im p o rta n t since th ey o rgan ize the e d u ca tio n an d
t rain in g o f scientists an d engineers. A u th o rities in ch arge o f reg u latin g c o m p e titio n
as w ell as o th er regu latin g au th o rities m a y have a m a jo r im p a ct o n tech n o lo g y
p o lic y and o n tech n o lo g ical d evelo p m en t. P u b lic au th o rities m ay, as elem ents o f
tech n o lo g y policy, organ ize tech n o lo g y assessm ent an d o th er w ays o f in v o lv in g
citizens.
T h ere are m a n y p o ssib le in stru m en ts to be used in p ro m o tin g sp ecific te ch n o lo ­
gies an d sectors. M o st efficien t m a y be co m b in atio n s o f in stru m en ts in field s w here
p u b lic p ro cu re m e n t is in volved . W h en the g o vern m en t has the lead in g user c o m p e ­
tence, it is in a better p o sitio n to ju d g e w h at k in d o f in stru m en ts w ill w o rk (E d q u ist
et a l 20 0 0 ). B esides p u b lic p ro cu re m e n t d irect e c o n o m ic in cen tives in term s o f
su b sid ies an d tax red u ctio n s m a y be o ffered to firm s. S u p p o rtin g research at
u n iversities in the science fields in w h ich the n ew tech n o lo gies are ro o ted m ay be
an im p o rta n t p art o f a p u b lic m issio n policy. T h e d an ger o f these k in d s o f p o licies is
that “ in d u strial co m p lexes” co m b in in g the vested interests o f a segm en t o f p u blic
users w ith th o se o f a segm ent o f in d u stry em erge an d that a lack o f tran sp aren cy is
exp lo ited b y vested interests. A m o re subtle p ro b lem is the k in d o f co n vergen ce and
agreem en t o n the d irectio n o f tech n o lo g ical trajecto ries that m ig h t d evelo p in such
co m p lexes, exclu d in g n ew an d m o re p ro m isin g ven u es (L u n d v a ll 1985).
In areas w h ere the m a in ap p licatio n o f the new tech n o lo gies is co m m ercial, the set
o f in stru m en ts used m a y be a c o m b in atio n o f sector o r te c h n o lo g y sp ecific econ o m ic
in cen tives w ith m o re o r less p ro tectio n ist trad e policy. A n exa m p le m ig h t be the high
d efin itio n T V p o lic y o f the E U in the early 1990s, w h ere the attem p t to define a
c o m p u lso ry an alo g ical stan d ard w o u ld have been a tech n ical trad e b a rrier to
em ergin g d igital stand ard s, co m b in ed w ith sp ecific e co n o m ic in cen tives fo r E u ro ­
pean p ro d u cers. Su ch packages m a y create a sheltered a tm o sp h ere fo r the firm s
in volved . M o re p ro m isin g m a y be p ro ject-o rg an ized su p p o rt b rin g in g different
firm s an d kn o w led ge in stitu tion s together in o rd er to fo cu s o n gen eric an d co m m o n
and new tech n o logical p ro b lem s w h ile m ak in g sure that the use o f the n ew k n o w ­
ledge takes place in a glo bal co m p etitive clim ate. E xp erien ce also sh ow s that m akin g
the p rojects w ell d efin ed b oth in term s o f content and tim e bu t o p en in term s o f w hat
sp ecific type o f tech nical so lu tio n s sh o u ld be aim ed at, lim its the n egative im p a ct on
co m p etition .
W hile the evalu atio n o f research is im p o rta n t in science p o licy there are sim ilar
general p o licy tools that are useful w hen d esign in g and red esign in g tech n o lo g y
policy. T ech n o logy forecastin g is a w ay o f cap tu rin g n ew tech n o lo g ical trends.
Asking lead in g experts am o n g scientists and a m o n g the m o st ad van ced p ro d u cers
and users ab out w hat tech nologies are risin g o n the h o rizo n helps to sco u t the next
 S C I E N C E , T E C H N O LO G Y, AND INNOVATION POLICY 6ll

generation o f “ strategic technologies" In order to limit the capturing o f public

interest by private firms, independent policy evaluation o f specific initiatives may be
useful Many evaluations end up addressing users o f the programs with questions
about the efficacy o f the program . Not surprisingly, such studies often end up
reporting that the program was very good and that more o f the same would be
welcome. In this situation, as in m any other situations, where too much agreement
among partners threatens to become a lock-in, it should be considered whether to
give “outsiders" a strong role as evaluators. It is as important for public policy as it is
for science-based firms to prom ote “ job rotation" and “ interfunctional teams."
As pointed out, science and technology policy are ideal types, which serve our
broad analytical purposes. In the real world o f advanced capitalist economies,
however, the policy focus, instruments and actors involved in science and technol­
ogy policy-making are not always easily grouped in one or the other o f these
categories. As we will exam ine now, innovation policy takes a step further by
bringing in an even broader set o f policy issues.1

22,2.3 • • *and to Innovation Policy

Innovation policy appears in two different versions. One— the laissez-faire ver­
sion-—puts the emphasis on non-interventionism and signals that the focus should
be on “ framework conditions" rather than specific sectors or technologies. This
often goes with a vocabulary where any kind o f specific measure gets grouped under
the negative heading “ picking the winners." The extreme version o f this type o f
innovation policy is one where basic research and general education are seen as the
only legitimate public activities and intellectual property right protection as the only
legitimate field for government regulation. In more moderate versions public
initiatives aiming at fostering “ entrepreneurship" and promoting a positive attitude
to science and technology in the population may be endorsed.
The other version may be presented as the “ systemic" version and by referring to the
concept o f “ innovation system." This perspective implies that most major policy fields
need to be considered in the light o f how they contribute to innovation. A fundamen­
tal aspect o f innovation policy becomes the reviewing and redesigning o f the linkages
between the parts o f the system. The first approach is built upon the standard
assumption made in economics that firms always know what is best for them
and that they norm ally (in the absence o f market failure) act accordingly. The second
perspective takes into account that competence is unequally distributed among firms
and that good practice in terms o f developing, absorbing and using new technolog} is
not immediately diffused among firms; and that failures may extend beyond
neoclassical “ market failure” to subsume “ failures” o f institutions to coordinate,
link, or address various systemic needs, etc. (see Ch. 7 by Edquist in this volume).
6l2 BENGT-AKE LUNDVALL AND SU SANA BORRAS

B o th o f these ap p ro ach es co ver all aspects o f the in n o v a tio n pro cess— in clu d in g
d iffu sio n , use an d m ark etin g o f n ew tech n o lo gies— an d in a sense th e y m a y be seen
as an im p o rta n t fo rm o f “ e co n o m ic p o lic y ” w h ere the fo cu s is m o re o n in n o v a tio n
th an on allo catio n . B o th ten d to p u t stro n g er em p h asis o n “ in stitu tio n s” and
“ o rg a n iz a tio n s” th an d o science and tec h n o lo g y policy. In the laissez-faire version ,
the p re d o m in a n ce o f the m ark et an d o f co m p e titio n b eco m es the m o st im p o rta n t
p rereq u isite fo r in n o v a tio n — there is in p rin cip le o n e sin gle re c o m m e n d a tio n fo r
in stitu tio n al d esign v a lid fo r all co u n tries.
In the system ic a p p ro ach the im p o rta n ce o f co m p e titio n is reco gn ized b u t so is
the need fo r clo ser co o p era tio n v ertically betw een users an d p ro d u c e rs an d so m e ­
tim es even h o rizo n ta lly a m o n g co m p etito rs w h en it co m es to d evelo p generic
tech n ologies. In the system s a p p ro a ch it is reco gn ized that the in stitu tio n a l set-up
d iffers acro ss n ation al eco n o m ies an d that this has im p lic a tio n s fo r w h at types o f
tech n o lo gies an d sectors thrive in the n atio n al con text. To d esign a su itab le in n o v ­
atio n p o lic y req u ires specific in sigh ts in the in stitu tio n a l ch aracteristics o f the
n atio n a l system .
In n o v a tio n p o lic y does n o t im p ly an y a p rio ri preferen ce fo r h igh versu s lo w
tech nology. T h e system s a p p ro ach in tro d u ces a vertical p ersp ective o n the in d u strial
system , seeing it as a n etw ork an d as valu e ch ains w h ere certain stages m ig h t be m ore
su itable fo r firm s in a sp ecific c o u n try
T h e respective theoretical fo u n d atio n s o f the tw o d ifferen t v ersio n s o f in n o vatio n
p o lic y are (1) an a p p lica tio n o f stan d ard n eoclassical e co n o m ics on in n o vatio n ,
and (2) a lo n g -te rm o u tco m e o f research o n in n o v a tio n an d e c o n o m ic evo lu tio n
(M etcalfe 1995; M etcalfe an d G e o rg h io u 1998). T h e in n o v a tio n system a p p ro a ch m ay
be seen as b rin g in g together the m o st im p o rta n t stylized facts o f in n o va tio n . It
m akes use o f em p irical m aterial an d an alytical m o d els d evelo p ed in in n o vatio n
research, as w ell as in in stitu tion al and e v o lu tio n a ry econ o m ics.
T h e m a jo r reason fo r in n o va tio n p o lic y b eco m in g m o re b ro a d ly used as a concept
w as the slo w -d o w n in e co n o m ic grow th a ro u n d 1970 and the persisten ce o f sluggish
grow th as co m p a red to the first p o st-w a r decades. T h e reason s fo r the slo w -d o w n in
the grow th in “ total facto r p ro d u c tiv ity ” w ere, an d still are, n o t w ell u n d ersto o d but
there w as a feelin g that it h ad to do w ith a lack o f c a p a b ility to e xp lo it tech n ological
o p p o rtu n ities. A t the sam e tim e, the restriction s im p o sed o n gen eral econ o m ic
p o lic y b y fear o f in flatio n m ad e it im p o rta n t to u n d erstan d the po ssib ilities to
p ro m o te grow th fro m the su p p ly side.
T h is im plies that the m ajo r o bjectives o f in n o va tio n p o licy are e c o n o m ic grow th
and in tern atio n al com petitiven ess. In the E u ro p ean U n io n d isco u rse these o b je ct­
ives are co m b in ed w ith “ social co h esio n ” an d equality. In n o v a tio n m ig h t also be
seen as a w ay to solve im p o rta n t p ro b lem s relating to p o llu tio n , energy, u rb an ism ,
an d poverty. B u t the m ain focus is o n the creatio n o f e co n o m ic w ealth.
A m o n g the in stru m en ts to be used are the reg u latio n o f intellectu al p ro p erty
rights and access to ven tu re capital. One fu n d am en tal d istin ctio n in in n o vatio n
 S C I E N C E , T E C H N O L O G Y , AND INNOVATION POLICY
613

Box 22*2 The neoclassical economics of innovation policy

According to neoclassical economics a necessary condition for public policy interven­

tion is market failure* If markets can do the job there is no need to intervene. Market
failure may have different causes but the ones most often raised in the context of
innovation policy are lack of incentives to invest in knowledge production* Knowledge
tends to be seen as a public good from which it is difficult to exclude others and also as
being non-rival since its user value may not suffer from the fact that others use it. When
knowledge is rival but non-excludable, intellectual property rights can be guaranteed
and enforced by governments. When knowledge is non-excludable and non-rival
governments should subsidize knowledge production addressed for public use or
take charge of producing the knowledge by itself.
The problem with this analysis is not the conclusions reached. There are certainly
good reasons for governments to support knowledge production and innovation in the
ways referred to— actually we have seen that most of these instruments were taken into
use long before neo-classical economics was established. The problem is that the
argument for support comes from a theory based upon assumptions that are incom­
patible with a dynamic economy where innovation is a widely spread and ongoing
process.

On rationality, markets, and competition

Innovation research has demonstrated that innovation is a ubiquitous phenomenon in
the modern economy. In such an economy the idea that the “representative firm” can
operate on precise calculations and choose among well-defined alternatives is a
dubious abstraction. The point, made by Kenneth Arrow and others, is that innovation
by definition involves fundamental uncertainty. Or as Rosenberg puts it “it is not
possible to establish the knowledge production function” since the output is unknown
(Rosenberg 1972: 172).
To assume that agents know what there is to know and to disregard competence
building tends to miss what is at the very heart of competition in a learning economy.
The assumption that markets are “pure” with amfs-length and anonymous relation­
ships between producers and users is logically incompatible with the fact that a major
part of innovative activities aims at product innovations. The only solution to the
paradox is that real markets are organized and constitute frameworks for interactive
learning between users and producers (Lundvall 1985).

policy lies between initiatives aim ing at promoting innovation within the insti­
tutional context and those aim ing at changing the institutional context in order to
promote innovation. The first category overlaps with instruments used in science
and technology policy. The second m ay include reforms o f universities, education,
labor markets, capital markets, regulated industries, and competition laws.
One fundamental question is how far these other policy fields can and should be
adapted to the needs to prom ote innovation. Universities and schools obviously
have other missions than to support innovation and economic growth. And the
614 BENGT-AKE LUNDVALL AND SUSANA BOREAS

m in istries and au th o rities in ch arge o f th em are n o rm a lly a ctin g a cco rd in g to a

d ifferen t lo gic th an the on e im p lied b y a fo cu s o n in n o v a tio n . O n e in terestin g
exam p le is la b o r m arket p o lic y w h ere it h as been assu m ed that w h a t m atters is
flexib le m arkets w h ere fle x ib ility refers b o th to w orkers" m o b ility an d to flexible
w ages. To th in k ab o u t h o w to d esign la b o r m arket in stitu tio n s so that th e y p ro m o te
co m p eten ce b u ild in g an d in n o v a tio n m ig h t lead to d ifferen t c o n clu sio n s regard in g
w h at is a “ w e ll-fu n c tio n in g la b o r m arket.5"
A secon d issue is ab o u t the lim its fo r p u b lic secto r in terv en tio n . A ssu m e, fo r
instance, that it can b e d em o n strated that the w a y firm s o rgan ize them selves
in tern ally has a m a jo r im p a ct o n in n o va tio n p e rfo rm an c e an d that e c o n o m ic grow th
can be exp lain ed b y differences in this respect. Is there a role fo r g o vern m en ts in
p ro m o tin g the d iffu sio n o f g o o d practices in this respect o r sh o u ld it still be left to
m an agem en t and ow n ers to cope w ith such p rob lem s? It is n o t se lf-ev id e n t that
go vern m en t sh o u ld help to d iffu se tech n ical so lu tio n s w h ile k eep in g aw ay fro m
su p p o rtin g the d iffu sio n o f m o re efficien t o rg a n iz a tio n a l so lu tio n s.
A s F igu re 22.1 indicates, the elem ents o f the in n o v a tio n system still include
u n iversities, research in stitu tio n s, tech n o lo g ical institutes, an d R & D lab orato ries.
H ow ever, the fo cu s o f p o lic y m oves fro m u n iversities an d tech n o lo g ica l sectors, as in
science an d tech n o lo g y po licies, tow ard all p arts o f the e c o n o m y th at h ave an im pact
o n in n o va tio n processes. T h is is the reaso n w hy, in F ig u re 22.1, the in stru m en ts o f
in n o v a tio n p o licy are also those o f science an d tech n o lo g y policy. In n o v a tio n p o licy
pays special atten tion to the in stitu tio n al and o rg a n iz a tio n a l d im e n sio n o f in n o v ­
atio n system s, in clu d in g co m p eten ce b u ild in g and o rg a n iz a tio n a l p erfo rm an ce.
In n o v a tio n p o lic y calls fo r “ o p en in g the b lack b o x 55 o f the in n o v a tio n process,
u n d erstan d in g it as a so cial an d co m p le x process.
M in istrie s o f e co n o m ic affairs o r m in istries o f in d u stry m a y be the ones p layin g a
co o rd in a tin g role in relation to in n o va tio n p o lic y b u t in p rin c ip le m o st m inistries
co u ld be in volved in efforts to redesign the n a tio n a l in n o v a tio n system . T h is is
actu ally the case fo r so m e sp ecific co u n tries, w h ich have exp erien ced a tru ly “ in n o v ­
atio n p o lic y t u r n 5 since the late 1990s, like fo r exam p le F in lan d , T h e N eth erland s,
an d D e n m a rk (B iegelb au er and B o rra s 2003). D eve lo p in g an in teractio n and d ia ­
lo gu e on p o lic y design betw een g o vern m en t au th o rities on the o n e h an d an d the
busin ess co m m u n ity, trade u n io n s an d k n o w led ge in stitu tio n s o n the o th er is a
necessary c o n d itio n for d evelo p in g so cially relevant and clear p o lic y p ro g ra m s that
can be im p lem en ted successfully.
T h e an alytical basis o f in n o v a tio n p o lic y co u ld be a c o m b in a tio n o f general
insights ab ou t w h at constitu tes g o o d practice, given the g lo b al co n text in term s o f
tech n o lo gy an d co m p etitio n , w ith specific insights in the ch aracteristics o f the
n ation al in n o vatio n system , T h e system can be analyzed in term s o f sp ecialization ,
in stitu tion al set-u p, and in sertio n in the global econ om y. A ttem p ts m ay be m ad e to
locate strengths, w eaknesses, threats, an d o p p o rtu n ities th ro u g h intelligen t b en ch ­
m arkin g. M issin g links and links co n stitu tin g lo ck -in s m a y be especially im p o rta n t
 S C I E N C E , T E C H N O L O G Y , AND INN O V AT IO N POLICY
615

Science policy
Focus: P ro d u c tio n o f s c ie n tific know ledge

In s tru m e n ts :
*P ubiic research funds granted in com petition
• (Semi-*) Public researchinstitutions (i,e:
laboratories, universities, Research centers..,)
• Tax incentives to firm s
• H igher education
• In te lle c tu a l Property Rights

Technology policy
Focus: A dvancem ent and co m m ercialization o f sectorial
te ch n ica l know ledge

In s tru m e n ts :
•P u b lic procurem ent
• P ublic aid to s tra te g ic sectors
• Bridging in s titu tio n s (between research w orld and industry)
• Labor force tra in in g and im provem ent o f technical skills
• S tandardization
• Technology forecasting
•B enchm arking in d u stria l sectors

Innovation policy

Focus: Overall innovative performance of the economy

Instruments:
• Im proving individual skills and learning a b ilitie s (through general education system and labour training)
• Im proving o rg a nizational perform ance and learning (i,e. ISO 9000 standards, quality control, etc.)
•Im proving access to in fo rm a tio n : in fo rm a tio n society
■E nvironm ental regulation
• B ioethical regulation
• Corporate law
•C o m p e titio n re g ulation
• Consumer p ro te ctio n
• Im proving social ca p ita l fo r regional developm ent: Clusters and industrial districts
• In te llig e n t bench m arking
• In te llig e n t, reflexive and dem ocratic forecasting

Figure 22.1 Relationship between science, technology, and innovation policy

to locate. Human resource development and use is another important dimension,

Finally, innovation policy analysis must increasingly give attention to the inter­
national dimension.

22.3 STI P o lic y E v a lu a tio n and

Im pa c t M ea su rem en t

It is characteristic for the evolution o f new policy fields that measurement and
quantitative guidance gives more legitim acy to the field. Within the field what can be
біб B E N G T - АКБ L U N D V A L L A N D S U S A N A B O R R A S

m easu red so m etim es gets m o re p o lic y atten tio n th an w h at ca n n o t b e m easu red .

A sp ecial p ro b lem an d im p o rta n t area fo r fu rth er research is to m e a su re the im p act
o f policy.

22.3.1 Measuring the Impact o f STI Policy

In the 1930s, B ern al m ad e the first attem p t to m easu re the effo rt m a d e in science by
relatin g R & D exp en d itu re to the n atio n al in co m e o f the U K . In the late 1950s and
early 19 6 0s, C h risto p h e r F reem an p layed a k e y role in d ev elo p in g the an alytical basis
o f science p o lic y a n d it is sig n ifican t that he also w as o n e o f the arch itects b eh in d the
Frascati m a n u a l that in 1963 gave the O E C D an d n a tio n a l a u th o rities m eth o d s to
m easu re R & D an d co m p are the effo rt across co u n tries (O E C D 1963b).
Today, n atio n al R & D statistics are qu ite detailed. T h e y sh o w the effo rt m ade
w ith in resp ectively p rivate an d p u b lic sectors as w ell as the fin a n c ia l so u rce o f the
investm ents. T h e exp en d itu re can be an alyzed a cco rd in g to p u rp o se . B ib lio g rap h ic
m eth o d s a llo w us to locate the scien tific fields in w h ich a sp ecific c o u n try h as its
relative stren gth — u sin g citatio n freq u en cy even the q u a lity o f research in different
co u n tries m a y be assessed.
In p rin cip le it is p ossib le to co n stru ct p ro d u c tiv ity m easu res fo r research using
scien tific articles in the n o m in a to r and resources u sed in term s o f m o n e y o r m a n ­
p o w er in the d en o m in ato r. O ne p ro b lem w ith u sin g such cru d e m easu res to guide
p o lic y is, o f course, that there are o th er o u tp u ts n o t so easily m easu red . T h e am o u n t
an d q u a lity o f students an d scientists train ed m a y be b ro u g h t in to the an alysis, w h ile
the in teractio n w ith users o u tsid e k n o w led ge in stitu tio n s m a y be less easy to quantify.
In the field o f tech n o logy, the data on patents are especially attractive since they
exist fo r lo n g p e rio d s a n d in clu de q u ite rich in fo rm a tio n a b o u t the te c h n o lo g y and
agents tak in g o u t the patent. Patent statistics can be u sed to co m p a re n atio n al
system s in term s o f tech n o logical sp ecializatio n — as revealed tech n o lo g ica l a d v a n ­
tage— an d it m ig h t even be p ossib le to d istin g u ish betw een m o re o r less im p o rta n t
paten ts u sin g citatio n patterns.
H ow ever, it has to be taken into accou n t th at patents p la y v e ry d ifferen t roles in
d ifferen t sectors and in so m e (e.g. p h arm aceu ticals an d b io tech n o lo g y ) th ey m ight
be m o re relevant w h en it co m es to ju d g in g p e rfo rm an ce th an th e y are in o th ers (e.g.
so ftw are and service p ro d u c tio n ). T h e m a jo r use o f p aten t statistics m a y th erefore be
to help m a p p in g the evo lu tio n o f n atio n al in n o va tio n system s rath er than as
p e rfo rm an ce in d icato rs to ju d g e the efficien cy o f tech n o lo g y policy.
T h e system ic view o f the in n o va tio n p o lic y m eans that these p re v io u s m easu re­
m ents are n ecessary but far fro m su fficien t to investigate the in n o v a tiv e p erfo rm an ce
o f an econom y. T h e O slo m an u al fo r gath erin g in fo rm a tio n and data o n in n o vatio n ,
w hich w as agreed u pon in 1990, is an im p o rta n t step in this d irectio n . In E u ro p e, the
 S C I E N C E , T E C H N O L O G Y , AND I N N O V A T I O N POLICY 6l J

Community innovation surveys have been collected several times in most o f the
member states (see Ch. 6 by Smith in this volume).
Among the m ore interesting inform ation that can be obtained through these
surveys is the share o f new products in total sales in firms in different sectors and
countries. This is a measure o f diffusion o f product inno vations in the economy and
may be seen as an im portant intermediate performance variable. Another field
where the diffusion o f technology has been mapped quite thoroughly is in relation
to information and com m unication technologies. These indicators are important
since for innovation policy one m ajor performance indicator should be the diffusion
and effective use o f new technologies.
The most im portant rem aining tasks for building indicators to support innov­
ation policy relate to the diffusion o f process innovations, innovation in services,
organizational innovations— and their diffusion— and, finally, to experience-based
learning. Even with better indicators in these fields, we cannot expect to get very
simple and clear conclusions from quantitative evaluation exercises.
Therefore case studies bringing together qualitative and quantitative information
and dialogue with policy practitioners will remain important sources o f insight
when designing policy. Richard Nelson has more than any other scholar developed
this approach (Nelson, Peck, and Kalachek 1967; Nelson 1982, 1984, 1988, 1993).

Box22.3 Innovation systems and innovation policy

Innovation systems is not an economic theory in the same sense as neoclassical or
evolutionary economics, but the concept integrates theoretical perspectives and em­
pirical insights based on several decades of research.
Innovation is seen as a cumulative process that is path-dependent and context-
dependent This is why innovation policy needs to build upon insight in a specific
context and why “best-practice” cannot be transplanted from one innovation system
to another. Innovation is also seen as an interactive process. The competence of single
innovating firms is important but so is the competence of suppliers, users, knowledge
institutions, and policy makers. The linkages and the quality of interaction is import­
ant for outcomes. This is why innovation policies that focus on subsidizing and
protecting suppliers of knowledge at best are incomplete—at worst they increase the
gap between technological opportunities and absorptive capacity. At least the same
attention needs to be given to users and to linkages.
Innovation systems may be seen as frameworks both for innovation and for compe­
tence building. Competence building involves learning and renews the skills and
heights necessary to innovate. Innovation processes are processes of joint production
where innovations and enhanced competence are the two major outputs. Learning
takes place in an interaction between people and organizations. The social climate
Includmg trust, power, and loyalty contributes to the outcome of learning processes.
Ш take into account the broader social framework
when the objective is to promote economic wealth creation.
6l8 B E N G T -A K E LU N D V A LL AND SUSA NA BORRAS

N o t least, in the era o f in n o v a tio n p olicy, w h ere in stitu tio n s m atter m o re th an ever, it
is d ifficu lt to see h o w q u an titative an alysis co u ld stan d alon e as the b asis fo r policy.

22.3*2 Evaluating STI Programs and Policies

W ith the gro w th o f S T I p o licies an d p ro g ra m s, p u b lic a u th o rities h ave been in creas­
in g ly interested in evalu atin g the effects an d im p acts o f p u b lic e xp e n d itu re in these
areas. E v alu a tio n is the system atic assessm ent o f p ro g ra m s o r related p u b lic e x p e n d ­
itures in term s o f h o w far th ey h ave attain ed th eir goals. E v a lu a tio n sh o u ld be
co n sid ered as an elem ent in a p o litical p rocess, nam ely, w h en p u b lic ad m in istra tio n s
try to elab orate co n clu sio n s an d lessons fro m past p e rfo rm an c e in o rd er to b eco m e
better in the fu tu re, o r to d ecide u p o n the fate o f the a c tiv ity in q u estio n . T h e
evalu ative exercise is ty p ic a lly co n d u cted b y extern al an d in d ep en d en t actors, w h o
use a ran ge o f m eth o d o lo g ies, in clu d in g the self-assessm en t o f th o se p erso n s in ­
v o lv ed in the im p le m e n ta tio n o f the p ro g ra m . T h ere are as m a n y e v a lu a tio n m eth ­
o d o lo gies as evalu ato rs, an d as m a n y p o lic y styles as p u b lic ad m in istra tio n s.
Several au th o rs em ph asize that evalu atin g S T I p o licies an d p ro g ra m s is p a rtic u ­
la rly d ifficu lt given the w id e effects th ro u g h o u t the system . It has been argu ed , for
exam p le, that m icro -le v el evalu atio n s (p ro g ra m -sp e cific) are m o re reliable than
m acro -level evalu atio n s w h ere issues such as w h eth er a sp ecific p ro g ra m o r p o licy
enh an ced the co m p etitiven ess o f an e c o n o m y are alm o st im p o ssib le to d eterm in e
(L u u k k o n e n 1998). Sim ilarly, it has been p o in ted o u t that m o st p ro g ra m s have
im p o rta n t effects b ey o n d their strict in itial goals. T h is is the case w h en S T I policies

Box 22.4 Normative principles for design of STI policies

Robustness: Decisions and social structures should withstand the occurrence o f differ­
ent future scenarios.
Flexibility. In the occurrence o f sudden socio-economic change institutions must be
able to change direction rapidly
Internal diversity Structurally dissimilar characteristics must be built in to allow
survival if the selection environment changes.
External diversity Variety o f links to different kinds o f agents will help adaptation when
change in the environment arises.
Window of opportunity: Attention to timing and sequence in face o f path-dependent
systemic context.
Incremental approach: The whole can be changed only through the cumulative impact
o f small steps.
Experimentation and prudence, new policy ideas should be submitted to trial in
localized contexts before full deployment.
Source Sandro Mendon^a
 S C I E N C E , T E C H N O L O G Y , A ND I N N O V A T I O N P O L I C Y 6l9

have helped the creation o f standards, have induced more risk-taking attitudes o f
innovators, have fostered long-term rather than short-term strategies in firms’
research, and have enhanced the acquisition o f new skills and knowledge (Peterson
and Sharp 1998). This has also been framed as the problem o f attribution. Closely
related is the problem o f synchronizing the evaluators’ time horizon with the
political time horizon o f the political consumers o f their work. Big programs’ effects
may not be realized for years if not decades.

22.4 STI P o l ic y in t h e US, Ja pan ,

a n d E uro pe

We have used O ECD documents to organize a stylized presentation o f the evolution

of science, technology and innovation policy. But while what is discussed at OECD is
one thing, what national governments actually do to affect science, technology and
innovation is quite another. No country has focused on just one o f the kind o f policies
described above. All countries have combined elements o f science, technology and
innovation policy. But the m ix and the policy design has been quite different between
countries. Here we will try to capture the most basic characteristics for respectively
the US, Japan, and Europe— understood as the big countries in Europe and as the
European Union. Finally, the future challenges for each o f them will be discussed.

22.4.1 Public Mission Technology Policy in the US

We have already m entioned the crucial Vannevar Bush report published in 1945
entitled “ Science: The Endless Frontier.” The neglect o f one o f its major recommen­
dations had a very im portant effect upon the evolution o f technology policy. Bush
recommended the establishment o f a coordinating authority at the national level,
the National Research Foundation ” But it took another five years before the
National Science Foundation” was set up. In the meantime, different sector author­
ities in charge o f contracting out research on nuclear power, defense, space, and
health had already established am bitious research programs o f their own and the
total resources o f N SF never came close to the budgets o f these sector-oriented
activities (M owery 1994).
Technology policy in the U S may therefore be seen as being organized in parallel
industrial complexes— vertically organized complex networks crossing disciplines,
technologies as well as industrial sectors, but with user interests in the public and
620 BENGT-AKE LUNDVALL AND SUSANA BORRAS

p rivate sector d riv in g an d d efin in g the policy. T h ese co m p lexes w ere o p e ra ted w ith
little c o o rd in a tio n betw een them and they w en t all the w a y fro m p ro c u re m e n t o f
sp ecific tech nical system s an d c o m p o n e n ts to the su p p o rt o f b asic research and
research train in g . In particu lar, the defense b u d g et w as u sed fo r p ro m o tin g activities
w ith little d irect co n n ectio n to sh o rt-te rm m ilita ry needs. R esearch in c o m p u te r
science an d softw are got su b stan tial su p p o rt fro m this b u d g et an d p ro d u c e d quite
gen eric kn o w led ge, in a d d itio n to p ro d u c in g sh o rt-te rm so lu tio n s to m ilita ry needs
(L an glo is and M o w e ry 1996).
O n e m a jo r issue in relatio n to the p u b lic m issio n te c h n o lo g y p o lic y is the lack o f
c o o rd in a tio n . O n e negative aspect is the resu ltin g b ia s in the d ire c tio n o f m ilita ry
an d space exp en d itu re. It seem s as i f it is m u ch easier to m o b ilize ta xp ayers' m o n ey
an d d evelo p m en t efforts to go to the m o o n than it is to solve the p ro b le m o f the
gh etto (N elso n 1977). In p art, this reflects the presence o f w e ll-o rg an iz e d lo b b ies and
interest g ro u p s a m o n g p rivate sector b en eficiaries; it also reflects the sh eer difficulty,
n o ted b y N elson , o f "'so lvin g5' such in tractab le p ro b lem s as u rb an p o v e rty o r p rim a ry
ed u catio n al ach ievem en t. A n o th e r p ro b lem is that the calcu latio n o f exp ected costs
an d benefits o f sp ecific p ro jects m ay either be co m p letely n eglected becau se n ation al
p rid e is at stake o r be system atically biased d o w n w ard s to m ak e a p ro ject lo o k m ore
attractive. A th ird w eakness m a y be a o n e-sid ed fo cu s o n the d ev elo p m en t o f science-
based in d u stries an d tech n o logies an d a general a ssu m p tio n that science and
tech n o lo g y is a q u ick fix fo r all k in d s o f p ro b lem s. M o re m u n d a n e in d u stries aim in g
at co n su m ers, an d p ro b lem s w h ere the so lu tio n s are less g la m o ro u s in tech n o logical
term s, m a y be neglected.
B u t the U S -ty p e n atio n al in n o v a tio n system certain ly h as its stro n g p o in ts b eyon d
its sheer scale advantages. T h e fact that several o f the agen cies are prep ared to finance
m o re o r less generic research an d research tra in in g w h ile still h a v in g the use o f
research in m in d ten ds to o verco m e the bias tow ard the su p p ly side. It tends to
establish a "ch a in -lin k e d m o d el o f in n o v a tio n 5w ith stro n g feed b ack elem ents (K line
an d R o sen b erg 1986). O ne stro n g elem ent is d iversity. T h e fact that d ifferen t agencies
"c o m p e te 55 in fu n d in g g o o d research m a y b e b en eficial in su p p o rtin g the d iv ersity o f
research efforts. T h e coexisten ce o f b ig p rivate fo u n d a tio n s that c o m m it resources
to research team s o n the basis o f their track record also co n trib u tes to diversity.
In these respects cu rren t E u ro p ean centralistic in itiatives in research p o lic y in
co n n ectio n w ith the sixth fra m e w o rk p ro g ra m and the E u ro p ea n R esearch Era
co u ld learn fro m the U S, w h ich ben efits not o n ly fro m scale b u t p erh a p s even more
fro m diversity.

22.4.2 Sector Technology Policy in Japan

W hile the in tern atio n al co m p etitiven ess o f U S firm s h as b een a co n cern in U S
tech n o lo gy policy, it has, w ith few exception s, been лп i m o b rit r i t h.л f I - . .. а. гиі. л. г
 S C I E N C E , T E C H N O L O G Y , A N D I N N O V A T I O N POLI CY 621

objective. This is different when we look at technology policy in Japan. More than
perhaps any other market econom y Japan has made use o f an explicit national policy
to promote specific sectors and industries with the ultimate aim o f stimulating
economic growth and competitiveness. According to Freeman (1987), the radical
character of the policy had to do with the fact that since the mid-1950s policy ended
up being designed by experts with an engineering background in the Ministry of
Technology (MITI) who were much less concerned with “comparative advantage”
than the economists in the Bank o f Japan.
But the M ITI has not acted alone. In the area o f telecommunications, NTT__a
public company in m onopolistic control o f telecommunication services— has
played an important role in coordinating technology development efforts in major
electronic firms such as Hitachi and N EC. The policies pursued have not involved
massive subsidies and, actually, the public sector has been much less involved in
financing R&D in the private sector than has been the case in the US (Nelson 1984).
The strategic prom otion o f the car industry, consumer electronics, and “mega-
tronics” typically com bined different policy instruments such as subsidies to re­
search and development o f generic technologies with elements o f “ infant-industry”
protection. Bringing together competing firms in consortia aiming at solving
common problems has been an im portant role for M ITI. This has been done on
the basis o f attempts to map new trends in technology and markets through, for
instance, technology forecasting.
One interesting aspect o f M IT Is technology policy has been that it did not
focus exclusively on high technology sectors. For instance, consortia initiated by
MITI aiming at prom oting the modernization o f textile and clothing industries
brought together firm s producing textiles and textile machinery with electronics
firms.

22.4.3 European STI Policies: From the Promotion of National

Champions to EU Framework Programs
Europe is certainly a m uch m ore diverse region than the US and Japan. University
systems are different in the U K , France, and Germany. The role of engineering in
industry and corporate governance differ as well. Bringing less rich countries such as
Portugal and Greece into the picture makes it even less homogeneous in terms o f
R&D efforts and innovation styles. Therefore, to treat Europe as one legion and
compare it with the U S and Japan in terms of, for instance, R8cD effort or patenting,
without being explicit about the dispersion o f the variable, is not helpful.
The formation o f a com m on European approach to science and technology is still
evolving. Figure 22.2 indicates the elements o f the European scientific and techno
logical architecture, within and outside the functional-administrative borders oft e
EU. As can be seen, some international research organizations were established m
622 BENGT-AKE LUNDVALL AND SUSANA BORRAS

the 1950s, w h ereas the m a in th ru st cam e in the 1970s. T h is fig u re d oes n o t in clu d e the
E u ro p ean stan d ard izatio n b o d ies (E T S I, C E N , an d C E N E L E C 2) an d the E u ro p ean
agen cies gran tin g in tellectu al p ro p e rty righ ts (like E P O , O H IM , an d C o m m u n ity
P lan t V ariety O ffice3), w h ich h ave also played im p o rta n t roles in p ro m o tin g scien ­
tific, tech n o lo gical, an d in n o vative synergies.
C E R N is the oldest an d o n e o f the m o st successful o rg a n iz a tio n s, an d is dedicated
to n u clear research. Its u n d en iab le scien tific success co n trasts sh a rp ly w ith the
tro u b les o f its T w in ,” the JR C (Jo in t R esearch C e n te r), u n d er the E u ro p e a n C o m ­
m u n ities an d also d ed icated to n u clear e n erg y research. D isp u te s b etw een France
and G e rm a n y ab o u t n u clear reacto r d esigns in the 1960s w eak en ed this in stitu tion ,
u n til the lau n ch o f the Jo in t E u ro p ea n T orus (JE T ) co n cern ed w ith n u clear fu sio n
(G u zzetti 1995). O th er E u ro p ea n b u t n o n -E U scien tific an d tech n o lo g ica l o rg an iza­
tio n s em erged in the 1960s an d 1970s, a m o n g th em the E u ro p ea n S o u th ern L a b o ra ­
to r y (A stro n o m y ), the E u ro p ean M o lecu la r B io lo g y L a b o ra to ry (E M B L ), the
E u ro p ea n Space A g en cy (E S A ), the E u ro p ea n Science F o u n d a tio n (E S F ), gathering
n atio n al research co u n cils, and A irb u s, a p u b lic -p riv a te en terp rise w ith heavy
in vestm en t in a v ia tio n tech n ology. A ll these b o d ies are in te rn a tio n a lly con stituted
an d d irected tow ard o n e sp ecific scie n tific -te c h n o lo g ic a l field.
In 1971, C O S T (C o o p e ra tio n in Science an d T ech n o logy) w as estab lish ed as an
in tergo vern m en tal p ro g ra m . T h e n o velty w ith C O S T w as that it covered several

AIRBUS
ESA (1970)
European Union (1975) EADS +
25 states 15 states British
aerospace
RTD Framework
Program

EUREKA
JET JRC (1985)
33 states + EU

ESF
(1974)
76 institutions
29 states

Figure 22.2 The scientific and technological architecture of Europe, 2001

Source: Borras (2003)
 SCIENCE* TECHNOLOGY,
AND IN N OVATION POLICY 623

scientific areas* and envisaged a very flexible form o f cooperation Later on

EUREKA developed this concept further, and became a successful tool o f European
collaboration outside the EU, m ainly due to the high degree o f participation by firms
and to its market orientation.
In the 1980s attempts to build strength in ICT technology were still national in
France, the U K and Germ any and the basic strategy was to promote national
champions. The success was limited and this was one important reason why the
ESPRIT program was developed in the early 1980s, under the realm o f the EU.
ESPRIT was inspired b y Japanese technology policy style and its actual design
came out o f intense consultation o f the EU commission with the top leaders o f the
fifteen biggest European IC T firms (Peterson 1991).
The first EU fram ework program was established in 1984 as a multiannual*
multisectoral (covers several scientific fields) and multinational program (grants
funding to projects submitted by researchers o f at least three EU countries). The
most recent fram ework program (the sixth) is very ambitious in terms o f promoting
European-wide networks o f excellence* as a means o f creating a European Research
Area* reducing national barriers. Furthermore, in 2002 European ministers o f
science set as their am bitious goal that the share o f R& D should reach 3 per cent o f
GNP in member states (2 per cent private and 1 percent public as the rale o f thumb).
This should be seen in the light o f declarations made by prime ministers at the
Lisbon Summit held in 2000* that Europe should become the world s most competi­
tive knowledge econom y by 20m, with social cohesion as a twin goal.
In general* it is a problem that the European construction calls for dramatic
declarations to build support for European ST 1 policy. It contrasts with Japan and
the US where less is said and m ore is done. The share o f total R8cD expenditures that
the commission distributes to member states is still quite small. Another problem is
that the Brussels adm inistration tends to take on more than it can master in terms o f
defining research programs* evaluating applications and administering projects; the
result being that it is quite dem anding to be a project coordinator for EU projects.
Last* but not least, another m ajor problem is that the general idea behind the
European Research Area is scale, rationalization* coordination, and concentration
of effort. Much less weight is given to the dimensions o f diversity and competition*
two key elements for successful innovation systems (Lundvail and Borras 1998;
Borras 2003).
The framework program s have been used as instruments to promote European
integration and there is no doubt that the programs have had an enormous effect in
terms o f building research collaboration o f a lasting kind across Europe. And in spite
of administrative problem s the m oney is still regarded as attractive. This is especially
the case in countries that have very little alternative free funding such as France and
in countries where the European efforts have been used to reduce national efforts
such as the UK.
624 B ENGT - AKE L DNDVAL L AND S US ANA BORRAs

22.4.4 The Challenges for the US, Japan, and Europe

T h ere is a ten d en cy to g lo rify the in n o v a tio n system an d the in n o v a tio n p o lic y o f the
c o u n try d o in g especially w ell in in tern atio n al c o m p e titio n . In the b e g in n in g o f the
1990s Ja p a n w as seen as a m o d e l and E u ro p e as a secon d -b est altern ative, w h ile the U S
w as seen as a th reaten ed b u t b ig p o w e r in the field o f tech n o logy. T o d ay the ro les h ave
been reversed, T h e truth o f the m atter is that each system h as its o w n stren gth s and
w eaknesses and that these d o n o t go aw ay in p e rio d s o f rap id e c o n o m ic grow th .
T h e U S e c o n o m y co u ld p ro b a b ly d o even b etter i f it in vested m o re b ro a d ly an d
m o re eq u ally in h u m a n resources. Fo r in stan ce it is rem ark ab le that th e u n w illin g ­
ness o f citizens in C a lifo rn ia to p a y taxes h as resulted in stu d en ts in sch o o ls in S ilico n
V alley h avin g m o re lim ited access to co m p u ters th an stu d en ts h ave in the N o rd ic
co u n tries. T h e w o rk o rg a n iz a tio n in in d u stry w ith a stro n g d iv isio n b etw een train ed
an d u n skilled w o rk ers— an d the resu ltin g in co m e gaps— p ro b a b ly m ak es the d iffu ­
sion an d use o f n ew tech n o logies less efficien t than w o u ld a m o re even d istrib u tio n
o f co m peten ce. W e believe that fo r the U S these are so m e o f the m a jo r ch allenges in
the b ro ad field o f in n o v a tio n policy.
O n e m a jo r w eakness o f the Jap an ese system o f in n o v a tio n has to d o w ith science
p olicy. T h e u n iversities d o not h ave the sam e in cen tives an d tra d itio n s fo r p ro m o t­
in g h igh q u ality research. In E u ro p e, the m a jo r ch allenge m ig h t be callin g fo r a
c o m b in atio n o f science an d tech n o lo g y policy. T h e so -called E u ro p e a n p a ra d o x —
that E u ro p e is d o in g w ell in science w h ile b ein g w eak in te c h n o lo g y — m ig h t be
so m ew h at o f f the m a rk due to its h igh in tra -E U d isp arities (P avitt 1998). Rather,
E u ro p e seem s to be w eak b o th in science an d tech n o lo g y in so m e o f the m o st rap id ly
g ro w in g fields an d m arkets— n o t least w ith in b io tec h n o lo g y and p h arm aceu ticals.

22.4.5 The Triad Game

A s lo n g as the C o ld W ar was a reality, the existence o f a c o m m o n e n em y su p p o rte d
scien tific co llab o ra tio n betw een U S , E u ro p e, and Jap an . G iv e n the cu rren t p re d o m ­
in an t un ilateral a p p ro a ch o f the U S an d the attem p ts to b u ild a co h eren t E u ro p ean
research area there are g ro w in g risks for co n flicts w ith in the triad regard in g ST I
policy. T h ere is especially a risk that access to k n o w led ge is used as a p o litica l
in stru m en t as it w as used b y E n g lan d fo r 500 years again st the C a th o lic states in
E u ro p e. Since m o re an d m o re tech n o logies can be argu ed to have m ilita ry relevance
this could b ecom e a m a jo r p ro b lem fo r the glo bal k n o w led ge society.
T h e g ro w in g em p h asis 011 the p ro tectio n o f intellectu al p ro p e rty righ ts in relation
to trade and exten sio n to softw are an d services m ay establish great obstacles fo r the
catch in g-u p co u n tries and red u ce their po ssib ilities o f p u rsu in g the successful
strategies o f Jap an and N IC s. A n interesting q u estion , in this co n text, is w h eth er
 S C I E N C E , T E C H N O L O G Y , AND
I N N O VATION po lic y 625

Europe and Japan are prepared to compete „ ,,h the US in becomtng centers to,
gtobal interaction wtth the res,. The more positive scenario » „ uld he one" h ”
dte three poles used therr capaatres promote ,h , drfhtsion and use of k n o„led «
,0 less pnvtleged parts o f the world. In anp case, the eaten, „ hich there i, 'ro
globalization o f knowledge production and diffusion, and the different form i
assumes, ts a topic that has started be analyzed (Arch,bug, and Michie , J
Georghiou 1998), but deserves m ore attention. 997>

22*5 C o n c lu sio n s

It should be clear from what has been said that the most pressing issues on the policy
agenda are specific to each national system. Even so, there are some issues that are
common for all countries, and these represent new research challenges for scholars
in this field. We have presented a stylized sequence, beginning with science and
ending with innovation policy. In the most recent debates about the learning
economy and the knowledge-based society we can see the contours o f a new policy
that we might call “ knowledge policy.” 4 It recognizes that innovation and compe­
tence building involve m any different sources o f knowledge and that innovation
itself is a learning process. This raises the need for new analytical efforts and for
rethinking the organization and implementation o f policy in several respects.
It is increasingly im portant to understand better the connection between science
and technology on the one hand and economic performance on the other. The rise
and fall o f the new econom y demonstrates that assumptions about simple and direct
connections are problematic. Between the new technologies and the performance o f
the economy the organizational characteristics o f innovation systems and firms
including “ slippery” elements such as “ social capital” affect the impact. This is an
issue that remains understudied. In terms o f public policy there is a need for
innovative thinking about how governments can support the diffusion o f good
and sustainable practices in cooperation with management and employees. In
terms of research opportunities, this links with the importance and need to devote
more analytical efforts exam ining how technical innovation interacts with organiza­
tional change. The academic traditions o f business organization and o f innovation
systems’ research have to come closer to each other in order to answer questions
regarding how organizational change affects innovation processes in the economy.
A second issue is about aggregate demand. In a period o f growing fears for
deflation and with little room left for expansionary monetary policy it might be
useful to reconsider what Keynesian policy could mean in a knowledge-based
6 l6 B ENGT - AKE L UNDVALL AND S US ANA BORRAS

e c o n o m y ” E stab lish in g large-scale te c h n o lo g ic a l-m ilita ry p ro g ra m s (like th e Star

W ars p ro g ra m s u n d er the R eag an and B u sh ad m in istra tio n s) m a y p e rh a p s be seen as
the m o d e m v ersio n o f b u ild in g “ p y ra m id s” ; h ow ever, o th er m o re so c ia lly o rien ted
an d lo w -scale o p tio n s co u ld also be co n sid ered . E stab lish in g firm - o r secto r-sp ecific
fu n d s th ro u gh tax exem p tio n s fo r u p g ra d in g the skills o f all catego ries o f em ployees
an d fo r m a k in g extra d evelo p m en t efforts in p e rio d s o f lo w e c o n o m ic a c tiv ity could
be o n e o p tio n .
A th ird c o m m o n co n cern in the era o f in n o v a tio n p o lic y is h o w to co o rd in ate
p o licies affectin g in n o va tio n . T h e p re va ilin g in stitu tio n a l set-u p m ean s that m in is­
tries o f fin an ce are the o n ly agencies tak in g o n a resp o n sib ility fo r c o o rd in a tin g the
m a n y specialized area p o licies. A re a-sp ecific m in istries, o n the o th er h an d , ten d to
id e n tify the interests o f th eir o w n “ c u sto m ers” an d take less in terest in global
o bjectives o f society. It co u ld be d ecid ed to establish n ew typ es o f in stitu tio n s such
as cro ss-secto r and in terd iscip lin a ry C ouncils on In n o vatio n a n d C om petence B u ild ­
ing at the su b n a tio n al an d n atio n al level (in F in lan d the p rim e m in ister is ch airm an
o f a N a tio n a l C o u n c il o f Science and T ech n o lo gy). T h is sh o u ld be co m p lem en ted b y
the m u ch -n eed ed research efforts to w ard s d ev elo p in g so p h isticated m easu rin g
m eth o d s ab o u t the in n o v a tio n system trends and ab o u t the im p a c t o f S T I p o licy
o n it. M o re ad van ced in n o v a tio n in d icato rs w o u ld be a cru cial in p u t fo r su ch holistic
p ersp ective o f p u b lic au th orities.
In 19 6 1, the O E C D exp ert g ro u p — Freem an , Sven n ilsso n , an d o th ers— presented
a k in d o f m a n u a l fo r h o w to d esign science p o lic y in su ch a w a y that it becam e
integrated w ith eco n o m ic p o lic y an d p ro vid e d a real im p a ct o n e c o n o m ic grow th.
P erh aps a su itab le clo sin g o f this ch ap ter m ig h t be to p ro d u c e a sim ila r set o f
reco m m en d atio n s fo r d esign in g n atio n al in n o v a tio n p olicy? B u t the m essage here
is that there is no w ay to d esign an effective in n o vatio n p o lic y w ith o u t an alyzin g the
d o m estic in n o v a tio n system , in clu d in g the w ay it p ro d u ces an d rep ro d u ces k n o w ­
ledge an d co m p eten ce, an d c o m p a rin g it w ith others. T h e stage o f d evelo p m en t and
the size o f the respective e co n o m y w ill affect the resu ltin g p lan o f actio n . In sm all
co u n tries an d d ev elo p in g co u n tries the stru ctu res an d in stitu tio n s that affect ab ­
so rp tio n and efficien t use o f tech n o lo g y are m o re im p o rta n t to u n d erstan d and act
u p o n than those p ro m o tin g the p ro d u c tio n o f the tech n o lo g ies at the fro n t. Big
co u n tries w ill necessarily be m o re fo cu sed on the p ro d u c tio n o f the n ew tech n o lo ­
gies, but they to o w o u ld have m u ch to g ain fro m tak in g in to acco u n t the a b so rp tio n
an d efficien t use o f in n o vatio n s an d n ew know ledge.

N otes

l Ergas has suggested the distin ction betw een “ m issio n -o rien ted " and “ d iffu sio n -o rie n te d "
policy designs, based respecti vely on the m assive su p p o rt to a sm all n u m b er o f scien ce-
tech nology fields, and on su p p o rtin g the scientific-technological in frastru ctu re (Ergas
 S C I E N C E , T E C H N O L O G Y , AND I NN O V A T I O N POLICY 627

1987), This analytical distinction has also been used when comparing policy styles in
science and technology policies across countries. (It has been argued that the Japanese and
French policies are mission-oriented, whereas the German policy has been depicted as an
example o f a diffusion-oriented policy style.) Again it is important to note that what is
referred to is ideal types and that simple groupings o f this kind might miss some o f the
most important complexities in the set-up o f national systems o f innovation.
2. ETSI: European Telecommunications Standards Institute; CEN: Comite Europeen de
normalisation; and CEN ELEC : Com ite Europeen de Normalisation Electrotechnique.
These three bodies have divided their standardization activities by technological sectors.
3. EPO: European Patent Office; OHIM ; Office for the Harmonization o f the Internal
Market; dedicated respectively to patents, and to trademarks and other intellectual
property rights.
4. The concept o f knowledge policy has evolved in connection with European policy making
and as a follow-up to the Lisbon ministerial meeting 2000. One o f the architects behind
the lisbon-strategy, M aria Rodrigues, defines knowledge policies as 'policies aimed at
fostering and shaping the transition to a knowledge-based society” (Conce^aio, Heitor,
and Lundvall 2003: p. xx).

R e f e r e n c e s

Andreasen, L., C oriat , B,, D en H artog , R, and K aplinsky , R. (eds.) (1995), Europe's N ext
Step— O rganisational In n ovatio n , Com petition a n d Em ploym ent , London: Frank Cass.
Archibugi, D., and M ich ie , J. (1997), Technology Globalisation a n d Econom ic Performance,
Cambridge: Cambridge University Press.
----- and L u n d v a ll , B.-A. (eds.) (2000), The G lobalizing Learning Econom y: M ajo r Socio­
economic Trends a n d E uropean In n ovation Policy >Oxford: Oxford University Press.
A ro cen a , R., and S u t z , J. (20 0 0 ), “ Interactive Learning Places and Development Policies in
Latin America,” D RU ID Working Paper no. 13.
Arrow, K. J. (1971), “ Political and Economic Evaluation o f Social Effects and Externalities,
in M. Intrilligator (ed.), Frontiers o f Q uantitative Econom ics , North Holland, 3-23.
Arundel, A., and Soete , L. (eds.) (1993), A n Integrated Approach to European Innovation and
Technology D iffusion Policy: A M aastrich t M em o ran d u m , Brussels. CEC: Sprint.
Be r n a l , J. D. (1939), The Social Function o f Science , London: Routledge & Kegan Paul
*Biegelbauer, P., and B orras, S. (eds.) (2003), Innovation Policies in Europe a n d the US. The
New Agenda , Aldershot: Ashgate,
Borras , S. (2003), The In n ovatio n Policy o f the EU, Cheltenham: Edward Elgar.
C a n t n e r , U., and P y k a , A. (2001), “ Classifying Technology Policy from an Evolutionary
Perspective” Research P olicy 30: 759-75**
C o lem a n , J. (1988) “ Social Capital in the Creation o f Human Capital,” A m erican Jou rn al o f
Sociology 94 (Suppl.): 95-120.
C onceicao , R , H e i t o r , M ., and L u n d v a ll , В.-A. (eds.) (2003), Innovation , Competence
Building a n d Social Cohesion in E urope: Towards a L ea rn in g Society >Cheltenham. Edward
Elgar.

* Asterisked items are suggestions for further reading.

628 BENGT-AKE LUNDVALL AND SUSANA BORRAs

C ozzens , S„ H ealey , A., and Z iman ,

P„ R ip , J. (eds.) (1990), The Research System in
Transition) Dordrecht: Kluwer Academic.
D a l u m , B., J o h n so n , B., and L u n d v a l l , B .-A . (1992), “ Public Policy in the Learning
Society/5 in Lundvall 1992: 296-317.
*D odgson , M., and B e s s a n t , J. (1996), Effective Innovation P o lic y : A N e w A p p ro a c h ) London:
International Thomson.
D ert o u tzo s , M. L , L e st e r R. K., and S olow , R. M. (1989), M a d e in A m e r ic a , Cambridge,
Mass.: The MIT-Press.
E dquist , C. (ed.) (1997), Systems o f Innovation: Technologies) Institutions and Organizations)
L o n d o n : Pinter.
------- and H o m m en , L. (1998), “ G o vern m en t T ech n olo gy P ro cu rem en t an d In n ovation
T h eo ry/5 T S E R -p u b lic a t io n , Lin ko p in g: U n iversity o f L in k o p in g, D ep artm en t o f Social
and T echnological C hange.
------------- and T s ip o u r i , L. (eds.) (2000), P u b lic T e c h n o lo g y P r o c u r e m e n t a n d In n o v a tio n .
L o n d o n : K luw er A cadem ic.
E r g a s , H . (1987), “ T h e Im p o rtan ce o f T echn ology Policy/5 in P. D asgu p ta an d P. Stonem an
(eds.), E c o n o m ic P o lic y a n d T ech n o lo g ica l P e rfo rm a n c e ) C am b rid ge: C am b rid ge U n iversity
Press.
F a g e r b e r g , J. (1995), “ Is there a Large C o u n try A dvan tage in H igh -T ech ?55 N U P I W orking
Paper no. 526, O slo: N U P I.
-------(2002), Techn ology, G ro w th a n d C o m p e titiv e n e s s , C h elten h am : E d w ard Elgar.
-------and V e r s p a g e n , B. (1996), “ H ead in g fo r D ivergence? R egio n al G ro w th in Europe
R econ sid ered /5 J o u r n a l o f C o m m o n M a r k e t S tu d ie s , N o. 34.
---------------and M a r jo l e in , C . (1997), “ Technology, G ro w th an d U n em p lo ym en t across
E u ro p ean R egio n s/5 R e g io n a l S tu d ie s 31(5): 457-66.
F r a n sm a n , M. (1995), J a p a n s C o m p u t e r a n d C o m m u n ic a tio n s In d u s try , O xfo rd : O xford
U n iversity Press,
F r e e m a n , C . (1982), “ Technological Infrastru ctu re an d In tern ation al C o m p etitiven ess/5
D raft p ap er subm itted to the O E C D A d h o c-gro u p o n Science, T echn ology an d C o m p e ti­
tiveness, A ugu st 1982, m im eo,
------- (1987)5 T ech n o lo g y a n d E c o n o m ic P e rfo rm a n c e : L esso n s fr o m Ja p a n , L o n d o n : Pinter.
* -------(1992), T h e E c o n o m ic s o f H o p e, Lo n d o n : Pinter.
-------(i995)> “ T h e N ation al In n o vatio n System s in H istorical P erspective/5 C a m b r id g e Jo u r ­
n a l o f E co n o m ic s 19 (1): 5-24.
-------and L u n d v a l l , B .-A . (eds.) (1988), S m a ll C o u n trie s F a c in g th e T e c h n o lo g ic a l R e v o lu tio n ,
L o n d o n : Pinter.
-------and P e r e z , C. (1988), “ Stru ctu ral C rises o f A d ju stm en t: B u sin ess Cycles and
Investm ent b eh avio u r/5 in G. D osi et a l , T ech n o lo g y a n d E c o n o m ic T h eo ry , L o n d o n : Pinter,
38-66.
G e o r g h i o u , L. (1998), “ G lobal C o o p eratio n in R esearch/5 R ese a rch P o lic y 27: 6 11—26.
G ib b o n s , M „ et al. (1994), T h e N e w P ro d u c tio n o f K n o w le d g e , L o n d o n : Sage.
G ittelm an , M ., and Ko gut, B. (2003), “ D oes G o o d Science Lead to V aluable K now ledge?
B io tech n o lo gy Firm s and the E vo lu tio n ary Logic o f C itatio n P attern s/5 M a n a g e m e n t
S c ie n c e 49(4): 366-82.
G uzzetti , L (1995)» A B r i e f H is to r y o f E u ro p e a n U n io n R ese a rch P o licy, L u xem b o u rg: O ffice
for O fficial P ublications.
 S C I E N C E , T E C H N O L O G Y , AND INNO VATI
ON POLICY 629

Hirschman , A. О (1969) >The Strategy of Economic Development, New Heaven; Yale Univer­
sity Press.
Jasanoff, S. (ed.) (i997)> Comparative Science and Technology Polity Cheltenham; Edward
Elgar.
*■---- (ed-) (2002), Handbook of Science and Technology Studies, 2nd edn., London: Sage.
Katzenstein, P, J. (1985), Small States in World Markets. Industrial Policy in Europe, New
York: Cornell University Press.
K u n e , S. J., and R o s e n b e r g , N. (1986), “An Overview of Innovation,” in R. Landau and
N. Rosenberg (eds.), The Positive Sum Game, Washington, DC: National Academy Press,
275-305.
Kristensen , P. H., and S t a n k ie w ic z , R. (eds.) (1982), Technology Policy and Industrial
Development in Scandinavia, Lund: Research Policy Unit.
K rull , W., and M e y e r - K r a m e r , E (eds.) (1996), Science and Technology in Germany
London; Cartemills.
K u enh , T. J., and P o r t e r , A. L. (eds.) (1981), Science, Technology and National Policy
London: Cornell University Press.
K u t zn e t s , S, (1960), “Economic Growth of Small Nations,” in E.A.G. Robinson (ed.),
Economic Consequences of the Size of Nations, Proceedings of a Conference held by the
International Economic Association, London: Macmillan.
La n g lo is , R., and M o w e r y , D. C. (1996), “The Federal Role in the Development of
American Computer Software Industry: An Assessment,” in D. C. Mowery (ed.), The
International Computer Software Industry, New York: Oxford University Press.
L ist , F. (1841), Das Nationale System der Politischen Okonomie, Basel: Kyklos (trans. and pub.
as The National о
System of Political Economy by Longmans, Green and Co., London, 1841).
Lu n d v a ll , B.-A. (1985), Product Innovation and User-Producer Interaction^ Aalborg: Aalborg
University Press.
---- (1988), “Innovation as an Interactive Process: From User-Producer Interaction to the
National Innovation Systems,” in G. Dosi, C. Freeman, R. R, Nelson, G. Silverberg, and
L, Soete (eds.), Technology and Economic Theory London: Pinter, 349-69.
-----(ed.) (1992), National Innovation Systems: Towards a Theory of Innovation and Inter-
active Learning, London: Pinter.
---- (1996), “The Social Dimension of the Learning Economy,” DRUID Working Paper, No.
1.
—— (2002), Innovation, Growth and Social Cohesion: The Danish Modeh Cheltenham:
Edward Elgar.
---- and J o h n s o n , B. (1994), “The Learning Economy,” Journal of Industry Studies 1(2):
23-42.
*— —and B o r r a s , S. (1998), The Globalising Learning Economy Implications for Innovation
Policy., Brussels: European Commission.
-----and T o m l in s o n , M. (2001), “Policy Learning by Benchmarking National Systems of
Competence Building and Innovation,” in G. P. Sweeney (ed.) Innovation, Economic
Progress and Quality of Life, Cheltenham: Edward Elgar.
----- Mendon^a, S. (2003), ‘Politica de Ciencia, Technologia e Inova^ao: Prindpios,
tendencias e questoes,’ mimeo, ISCTE University.
Luukkonen, T. (1998), “The Difficulties in Assessing the Impact of EU Framework
Programmes,” Research Policy 27(6): 599-610.
630 BEN G T-A KE LUND VA LL AND SUSA NA BORRAS

M a r t in , B ., and S a l t e r , A, (1996), T h e R e la tio n s h ip B e t w e e n P u b lic ly F u n d e d B a s ic R ese a rch

a n d E c o n o m ic P e rfo rm a n c e . A S P R U R e v ie w (Ju ly 1996). R ep o rt p rep ared fo r H M Treasury,
#M e t c a l f e , S. (1995), “ The E co n o m ic Fo u n d atio n s o f T ech n olo gy P olicy: E q u ilib riu m and
E vo lu tio n ary Perspectives,” in P. Sto n em an (ed), H a n d b o o k o f th e E c o n o m ic s o f In n o v a t io n
a n d T e ch n o lo g ica l C h a n g e , O xford : O xfo rd U n iversity Press, 4 0 9 -5 12 .
-------and G eo rg h io u , L. (1998), “ E q u ilib riu m and E v o lu tio n ary Fo u n d atio n s o f T echnology
Policy,” S T I R e v ie w 22: 7 5 -10 0 .
*M o w er y , D. C . (1994), “ U S P ost-w ar Techn ology P o licy an d the C reatio n o f n ew In d u s­
tries ” in O E C D , C re a tiv ity , In n o v a t io n a n d Jo b - C r e a t io n , Paris: O E C D .
-------and S a m p a t , B. N. (2001), “ U n iversity Patents an d Patent P o licy D ebates in the U SA ,
19 25-19 8 0 ,” In d u s t r ia l a n d C o r p o r a te C h a n g e 10: 7 8 1-8 14 .
N a t io n a l A c a d em y of E n g in e e r in g (1993), P r o s p e r in g in a G lo b a l E c o n o m y , W ashington,
D C : N ation al A cad em y Press.
*N e l $on , R. (1959), “ The Sim p le E co n o m ics o f Basic E co n o m ic Research,” J o u r n a l o f
P o litic a l E c o n o m y 6 j : 323-48 .
-------(1977), T h e M o o n a n d the G h etto : A n E s s a y on P u b lic P o lic y A n a ly s is , N ew York: N orto n .
-------(ed.) (1982), G o v e r n m e n t a n d T e c h n ic a l P ro g ress; A C ross I n d u s t r y A n a ly s is , N ew York:
P ergam on Press.
-------(1984), H ig h -T e c h n o lo g y P o lic ie s: A F iv e - N a t io n C o m p a r is o n , W ashin gton D C : A m e ri­
can En terprise Institute.
-------(1988), “ Institution s Su p p o rtin g Technical C h an ge in the U n ited States,” in G . D osi,
C. Freem an , R. R, N elson , G . Silverberg, an d L. Soete (eds.), T e ch n o lo g y a n d E c o n o m ic
T h e o r y L o n d o n : Pinter, 3 12 -2 9 .
-------(ed.) (1993), N a t io n a l In n o v a t io n S ystem s: A C o m p a r a t iv e A n a ly s is , O xfo rd : O xford
U n iversity Press.
------- P e c k , M . J,, and K a l a c h e k , E. (1967), T e c h n o lo g y E c o n o m ic G ro w th a n d P u b lic P o lic y
W ashington, D C : The B ro o kin gs Institution .
O E C D (1963a), S c ie n c e , E c o n o m ic G ro w th a n d G o v e r n m e n t P o licy, Paris: O E C D .
-------(1963b), “ P ro p o sed Stan dard Practice fo r Su rveys o f Research an d D e v e lo p m e n t”
D irectorate fo r Scientific A ffairs, D A S/PD /62.47, Paris: O E C D .
-------(19 71), S c ie n c e , G ro w th a n d S o c iety ; Paris: O E C D .
-------(1980), T e c h n ic a l C h a n g e a n d E c o n o m ic P o lic y , Paris: O E C D .
-------(1981), S c ie n c e a n d T ech n o lo g y P o lic y f o r the 19 80 s , Paris: O E C D .
-------(19 9 1), T ech n o lo g y a n d P r o d u c t iv it y Paris: O E C D .
-------(1992), T ech n o lo g y a n d th e E c o n o m y ; Paris: O E C D .
-------(1994), T h e O E C D Jo b s S tu d y : F acts , A n a ly s is , S tra te g ie s , Paris: O E C D .
-------(1996), T ra n sitio n s to L e a r n in g E c o n o m ie s a n d S o c ieties , Paris: O E C D .
* -----(2000), K n o w le d g e M a n a g e m e n t in the L e a r n in g S o c ie t y Paris: O E C D .
-------(2001), T h e N e w E c o n o m y : B e y o n d the H y p e , Paris, O E C D .
O s t r y , S. (1996), “ P olicy A pproach es to System Friction: C on vergen ce Plus,” in S. Berger and
R, D o re (eds.), N a t io n a l D iv e r s ity a n d G lo b a l C a p it a lis m , Ithaca and L o n d o n : C orn ell
U n iversity Press.
P a v i t t , K, ( i 9 9 5 )> A cadem ic Research, Technical C hance and G o vern m en t Policy,” in
J. Krige and D. Pestre (eds.), S c ie n c e in th e 2 0 th C e n t u r y F larw o o d A cad em ic Publishers,
143-58,
— (1998), “ T he Inevitable Lim its o f EU R & D Fu n d in g,” R esea rch P o lic y 27(6): 559-68.
 S C I E N C E , T E C H N O L O G Y , AND I N N O V A T I O N POLI CY 631

*-— (2001), “ Public Policies to Support Basic Research: What Can the Rest o f the World
Learn from the US Theory and Practice (And What they Should not Learn);’ Industrial
and Corporate C han ge 10(3): 761-79.
Per ez, C. (2002), Technological R evolutions a n d Fin a n cia l C apital, Cheltenham: Edward
Elgar.
Per ko u x, F. (1969)» L ” Econom ic du zoem e siecle , Paris; Presses universitaires de France.
Peter so n , J. (i99i)> “Technology Policy in Europe: Explaining the Framework Programme
and Eureka in Theory and Practice,” Jo u rn a l o f Com m on M arket Studies 29(3): 269-90,
,
------ and S h a r p , M. (1998), Technology Policy in the European Union London: Macmillan.
Pu tn am , R. D. (i993)> M a k in g D em ocracy Work: C ivic Traditions in M odern Italy., Princeton:
Princeton University Press.
R ip , A., and v a n d e r M e u l e n , B. J. R. (1997), “The Post-Modern Research System” in
R. Barre, M. Gibbons, ]. Maddox, B. Martin, and R Papon (eds), Science in Tomorrow's
Europe, Paris: Economica International, 51-67.
R omer , P. M, (1990), “Endogenous Technological Change,” Jou rn al o f Political Econom y 98:
71-102.
R o se n b e r g , N. (1972), Technology a n d A m erican G row th , New York: M. E. Sharpe.
Ro th w ell , R., and Z e g v e l d , W. (1980), Innovation a n d Technology Policy, London: Pinter.
S h a r p , M. (2003), “The UK-Experiment: Science, Technology and Industrial Policy in

Britain 1979-2000,” in P. Biegelbauer and S. Borras (eds), Innovation Policies in Europe

and the US, The N e w A gen da, Aldershot: Ashgate, 17-41.
Svennilson , I. (i960), “The Concept of the Nation and its Relevance to Economic,” in E. A.
G. Robinson (ed.) E conom ic Consequences o f the Size o f Nations, Proceedings of a Confer­
ence held by the International Economic Association, London: Macmillan.
W a g n e r , M. R (1998), D et polytekniske gennem brud, Aarhus: Aarhus Universitetsforlag.
Westling , H, (1991), Technology Procurem ent: For Innovation in Swedish Construction,
Stockholm: Council for Building Research.
---- (1996), C ooperative P ro cu rem en t M arket Acceptance fo r Innovative Energy-Efficient
Technologies, Stockholm: NUTEK.
W o o lco ck , M. (1998), “ Social Capital and Economic Development: Toward a Theoretical
Synthesis and Policy Framework,” Theory a n d Society 27(2): 151-207.
Yaku sh iji , T. (1986), “Technological Emulation and Industrial Development,” Paper pre­
sented at The Conference on Innovation Diffusion in Venice, mimeo.
Z ucker , L., and D a r b y , M . (1998), “ Intellectual Human Capital and the Birth o f US
Biotechnology Enterprises,” American Economic Review 88(1): 290—306.
 I n d e x

Note: Bold entries refer to boxed features.

Abramovitz, Moses 366 Argentina, and catch-up 525-34,553~4

and catch-up 523,524 ,525 Argyris, C 125-6
and economic growth 489, 490, 516 Arrow, Kenneth 148, 212, 279,491, 613
and technological congruence 535 Arthur, W B 463,472
absorptive capacity 508,561 Arundel, A 169, 281
and catch-up 20 Asheim, Bjorn, and regional systems of
and educational level 341 m6 innovation 184,300
and firm size 333 asset-augmenting activity, and multinational
and industry-university links 234 enterprises 327” 9>329-30* 332, 334*337
and innovation 11 asset-exploiting strategies, and multinational
and inward investment 338 enterprises 326-9,329-30,333* 334
and knowledge transfer 77 Australia 428
and low- and medium-tech industries 417 and patent laws 189
and networks 67 and university funding 210
acquired technology 156-8 and vocational and technical education 194,
Actor-Network Theory 74 195
adaptation, and organizational change 116,133-5, Austria:
i 37_9 and collaborative networks 197
adhocracy model 120 and universities 215, 219
and organizational learning 127-32 and vocational and technical education 194,
and radical innovation 119,128,132 195
Silicon Valley 127,129,130
and spaghetti organization 131 ba, and organizational knowledge creation 125
Advanced Micro Devices (AMD) 48 Baden-Wurttemberg 301-2,307
aerospace, and collaborative networks 59 Bandai 416
Aghion, P 502,503,506,509 Bank of England 254-5
Agricultural Research Laboratory Bank of Japan 42,621
(Denmark) 601 banking system:
Agricultural University (Copenhagen) 601 and credit creation 244
Airbus 622 and industrialization 517
Alcoa 363 and Japanese industrial organization 41,42,45,
Alderman, N 308 519
Almeida, P 328 Barras, Richard 440
ambidextrous organizations 119,136,141 П4 Bathelt, H 310
American Research and Development 251 Batt, R 587-8
American Telephone and Telegraph 364 Baumol, W J 523-4, 581
Andreesen, Marc 370 Baxter, William 274
Aoki, M, and J-form 127 Bayer 323
Appelbaum, E 587-8 and foundation of research laboratory 362
Apple 368 Bayh-Dole Act (USA, 1980) 209—10, 276
architectural innovation 21115,141 04 effects of 229-31
 in d e x 633

international emulation o f 227, 232 business services, see knowledge-intensive

origins of 228-9
Belgium:
and multinational enterprises 326
and university-industry links 219
Belleflame, C 441
Ben-David, J 215,23508
Benetton 425, 426
Berger, A 254
Bernal, J D 604, 616
Berne Convention for the Protection o f Literary
and Artistic Works 270, 271
bibliometrics 87,150 ,152,153, 221
biomedical sector:
federal support for 371
and patents 228, 229
and university research 221, 223
biotechnology industry 555-6
and collaborative networks 6o, 65, 66, 67,
247
and entrepreneurial regional innovation
system 304-5
and geographical clustering 291
and informal networks 71-2
and international linkages 334
and knowledge sharing 75
and knowledge spillover 297
and patents 229, 616
post-war development 371-2
and science, technology, and innovation
policy 602
and university research 94, 221, 391,392, 397
Boulton, Matthew 357
Boyer, R 580,581
Braczyk, H, and regional systems o f
innovation 12,184
Brazil, and catch-up 525-34
Breschi, S, and sectoral innovation
systems 183-4, 200
British Rail 442
British Telecom 437
‘Brooks Report' (OECD) 603
Brown, J S 309
Brown, S L, and organizational change 137
Bruland, Kristine 464
Brusco, Sebastiano 37
Brynjolfsson, Eric 473
Burgieman, R A, and strategic adaptation 137
Burns, T, and organizational types 118
Burt, R $, and networks 61
Bush, Varmevar 8 ,212,6 0 4 -5,6 19
business cycles 6,242,248,249
and clusters o f innovation 15
and economic development 244
business services
‘buzz’:
local 310,311
and localized knowledge circulation 297-8
call centers 438, 439
Calvert, ] 220
Canada:
and research and development intensity 155
and universities 195, 216, 217, 218, 232
and vocational and technical education 195
Canon i d , 141
capital markets 135, 254,306, 502, 613
capitalism:
differential impact o f 514
evolutionary nature of 260
and innovation 18, 572
intellectual 277-8
varieties o f 132, 205 024,305
Carlsson, B, and technological systems of
innovation 183
Carr, E H 190
carrier industries 417
Caselli, F 473
catch-up 516
and absorptive capacity 20
and convergence 514-15,523
cross-national comparisons 537-8
access to foreign technology 533
and education 525-30, 553~5
factors affecting 531-4
and foreign direct investment 530-1,533
and Gross Domestic Product 525
and research and development 530
and developing countries 520,523,525,535,536,
537
and developmental state 516,534
and diffusion 514, 516
and education 521, 525-30,533,537>538» 553~5
European, pre-World War I 516-18
explanations o f 523-4
favorable post-war conditions 523
and infrastructure 518, 523, 535, 537
and innovation 19, 5*5>
Japan 272-3,518-19,520
obstacles to 624
and patent system 275,284
and policy 534-7
role o f government 374>52i> 552
and service sector 535
Singapore 518, 520-1
and social capability 5*3>5*4
South Korea 518, 520-1, 522
634 INDEX
catch-up (confd)
strategies 518
Taiwan 518, 520-1
and technological congruence 523,524
and trade liberalization 521--2
Celera, and Human Genome Project 56, 57
Celestica 49
Centre National de la Recherche Scientifique
(CNRS) 193
CERN 369,370, 622
Chandler, Allred 262 пз
continuity thesis 106
and growth of industrial enterprise 39
and Japanese challenge 40
and M-form company 418
the managerial revolution 32,38
and organizational design 102,117
Chang, Ha-Joon 534-5
change:
drivers of 414-15
demand differentiation 415-16
technology 416-17
and innovation process 102,104-7
and organizational innovation 133
incremental view of 134-5
punctuated equilibrium 133,135-6
strategic adaptation 137-8
resistance to 107-8,134
chemical industry:
and Second Industrial Revolution 359-60,
362-6
sectoral structure 392-3
Chenery, H B 488
Chesnais, Francois 603
Chiaromonte, F 500
Child, J, and strategic adaptation 137
Chile, and catch-up 525-34
China, and catch-up 525-34, 554
Christopherson, S 306
Cisco Systems 47-8, 49,107
Civil Aeronautics Board (USA) 477
clusters:
and biotechnology 304-5
geographical 291
and innovation 14,15,18, 20
and innovation diffusion 6
inter-sectoral 398
regional 225-7, 299,300,302
temporal 498-9
Coca-Cola 415
Coe, D T 507-8
cognition 20
and organizational innovation 123-4
and organizations 116,139,140
shared 125,126
Cohen, W M 417
Coleman, W 473
collaboration, see networks
Community Innovation Survey (CIS) 148-9,166,
616
criticism of 168-9
innovation definition 164
and measurement of innovation 162-5
and organizational interaction 196-7
results from 165-8
and service sector:
innovation in 446,447
research and development 437
sources of information for innovation 448
usefulness of 169-70
Company Reporting Ltd 437
Compaq Computer 48
comparative advantage 407
and competitiveness 545
Denmark 409-10
and developing countries 428-9
and France 557
and globalization 553
and Japan 621
and United States 555-6
compensation theory, and employment 572,
580-2
competence building:
and education 192,194-5
and labor market institutions 614
and systems of innovation 190-2,194-5,617
competition, technological:
and innovation 14-15
and product-life-cycle theory 15-16
competitive advantage 545, 559
and economic system 305-6
and institutional frameworks 305-7
and Japan 43, 45,129
and regional innovation systems 311
sources of 419, 560, 562
competitiveness:
and capability construction 544, 561
definition 544
and diffusion 548
and employment 569
at firm level 559-ёо, 561-2
at industry level:
capabilities 551,552
demand conditions 551
and education 553—5
and industrial leadership 551
institutions 551, 552
inter-company networks 551
 INDEX 635

and technological specialization 554-7 Court of Appeals for the Federal Circuit
and innovation 87,122, 545, 546, 561, 583 (CAFC) 228-9, 274, 277
and innovation policy 612, 618 craft labor, and innovation 35-7
and intellectual property rights 266» 274» 275, Cray Systems 368
276 creative destruction 32,105,115,126,359,382,
and Japanese policy 621 503
measurement o f 544 credit:
as national goal 544 and innovation 242, 243, 244
at national level: rationing of 521
and capital accumulation 547, 548-50 Crevoisier, О 311
and currency values 546 critical technology 349-50
and innovation 546-50 cross-shareholding, and Japanese industrial
and international trade 545-7 organization 42
and technology gap 547 currency, and competitiveness 546
and prices 546,581
and product and process innovation 573 da Vinci, Leonardo 5
at regional level 294,301, 303, 557-9 Darby, M 297, 607
research opportunities in 562-3 David, Paul 213,463,473-4
revival of interest in 543-4 Dawkins, Richard 495
and science parks 226, 303 decision-making:
sources o f 561 and adoption o f new technology 465-6
and speed to market 425 and finance 246
and United States 620-г in firms 39,42
Computer Associates International 48 and innovative investment 257-8
computer industry: organizational 123,124
and coevolutionary processes 397 and radical innovation 108
post-war development 368-71 and scientific knowledge 606
sectoral structure 393 DeFillippi, R 127,130
and standard setting 68 Dell Computer 48,49
Computer Research Associates 368 demand, and sectoral innovation systems 391
Conlisk, J 500-1 Denmark 428
Constant, E, and knowledge production 97 and collaborative networks 197
consultancy, and systems o f innovation 191 and comparative advantage 409-10
contingency theory, and organizational and cooperative movement 601
structure 117-21 and innovation policy 614
contract manufacturing 99 and national system o f innovation 199
control, and ownership 38 and patent laws 189
convergence: and research and development 194
and catch-up 514-15, 523 and vocational and technical education
conditional 506 194
cross-country 554-5, 557 developing countries:
and growth rates 505 and catch-up 520,523* 525.535»53^, 537
and intellectual property rights regimes and diffusion 460
273 and education 553-4
technological 91-2, 99,337 and employment 583
Cooke, P, and regional innovation systems 184, and foreign direct investment 332,338
300-5 and intellectual property rights 271, 273, 275,
Cooperation in science and technology 284
(COST) 622-3 and low- and medium-tech industries 410,
cooperative movement, Denmark 601 428—9
core competence 93,126 ,134 and science, technology, and innovation
Cornwall,! 4 9 h 523 policy 601, 626
corporate change 418-21 development blocks 389» 429
corporate finance 250-5 developmental state, and catch-up 51b, 534
636 INDEX
diffusion 478-9
and adoption decision 465-6
and American agricultural industry 602
and catch-up 514, 516
and clusters 6
and communities of practice 309
and competitiveness 548
conceptual frameworks 461-5
and consumer heterogeneity 466-7
and consumer learning 467
as creative process 15
and Danish cooperative movement 601
definition 459-60
and developing countries 460
and economic growth 486, 500,508
and employment 573, 582-3, 589, 590
and feedback effects 460,462, 464,470,479
in First Industrial Revolution 353,355
government fostering of 428
and government policy 614, 625
and imitation 21 пз
and immigration 600
and incremental innovation 311, 504
informal channels of 558
and information and communications
technology 582, 584, 590, 617
and innovation 15, 21 03,150
and Internet 370
and interpersonal communication 465
and learning by observing’ 459
and linear model of innovation 479
measurement of 616-17
modeling of 465-8
and multinational enterprises 318,323, 327
and networks 59,73,335,391
and Organization for Economic Cooperation
and Development 602
as part of innovation process 460
of patent system 269, 270-1
and patents 279, 280
and QWERTY controversy 463
rate of 459, 460, 466-7
determinants of 468-79
and regional innovation systems 312 05, 558
research opportunities 479-80
and science, technology, and innovation
policy 599, 626 m
and sectoral systems of innovation 381, 384,
3 9 i - 3 >399
and service sector 617
and skills 624
and software industry 398
and systems of innovation 183,184,186,189,
190, 200,201
and technological revolution 249
and technology transfer 479-80
and Third Industrial Revolution 367,369
and universities 211-12
Digital Equipment Corporation 251,368
digitalization 96, 97,105
discontinuity, and organizational change 133,
135-6
Dosi, G 497*500
Dow 365
Downie, J 543, 560
Du Pont, and foundation of research laboratory
363
Duchin, F 582
Duguid, P 129,309
Duisberg, Carl 362
Dvorak keyboard layout 463
Eaton, J 508
economic development:
and business cycle 244
and entrepreneurship 32
and innovation 14, 241, 248, 255
and universities 209, 210
and regional development 225-7
economic growth:
differential rates of 19, 20
and diffusion 486, 500, 508
and endogenous growth theory 250, 255,488,
501-5
critique 506-7
empirical work on 505-8
origins of 490
and evolutionary theory 488-9, 492-3, 495,
505
evolutionary philosophy 496-7
formal growth models 499-501
and innovation clusters 498-9
non-formal theorizing 497-9
technological trajectory 497-8
extensive 351
factors affecting 487
and finance 250, 255-6, 258-9
and government policy 492
and growth accounting 488-9, 490
historical rates of 487-8
and imitation 19
and innovation 18-19, 20, 492
intensive 351
and investment 501-2, 549
and dong-wave’ 499
neoclassical m odel 488, 492, 493, 497, 500
and patent system 284
and Post-Keynesianism 489-90

research opportunities in 508-9
and science, technology, and innovation
policy 603, 604-5, 608, 612, 621, 626
and sectoral mix 488
and significant innovations 493-4
sources of 18-19
and spillover effects 490-1
and structural change 487-8
and technological change 255, 488-92
and technology 350, 492
and total factor productivity 488, 489
traditional economics approaches 488-92
and universities 211, 212, 232
economic sociology 117
Edquist, Charles, and national systems o f
innovation 186, 212
education:
and absorptive capacity 341 ni6
and catch-up 521, 525-30* 533>537* 53®* 553-5
and competence building 192,194-5
and competitiveness 553-5
and developing countries 553-4
and First Industrial Revolution 35
and innovation 192
and management 39
and science, technology, and innovation
policy 606, 610
as strategic asset 210
see also universities
Eisenhardt, K M 68,127
and organizational change 137
electronics industry, and Third Industrial
Revolution 367-9
Emilia-Romagna 37,300
employmen t, and impact o f innovation 7, 568,
571-2,573-5,589
and adjustment to 571
and compensation theory 572, 580-2
and competitiveness 569
and developing countries 583
different perspectives on 569-71
and labor market 591
and national system o f innovation 590
and organizational innovation 573, 576, 590
and process innovation 572-3, 579, 583, 590
and product innovation 569, 573-5, 579-®°»
5®3, 59° , 592 пз, 593П9
and quality o f employment 583-4, 5®®-9
impact o f organizational change 587-8
skill-biased technical change 584-7, 590
wage polarization 587, 590
and quantity o f employment 575-6
firm level 576-9
industry level 579-80
in d e x 637
macroeconomic level 580-2
simulation studies 582-3
and research and development 364, 591-2
and service sector 433-4, 449> 450> 580
and technological change 569-71,5®°, 582,588,
5® 9, 590-1
and trade 590
endogenous growth theory 18,197, 250, 255
and economic growth 488,501-5
critique 506-7
empirical work on 505-8
origins o f 490
research opportunities in 509
Energy, Department of (USA) 56
entrepreneurial regional innovation system
(ERIS) 304-5
entrepreneurs:
and economic development 32
and innovation 5, 6,9,32, 242-3,259
and limits to firm growth 38
and networks 66-7, 72
and New Economy model 46
Ericsson 101,198
Ermoliev, Y 472
ESPRIT program 623
ethnic communities, and networks 63, 78
EUREKA 98,623
European Commission:
and measurement o f innovation 160,162,
167
see also Community Innovation Survey
European Community Patent Convention
(1975) 272
European Innovation Monitoring System 167
European Molecular Biology Laboratory
(EMBL) 622
European Patent Convention (EPC) 272
European Patent Office (EPO) 272
European Research Area 623
European Science Foundation (ESF) 622
European Southern Laboratory 622
European Space Agency (ESA) 622
European Union 98,429
employment impact o f innovation 582
and European Research Area 623
framework programs 623
and innovation policy 612
Lisbon Summit (2000) 623
and research and development intensity 155
and science, technology, and innovation
policy 155, 610, 621-3, 624
and service sector research and
development 436-7
see also Community Innovation Survey
638 INDEX
evolutionary theory 486
and economic growth 488-9. 492- 3. 495. 505
evolutionary philosophy 496-7
form al growth models 499-501
and innovation clusters 498-9
non-form al theorizing 497-9
technological trajectory 497-8
and industrial dynamics 17
research opportunities in 509
and sectoral innovation systems 386-7
exports 4 11-12, 546, 550, 583
and catch-up 521
and high-tech products 319
Fabricant, S 491
Fagerberg, Jan:
and catch-up 524, 535
and competitiveness 545.547. 548
and differential growth rates 19
Fairchild Semiconductor 46
Faraday, Michael 359
Farrell, J 472,473.477
fast food sector 435, 438, 442
feedback:
and diffusion 460, 462, 464, 470, 479
and economic growth 490
and innovation 95-6,185, 332, 466
and managing uncertainty 101
finance:
and competitive interactions 247
contextual factors 259
corporate 250-5
and decision-making 246
and economic development 248-9
and economic growth 250, 255-ё, 258-9
economics o f 250-6
and firm s stage o f development 254
and industry entrants 246-7
and industry evolution 248
and information asymmetries 257
and innovation 240-1
availability o f 248
economics o f 241-9
and investment 257-8
research opportunities and
difficulties 256-7, 259-62
sectoral patterns o f 246-7
and investment behavior 253
and organizational learning 246
and relationships with industry actors 247
and research and development 253-4
sources o f 250-3
and systems o f innovation 191
and techno-economic paradigms 248, 249
and technology-based small firms 254-5
and theory and dynamics o f economic
change 257-9
and venture capital 251-3
financial services, and geographical
clustering 291
Findlay, R 338,341 m6
Finland:
and catch-up 525-34
and innovation policy 614
and knowledge-intensive business services 451,
452
and universities 216, 219
firms:
and absorptive capacity 11,333
and catch-up 536
and competitive interactions 247
and competitiveness 559-60
and corporate finance 250-5
and finance 2 9 ,30 ,34 , 50, 51, 254-5
growth determinants 559-60
and heterogeneity 390
and historical transform ation 31
and innovation 5
employment impact o f 576-9
implications o f 14
process o f 10 -12
large 32
and limits to growth 31, 38
optimizing 30 -1
organization 29, 30, 33,34, 5 0 -1
and organizational learning 30, 32, 34, 245-6
and organizational m em ory 11
path dependency o f 95
resource-based view o f 33
skill base 34
social characteristics o f innovative 30-4
industrial districts 35-7
Japanese model 41-5
managerial corporation 38-40
New Econom y model 45-9
strategy 29, 30, 34, 50, 51. 418
Flextronics 49
flight simulation industry, and collaborative
networks 68
food industry:
and absorptive capacity 417
innovation in 426-7
as low- and medium-tech industry 409-10
and vertical alignment 423
Ford, Henry 520
foreign direct investment 15
and asset-augmenting activity 327-8
and catch-up 530-1, 533
 INDEX 639

and developing countries 332, 338 Gerscbenkron, Alexander 516

growth in 322 and catch-up 517-18, 524
and multinational enterprises 318, 322-3 Gershuny, J I 442
and product-life-cycle theory 327 Gibbons, Michael 213
and spillover effects 338, 508 Gittelman, M 607
France 394 globalization:
and catch-up 525-34 and comparative advantage 553
and government-funded programs 98 cross-border commercialization of national
and patent laws 269 technology 319-20
and research and development 193 impact o f 424
and technological specialization 557 implications o f 318
and technology policy 608 and innovation 320
and universities 216, 217 and multinational enterprises 318, 322-6
and vocational and technical education 194 and strategic technology partnering 320-2
Frascati Manual 151,153-4, 4*o, 616 and tacit knowledge 292,310
Freeman, Christopher 3,311, 497, 603, 626 Glynn, M A, and innovation 123
and Japanese technology policy 621 Golder, P N 465, 467
and national system o f innovation 183 Gomulka, S 523
and science policy 604, 616 governance:
Fujitsu 44 models o f corporate 122
Furman, J L, and national innovative capacity 197 and networks 60-1,77-8
and regions 299
Galbraith, John Kenneth, and technostructure 40 government:
Galli, R 189 and catch up 374, 521,552
Galunic, D C 127 and diffusion 614, 625
Gann, D M 308 and economic growth 492
General Agreement on Tariffs and Trade, and and high-tech industries 409, 427-8
intellectual property rights 274, 275 and innovation process 96, 98
General Electric 48,363, 364 low- and medium-tech industries 427-9
general purpose technologies 416-17, 559 and regional economic development 225-7
and economic growth 504-5 and research and development 98
genetic engineering, and science, technology, and small and medium-sized enterprises 428
and innovation policy 602-4 and Third Industrial Revolution 367, 372
geography: and university-industry collaboration 209,
and innovation 29 1-4,30 9 -10 210, 224-5
and analytical knowledge 297 see also science, technology, and innovation
dual 310-11 policy
and synthetic knowledge 296-7 Grabher, G 60-1
and sectoral innovation systems 399, 400 Granovetter, M 59, 61, 62
and system boundaries 199-200 Great Depression 40
see also regional innovation systems Greece, and catch-up 525~34
German Association for Patent Protection 361 Greenan, N 473
Germany: Greenspan, Alan 604
and catch-up 525-34 Griliches, Zvi 148, 464, 507
pre-World War I 515, 517-18 growth accounting, and economic growth 488-9,
chemical industry 359-60,362-6 490
and education 194,554-5 Grundtvig 601
and patent laws 189 Guellec, D 473
and research and development 18 8 ,19 3,323
and Second Industrial Revolution 358, 359-63 Hagedoorn, John 152
and technological specialization 556 Hamilton, Alexander 534
and technology policy 608 Hamilton, K 220
and universities 218, 219, 232,360-1 hard disk drive industry 247» 248
and vocational and technical education 195 Hedlund, G 127,327
640 INDEX

Heertje, A 508-9 industrial complexes, and technology policy 6x0,

Helpman, E 507-8 619—20
Hemingway, Ernest 429 industrial districts:
Henry VIII 600 British 35-7
Hewlett-Packard 48 and collaborative networks 59
Hicks, D 220 and competitiveness 558
high-tech industries: and the Third Italy 37
economic impact o f 407-8 industrial dynamics 16 ,17
and exports 319 and low- and medium -tech industries 424-5
government attraction to 409, 427-8 and sectors 382
and industrial classification 16 Industrial Enlightenment, and the First Industrial
and innovation in 414, 415 Revolution 355
and research and development 155 industrial growth, and product-life-cycle
role o f technology 419 theory 15
and science, technology, and innovation industrial organization:
policy 609 industrial districts 35-7
Hirschman, A О 6o8 Japanese model 41-5
historical transformation, and firms 31 managerial corporations 38-40
Hitachi 43, 44, 621 New Econom y model 45-9
Hobday, M 8, 533 see also industry, classification o f
Hoechst 323,362 Industrial Revolution, First 350-1
Hogge, Ralph 600 and economic growth 487
home-base augmenting activity, and and education 35
multinational enterprises 327 and general purpose technologies 416
home-base exploiting activity, and multinational and Industrial Enlightenment 355
enterprises 326 innovation in 351-5, 357-8
Hong Kong, and catch-up 525-34, 553-4 innovation system 374
Hounshell, David 39, 40 and institutional change 356-7, 357-8
Howell, D 584 and learning 355
Howitt, P 502, 503, 506 and technological diffusion 353, 355
Human Genom e Project, and organizational Industrial Revolution, Second:
innovation 56-7 characteristics o f 358-9
Humble Oil 365 and chemical industry 359-6 0 ,36 2-3
interwar innovation 364-6
IBM 4 4 ,4 5 ,4§> 73> 107,108,368,369,370.393> 397, and firm structure 359
494 and general purpose technologies 416
imitation 382, 500 innovation system 374
active 22 mo and origins o f industrial research 36 0 -1,36 2-3,
and economic growth 19 364
and innovation 8,15 and role o f science 359-60
immigration, and diffusion 600 and universities 36 0 -1
Imperial Institute o f Physics and Technology Industrial Revolution, Third:
362 and biotechnology 371-2
income: and electronics industry 367-9
and education 624 and general purpose technologies 416
international divergence in 514, 526, 527 and information and com m unications
and productivity growth 544 technology 369-71
and wage polarization 584, 587, 590 innovation system 374
incremental innovation 7 -8 ,12 2 ,12 8 ,13 2 ,16 2 , and the Internet 369, 370, 371
248, 390, 494, 495>496~~9>572, 573 new innovation resource base 372-3
and diffusion 311, 504 and pharmaceutical industry 37 1-2
India: and post-war transformation 366-7
and catch-up 525-34 and role o f government 367
and software industry 4 10 -11 and technological diffusion 367, 369

industrialization:
and banking system 517
and foreign direct investment 521
and patent system 284
and service sector 437-9
see also Industrial Revolution
industry:
classification o f 155» 4° 9>411-14» 429-30
and factory intensity 411
and technological profile 409-10
drivers of change 415-17
knowledgebases 294-9
technological classification o f 16
inertia:
and entrepreneurs 6, 9
and innovation 9-10
and organizational dhange 133,134 ,136
overcoming 116
systemic 332
information and comrriunications technology
(ICT):
adoption costs 473-4
and coevolutionary processes 397
diffusion of 582, 584,1590, 617
and government-funded programs 98
and innovation process 97, 562
and intellectual property rights 277-8
and New Economy model 45-9
and research and development 154
and service sector innovation 439-45, 448
and skills 584-7, 590
and standards 472
and Third Industrial Revolution 369-71
infrastructure 39, 98
and catch-up 518, 523, 535, 537
and competitiveness 553
and First Industrial Revolution 357
institutional 299, 301,302
knowledge 203, 299,304,332
research 12 ,14
and Second Industrial Revolution 360
and Third Industrial Revolution 370, 372, 373
innovation:
and analytical knowledge 296
and catch-up 19, 515, 519-20
chain-link model of 150
and clusters 14,15
temporal 498-9
and competitiveness 87,122, 545, 546, 561. 5^3
context o f 8
contingent nature o f 87, 96
as continuous process 5-6
cross-disciplinary nature of 2-4
definition 4 ,12 3 ,14 9 ,16 4
INDEX 641
and economic change 6
and economic development 14, 255
and economic growth 18-19, 20, 255
and education 192
geography o f 291-2,309-10
and analytical knowledge 297
dual 310-11
and synthetic knowledge 296-7
incremental 7-8,122,128 ,132,16 2, 248, 390,
494»495»496-9» 572, 573
and industrial classification 16
and industrial districts 35-7
and invention 4-5
and Japanese model 35-7
and linear model 8-9,93,95, 212,229,302,428,
479
as matching process 88
New Economy model 45-9
and path dependency 10 ,12 ,13,9 5,10 9
process o f 9-12
radical 7-8, 202,494
and adhocracy 119,128,132
and collaboration 65
decision-making about 108
and economic structural change 104
and knowledge application 296
and learning economy 311
role o f 19-20
sectoral differences 246-7
significance o f 493-4
and synthetic knowledge 295-6
systemic nature o f 12-14
types o f 6-8
autonomous 122
systemic 122
and uncertainty 9,10,30, 88,100-1, 493
variability o f 14
see also organizational innovation; process
innovation; product innovation
innovation policy, see science, technology, and
innovation policy
innovation processes 87-9,108-9
continuous nature o f 5-6
and firms 10-12
heterogeneity in 95-6
and knowledge production 86,88
and complexity 96
and industrial research and development 91
industry-university links 93-5
and vertical disintegration 91-З» 99
and market demand 86, 88-9,101-8
and disruptive change 102,104-7
and organizational practices 102-4
and tribal warfare 102,107-8
 INDEX
innovation processes (cont’d )
and project management 9 1-2
and Schumpeter 9-10
transformation o f knowledge into working
artifacts 86, 88, 96-7
government-funded programs 96, 98
and managing uncertainty 9 6 ,10 0 -1
systems integration 96, 99-100
technology trajectories 96,97-8
innovation systems, see systems o f innovation
(SI)
Institute for Scientific Information 153
institutional regional innovation system
(IRIS) 303,304
see also regional innovation systems
institutions:
change in 375
and competitiveness 551, 552
conceptual diffuseness o f term 186
definition 188
distinct endowments o f 294
and innovation policy 612, 6 13-14
and organizational innovation 132 -3,134
and regional innovation systems 299
and sectoral innovation systems 385, 394-5
and systems o f innovation 185-6,188-9, 191, 196
and technological change 394
Intel 4 4 ,4 8 ,4 9 ,36 9 ,4 9 4
and strategic adaptation 137
intellectual property rights 266-7, 284-5
and academic research 232
and Bayh-Dole Act 228-32
Berne Convention 270, 271
and developing countries 271, 273, 275, 284
and economic progress 284
history o f 268
international patent era 271-3
multinational patent era 270-1
national patent era 267-9
pre-patent era 267
pro-patent era 273-8
and information and communications
technology 277-8
and intellectual capitalism 277-8
and investment 279
nature o f 278
Paris Convention 270, 271, 273
and post-war United States 367
role o f 278-80, 285
national differences 283-4
sectoral differences 282-3
and Second Industrial Revolution 361
Trade-Related aspects o f Intellectual Property
Rights 273, 274, 275, 284
and universities 606
see also patents
International Association o f Science Parks 226
International M onetary Fund 535
international trade:
and competitiveness 545
and employment 590
liberalization o f 521-2
and wages 587
see also General Agreement on Tariffs and
Trade; World Trade Organization
Internet 98, 369, 3 7 0 ,37 h 397-8
diffusion o f 477
and wireless connection 463-4
interpersonal relations, and networks 60
invention, and innovation 4 - 5 ,2 1 n2
investment;
and compensation mechanism 581
complementary 473-4
and economic growth 501-2, 549
and inform ation asymmetries 257
and innovation 242
and intellectual property rights 279
and liquidity 253
and research and development 279,323-8,410,
491, 500, 503,605
and service sector 436-7
and uncertainty 257-8
see also foreign direct investment; inward
investment; multinational enterprises,
offshore research and development
investment; venture capital
investment regimes, and competitive
advantage 306
invisible colleges, and science networks 73
inward investment 305, 546
host country effects 338
see also foreign direct investment; investment
Ireland:
and catch-up 525-34, 537-8
and service sector 434
and university-industry links 219
Isaksen, A, and regional systems o f
innovation 184
Italy:
and catch-up 525-34
and industrial districts 37
and patent laws 189
and research and development 193
and territorially embedded regional innovation
systems 300
and universities 216, 219, 232
and vocational and technical education
194
 INDEX 643

I-form model, and organizational learning 127, knowledge:

128,132 accessibility o f 388
Jabil Circuit 49 analytical 295, 296, 297
Jacobsson, S, and systems o f innovation 189 collective 124-5
Japan: costs o f transfer 76
and banking system 41, 42, 45, 519 cumulativeness o f 388-9
and catch-up 272-3, 515, 518-19,520 distributed 58
and comparative advantage 621 explicit 75,76
and competitive advantage 43, 45,129 geography o f 292-4
and cross-shareholding 42 global networks 310
diffusion o f major innovations 469 industrial bases 294-9
economic development 366, 373, 375 and learning economy 293
economic stagnation 45 localized circulation o f 297-8
and education 554-5 production o f 88, 89-90
financial system 522 and collective learning 124-7
and government-funded programs 98 and complexity 96
and industrial organization model 4 1-5,12 9 and industrial research and development 91
and lean production 420, 520 industry-university links 93-5
and lifetime employment 41, 42-3 and J-form 128
Meiji restoration 518 and organizational cognition 124
Ministry for Trade and Industry (MIT1) 519, and specialization 99
621 and vertical disintegration 91-3, 99
and organizational innovation 520 relationship between local and global 308-9
and patent system 272-3, 281-2 and sectoral innovation systems 385, 387-90
post-war industrial reorganization 41-2 spillovers 297-8
and research and development 188,193 synthetic 295-7
and science policy 624 tacit 75-6,125, 296
and service sector research and geography o f 292-3
development 436 and globalization 310
and technology policy 608, 609, 610, 621 and joint innovation 293-4
and universities 216, 217-18, 219, 220, 232 production o f 308-9
Japanese Patent Office 272, 273 and spatial proximity 294,308
Jefferson, Thomas 269 transfer o f 74 -7,7% 294
Johnson, A 189,198 knowledge-intensive business services, and
Joint European Torus (JET) 622 innovation 449-52
Joint Research Center (JRC) 622 Kogut, Bruce 33, 66, 607
joint-stock organization, and First Industrial Kortum, Samuel 48, 508
Revolution 356 Krugman, P 19, 544, 547
Jones, C 506-7 Kuhn,T 497
Jones, E L, and European development 351 Kumho 413
Jovanovic, Boyan 249
just-in-time production 420, 422, 515, 520 labor market:
and competence building 614
Kaiser Wilhelm Institute for Chemistry 362 and employment 569, 591
Kaldor, N 490 flexibility o f 135
Kaniovski, Y 472 and innovation policy 614
Katz, M L 58,472 and neoclassical economics 581
Keller, W 508 and wage polarization 587
Kleiner, Eugene 251 see also employment, and impact o f innovation
Kleinknecht, A, and patents 160 labor mobility 132, 219, 3$7
Kline, Stephen 165, 620 interfirm 44
chain-link model o f innovation 150 and Silicon Valley 130
innovation as continuous process 5 Lam, Alice 294,30b
and linear model 8-9 Lazonick, William 121,349
644 INDEX

Le Bas, C 329-30 List, Friedrich 534

leadership:
industrial 551, 563
and innovation 9 ,10
and learning 22 m3
scientific 366
technological 14, 333,358,360
lean production 75, 127 ,4 20 ,4 22 ,5 2 0
learning:
collective 294
and economic growth 490
and firms 3 0 ,3 2 ,3 4
and innovation 88
and innovative firms 50, 51, 245-6
industrial districts 35,36
Japanese model 41, 43, 44
managerial corporations 40
New Econom y model 46, 49
and knowledge transfer 294
localized 298-300, 311
and organizational forms:
adhocracy model 12 7 -9 ,13 0 ,13 1-2
institutional context o f 132 -3,134
J-form model 127,128
Japan 129
spaghetti organization 131
and organizational innovation 123-7
and relational proxim ity 308-9
and systems o f innovation 18 4 -5 ,19 1-2
and tacit knowledge 293
Leavitt, William 600
Leonard-Barton, D 59 ,10 7,126
Leontief, W 582
Lerner, Josh 48
Levinthal, D A 126, 417
Levitt, Theodore 437
liberalization 395, 521-2
Licht, G 448
Lichtenberg, F 508
Liebowitz, S J 463
limited liability, and First Industrial
Revolution 356
linear model:
and Bayh-Dole Act 229
and diffusion 479
and government policy 428
and industry-university links 93, 95
and innovation 8-9
and regionalized national innovation
systems 302
and universities 212
Linux community 73, 398, 472
liquidity 50, 252-3
Lisbon Summit {2000} 623
Liu, Xielin 189
Los Alam os project 56 ,6 0 1-2 , 605
Lou^a, F 497
low- and m edium-tech industries (LM T) 408,
429-30
and absorptive capacity 417
characteristics o f 408
and classification problems 16 ,4 11- 14
and corporate change 418-21
and demand differentiation 415-16
and developing countries 4 10 ,4 28 -9
and factory intensity 4 10 -11
and government policy 427-9
and industry change 422-5
innovation in 425-7
and new technological paradigms:
carrier industries 417
general purpose technologies 4 16 -17
and oil industry 410
and technological profile 409-10
low-tech industries, see low- and medium-tech
industries (LM T)
Lucent Technologies 48
Lundvall, Bengt-Ake:
and institutions 186
and learning economy 293
and national system o f innovation 12 ,18 3,18 6 ,
299, 300
and subcontracting 92
Luque, A 475-6
M acaque monkeys, and innovation 459
M cDonald’s 438
Machiavelli, Niccolo 107
machine tool industry 9 1-2, 417
and re v o lu tio n a ry processes 398
and innovation process 390
and institutions 395
McKendrick, N 357
M addison, Angus 487, 505
Majumdar, S 474-5
Malaysia, and catch-up 525-34
Malerba, Franco:
and sectoral innovation systems 183-4, 200
and technological regimes 382
Malmberg, A:
and localized capabilities 294
and tacit knowledge 292-3
Malthus, Thomas 571
management:
and competitiveness 560
and First Industrial Revolution 356-8
new practices 587-8

professionalization of 38-9
project-led 91
revolution in 38
and strategic adaptation 137-8
managerial corporations, and United
States 38-40
Manhattan Project 56, 601-2, 605
Mansfield, E 410, 475, 476
March, J G, and learning myopia 126
Margolis, S E 463
market economy:
coordinated 132, 305, 306
liberal 132,305,306
see also neoclassical economics
marketing:
and diffusion 464-5
and IBM 368
investment in 412
and Josiah Wedgwood 356
and low- and medium-tech industries 425
and Own Brand Manufacturing 533
markets:
and innovation process 101-8
and systems o f innovation 190
Marris, R 543
Marshall, Alfred:
and industrial districts 35, 36
and the innovating firm 31
on Lancashire textile industry 37
and limits to firm growth 31,38
Marsili, О 384, 411, 426
Marx, Karl 14,18, 353, 489, 572, 580
Maskell, P 292-3, 294
mass production 39-40,420,421,437,505,515,520
Max Planck Gessellschaft 193
measurement (of innovation) 65, 69, 79,148-9
and commensurability 149
Community Innovation Survey 148-9,162-5,
166
criticism o f 168-9
innovation definition 164
results from 165-8
usefulness o f 169-70
economic measures o f 151-2
indicators o f 152-3
bibliometric data 153
conceptual foundations 150-1
and innovation surveys 16 0 -1,16 9
‘object1 approach 16 1-2
patents 65, 69,158-60
research and development data 153-8
‘subject* approach 162-5
and novelty 149,150
and theories o f innovation 150-1
INDEX 645
mechanistic organizations 118,119 ,137
Meliciani, V 409
mercantilism 35
Mexico, and catch-up 525-34, 553-4
Mezias, S J, and innovation 123
Micron Technology 48
Microsoft 48, 437
Miles, R E 127
military programs 626
impact o f 548
and innovation 98, 367, 368,370
and semiconductor industry 393-4
and technology policy 620
Mintzberg, H 119 ,120,127
Moch, D 448
Mode 2 research 213-14, 221, 233
modular innovation 21 ng
modularization 99,438
Mokyr, J 355
motor vehicle industry, and vertical
alignment 422-3
Motorola 121, 274
Mowery, David C 58, 477, 551
multinational enterprises (MNEs):
and cross-border generation of
innovation 322-6
and cross-border knowledge flows 318-19
and diffusion 318,323,327
and foreign direct investment 318, 323
and globalization 318, 322-6
offshore research and development
investment 323-7
and asset-augmenting activity 327-8,
329-30
and asset-exploiting strategies 326-7,328,
329-30
and firm size 333
and home-base augmenting activity 327
and home-base exploiting activity 326
host country effects 338
impact on home country 338-9
and innovation systems 332
and integration costs 331
and market structure 333“ 4
opportunities and constraints 331-3
and organizational issues 334
spillover effects 337-8
overseas innovative activities 326-30
and regional innovation systems 558-9
and strategic technology partnering 320-2,
334-6, 339 пз
NASDAQ 47
national innovative capacity 197, 55^ 3
646 INDEX

National Institutes of Health (USA) 56 ,30 4 ,371 and interpersonal relations 60

National Research Council 57 and knowledge transfer 74-7, 79
National Science Foundation 57, 370 and m ulti-party relationships 72-4
formation of 619 and openness 12
and sendee sector research and and organizational innovation 56-7
development 436-7 and organizational learning 126
national systems o f innovation (NSI) 183,198-9, and patenting activity 65-6, 69,79
200 regional 299, 300
and employment 590 and risk-taking 68
‘hollowing out’ of 338-9 and sectoral innovation systems 385,390-4
and regional innovation systems 303-8 and stagnation 78
and sectoral innovation systems 386 and startup firms 66-7
see also regional innovation systems; sectoral and strong ties 12-13, 60, 61, 69, 79
innovation systems; systems; systems of and subcontracting 62, 69-70
innovation and systems o f innovation 191
N EC 45, 621 and technological uncertainty 68
Nelson, Richard R 17, 372, 508, 523, 617 and trust 58, 65, 70, 72
and evolutionary theory 497, 499-500 varieties o f 60-4, 66, 69, 70
and industrial leadership 551 and vertical alignment 422-3
and institutions 186 and weak ties 12 -13 , bo, 61, 69, 79
and national system of innovation 183,186 New Econom y 44, 266-7, 277, 495, 625
and organizational capabilities 33 and entrepreneurs 46
and organizational routines 126 and informal networks 71
and technology policy 620 and O ECD 604
neoclassical economics: and United States 45-9, 277
and economic growth 488, 492, 493, 497, 500 new economy innovation system (N EIS) 303,
endogenous growth models 501-5 304>З07
and innovation policy 613 Nishi, Yoshio 44
and labor market 581 N M T 450 198
Netherlands: Nobel Prizes 365
and innovation policy 614 Nokia 198
and universities 193, 216, 217 Nonaka, I 125,127
and vocational and technical education 194, Nordhaus, W D 279
195 North, D C, and institutions 188
network goods, and diffusion 470-3 north-south models, and economic growth 19
networks; Norway:
and absorptive capacity 67 and collaborative networks 197
and Actor-Network Theory 74 and research and development 193,194
advantages o f 59-60, 68-9, 79 intensity o f 155
collaborative 57, 58 and universities 215, 217
and coordination 334 novelty, and measurement o f innovation 149,
developing and maintaining 331 150
and diffusion 59,73, 335, 391 Nvidia 49
and diffusion rates 470-3
duration o f 78, 80 П4 oil industry:
dynamics o f 67-8 and frames 419
efficiency o f 67 as low- and medium-tech industry 408
and formal ties 65-70 and research and development 410
functions o f 60 Okimoto, Daniel 44
and governance 60-1, 77-8 Olsen, Kenneth 251
and informal ties 70-2 open source software 58
and industry-university links 93-4 openness:
and innovation 20,151 and innovation 10 -12
and interfirm collaboration 57-8 and systems 13
 in d e x 647

O'Reilly* C A, and ambidextrous and productivity 8, 515, 588,590

organizations 136 ,14 1114 and Second Industrial Revolution 358, 360
organic organizations 118 ,119 and service sector 447-8
organization: and technological change 133,135,136, 590
and the innovating firm 29, 30, 33 organizational memory n, 17,20,124
and the managerial revolution 38-9 organizations:
see also organizational innovation and regional innovation systems 299
Organization for Economic Cooperation and and systems o f innovation 188,196
Development (OECD): Original Equipment Manufacturing (OEM) 533
and diffusion 602 Oslo Manual 162,165, 616
Frascati Manual 151* 153-4, 410, 616 Oster, S M 476
and high-tech industries 407-8 O’Sullivan, Mary, and social conditions of
and industrial classification 155, 409, 429-30 innovative enterprise 33-4
innovation Manual 150 Oticon 130,131
and innovation policy 602 Own Brand Manufacturing (OBM) 533
Oslo Manual 162,165, 616 Own Design and Manufacturing (ODM) 533
and science, technology, and innovation policy ownership:
603-4 and control 38
and systems o f innovation 194 cooperative 601
and universities 215-19 and cross-shareholding 42
organizational innovation 7, 2 0 ,115 -16 ,138 -4 0 and strategic control 50
ambiguity of term 138,140 m
and catch-up 515 Paley, William 495
definition 115 paradigms:
and developing countries 563 techno-economic 248, 249, 569
diffusion of 617 technological 416-17,497-8
and economic growth 488 Paris International Convention for the Protection
and employment 573, 576, 587-8, 590 o f Industrial Property 270, 271, 273
Japan 520 Pasinetti, L 580
and organizational change 116 ,133,138 -9 Patel, P 22G, 329
adaptation 116 ,134 -5 Patent Act (USA, 1836) 269
incremental view o f 133,134-5 Patent Cooperation Treaty (1978) 271
and inertia 133,134,136 patents:
punctuated equilibrium 133,135-6 and catch-up 275, 284
strategic adaptation 133,137-9 and collaborative networks 65-6,79
and organizational learning 116 ,123-4 ,138 and diffusion 279, 280
adhocracy model 127-9, 130, 131-2 duration o f 279, 280
and collective learning 124-7 and economic growth 284
institutional context o f 132-3,134 and economic progress 284
J-form model 127,128 and Tvergreening’ 287 ni8
Japanese model 129 and First Industrial Revolution 352
learning models 127 history o f 268
spaghetti organization 131 international patent era 271-3
and organizational structure 116 ,117 ,13 8 multinational patent era 270-1
ambidextrous organizations 119 ,13 6 ,14 1Щ national patent era 267-9
archetypes 119 ,12 0 pre-patent era 267
and contingency theory 117-21 pro-patent era 273-8
and differentiation 119 and industrialization 284
and governance models 122 and information disclosure 280
integration 121-2 and innovation 280-1
internationalization of 319-20
mechanistic organizations 118 ,119
and measurement o f innovation 65, 66,69* 79>
models o f 127
organic organizations 118 ,119 158-60, 616
and multinational enterprises 325-b
and strategy 121-2
648 INDEX

patents: (cant'd) policy, see science, technology, and innovation

and national innovative capacity 197 policy
origins o f term 2S5 щ Pollard, S 356
and regional concentration 558 Pomeranz, Kenneth R 351
role o f 278 Porter, M E 10 7 ,4 18 ,4 19
scope o f 279-80 and national innovative capacity 197, 552-3
and Second Industrial Revolution 364 Portugal, and catch-up 525-34, 537-8
and United States 48, 188-9, 269, 364 Post-Keynesianism, and economic
and universities 80 222, 225, 228, growth 489-90
229-31 Powell, Walter W 67, 78
use o f 281-2 practice:
national differences 283-4 communities o f 7 1,12 5 ,30 3,3 0 8 , 309
sectoral differences 282-3 innovations in 466
variations in laws o f 188-9 networks o f 58,71
see also Bayh-Dole Act; intellectual property prices, and competitiveness 581
rights privatization 395
path dependency: process innovation 7,18 2, 572-3
and innovation 10 ,12 and competitiveness 573
and innovation process 109 diffusion o f 617
and large firms 555 and employment 572-3, 579, 583, 590
and manufacturing firms 95 and Japanese catch-up 519-20
and systems 13 and productivity 572
Pavitt, Keith 3 and sector characteristics 384
and creative accumulation 382 and systems o f innovation approach 185
and firm types 162 product innovation:
and industrial classification 16 ,3 11 m , 384, and collaboration 57, 65, 6 8 ,19 6 -7
4 11-12 compared with process innovation 572-3
Penn World Tables 505 and competitiveness 573
Pennsylvania Railroad 360 definition 7,18 2, 572
Penrose, Edith 32, 543, 560 diffusion o f 617
and Japanese challenge 40 and employment 569, 573-5, 579-80, 583, 590,
and organizational learning 245 592 пз, 593 П9
on theory 51 and forecasts o f commercial success 100
Perez, Carlota 18, 249, 499 impact o f 572, 581
Perroux, F 608 and industrial research and
Perry, Com m odore 272 development 360-3
pharmaceutical industry 555-6 and the Industrial Revolution 352
and coevolutionary processes 397 and innovation surveys 16 0 -1,16 4 -5
and innovation process 97-8, 389-90 and Japanese catch-up 519
and institutions 395 and Josiah Wedgwood 356
and knowledge spillover 297 and low- and medium-tech industries 415-16
and patents 280, 281 and organizational change 137
post-war development 371-2 and organizational structure 117 ,12 7
and research publication 213 and product-life-cycle theory 15
and university research 93, 95, 222 and service sector 435-6, 440, 442, 445
Pharmaceutical Manufacturers Association and systems o f innovation approach 185
(USA) 371 and trust 72
Philippines, and catch-up 525-34, 553-4 and university research 57, 212, 221-2
Fiore, Michael 37 product-life-cycle theory 15-16 , 545
Pisano, Gary 33 and foreign direct investment 327
Poignant, Raymond 603 productivity:
Polanyi, M 75,125 and capital per worker 490
Polaroid 107, 274 and income 544
Polaroid Carp v Eastman Kodak 274 international divergence in 514

labor 550
and organizational innovation 8, 515, 588, 590
and process innovation 572
and research and development 154, 491, 506
and technological change 474» 489, 504
and technology transfer 338
total factor 489, 500, 506, 507, 612
Project Apollo 56
project management 9 1,165
public research organizations 192, 222
funding of 606
industrial links 104» 194, 224,392
and research and development 193
and science-based sectors 384
publishing industry, and networks 78
punctuated equilibrium, and organizational
change 133,135-6
Qualcomm 48
quality control:
and Japanese organizational innovation 43,
520
and service sector 438
quality ladders 416, 502-3
quality requirements, and systems o f
innovation 191
Quebec, and vocational and technical
education 194
QWERTY keyboards, and diffusion 463
radical innovation 7-8, 202, 493-4
and adhocracy 119 ,128,132
and collaboration 65
decision-making about 108
and economic structural change 104
and knowledge application 296
and learning economy 311
Rajan, R 255-6
regional economic development, and
universities 225-7
regional innovation systems 184, 310,312 П5
Anglo-American model 306-7
and competitive advantage 311
and competitiveness 557-9
and diffusion 312115,558
and entrepreneurial regional innovation
system 303-4
and geography o f knowledge 292-4,309-10
and industrial knowledge bases 294-8,310
and institutional regional innovation
system 303, 304
and localized learning 298-300,311
and new economy innovation system 303
origins o f concept 299
INDEX 649
and patenting concentration 558
and regional culture 300
relationship with national innovation
systems 303-8
systemic character o f 299-300
and universities 230
varieties o f 300
regionalized national 302-3
regionally networked 301-2
territorially embedded 300—1
see also national systems o f innovation; sectoral
innovation systems; systems; systems of
innovation
regions:
governance o f 299
importance o f 299
regulatory systems, and diffusion 477
research and development 381-4
and catch-up 530
and collaborative networks 57, 58, 59, 65-6
and development activities 97
and employment 364
and finance 253-4
and government-funded programs 98
and innovation 16
and intellectual property rights 279
intensity o f 155-8,382
and investment 279,323-8, 410, 491, 500, 503,
605
international 323-6,328
and knowledge production 91
measurement o f 1 5 1,153“ 4> 616
and acquired technology 156-8
and research and development
intensity 155-8
modeling o f 502
and multinational enterprises 323-6,337
and asset-augmenting activity 327-8,
329—3 °
and asset-exploiting strategies 326-9,
329-30
and firm size 333
and home-base augmenting activity 327
and home-base exploiting activity 326
host country effects 338
impact on home country 338-9
and integration costs 331
and market structure 333~4
opportunities and constraints 331-3
and organizational issues 334
spillover effects 337-8
origins in Second Industrial Revolution 360-1,
362-3,364
overinvestment in 279
650 INDEX
research and development (c o n f d )
post-war transformation 366-7
and productivity 154,491,506
and public research institutes 193
and service sector 436-7, 448
social rate o f return 491
spillover effects 490-1, 502-3, 507-8
and strategic technology partnering 320-2,
334~6 >339 пз
and systems o f innovation 19 0 ,19 2 -4
and internationalization 332
and uncertainty 10 0 -1
and universities 110 m 3 ,194-5
resource-based theory, and the firm 33 ,15 1
retail sector 427, 439, 445
Reuters 437
Reverse Product Cycle, and services
innovation 440-2
Ricardo, David 571-2,580
Rickne, Annika, and systems o f innovation 189
risk taking:
cultural attitudes towards 478
and networks 68
and science, technology, and innovation
policy 619
Robinson, S 488
Rogers, E M 461-2, 469, 478
Romeo, A A 476
Romer, P 197, 501, 502, 506
Rosenberg, Nathan 613, 620
chain-link model o f innovation 150
and diffusion 460, 462, 477, 479
innovation as continuous process 5
and institutions 186
and linear model 8-9
and machine tool industry 91-2, 417
and national system o f innovation 183,186
and universities 215
Route 128 400
and regional economic development 225
and regional innovation systems 307
Sabel, Charles 37
Sahlman, William 248, 251
Saloner, G 471, 473, 477
Salter, A J 308
Sanmina-SCI 49
SAPPHO project 87
Saxenian, AnnaLee 46, 71, 227, 307
Say, Jean-Baptiste 571
Schmoolder, J 7, 364
Schon, D 125-6
Schreiber, Servan 608
Schumpeter, Joseph:
and business cycle 244
and creative destruction 10 5 ,115 ,35 9
and credit creation 243, 244
and economics o f innovation 241-2
m icroeconomics o f 242-4
and structural econom ic change 244
and entrepreneurs 32 ,2 4 2 -3
and finance 240,244
and innovation 6, 87, 572
temporal clusters 498-9
types o f 6-7
and innovation process 9 -10
and large firms 32 ,2 4 3-4 ,3 6 4
and limits to firm growth 38
and radical innovation 7-8
and technological com petition 14 -15
wave theory 350
science:
and decision-m aking 606
and innovation 375
and invisible colleges 73
and linear model o f innovation 8-9
and Second Industrial Revolution 359-60
see also science, technology, and innovation
policy
Science Citation Index 153
science parks 226-7, 303
science policy, see science, technology, and
innovation policy
Science Policy Research Unit (SPRU) 2 ,3 ,16 1- 2 ,
163
science, technology, and innovation policy 599
Danish cooperative movement 6 o o -i
and developing countries 601, 626
and diffusion 599, 626 m
and economic growth 6 03,604-5,6 08,612,621,
626
and education 606, 610
evaluation o f 618-19
ideal types o f 602
and innovation policy:
American agriculture 602
analytical basis o f 614-15
and competitiveness 612, 618
as economic policy 612
and innovation process 614
and innovation systems 617
and institutions 612, 6 13-14
instruments o f 612-13
laissez-faire version 6 11-12
and neoclassical economics 613
objectives o f 612
and public sector intervention 614
and service sector 451-2
 INDEX 651

systemic version 611-12 and national innovation systems 386

and systems o f innovation 191 and networks 385,391-4
measuring impact o f 615-17 policy implications o f 398-400
normative principles o f 618 research opportunities 401-2
research opportunities 625-6 and sectoral boundaries 385,387-90
and science policy 604-7 sectoral differences in innovation 381-4
funding allocation 606 and technological regimes 382-3
objectives 605 see also national systems o f innovation;
origins o f 604-5 regional innovation systems; systems;
post-war development 605 systems o f innovation
and progress 605 sectors:
public sector actors 606 and appropriability 382, 384
quality o f research 607 characteristics o f 408-9
research evaluation 606-7 classification o f 116,155, 409, 411-14, 429-30
resource allocation 605 definition 385, 408
and scientific autonomy 605-6 differences in innovation 381-4
and technology policy: and employment 580
and catch-up 608 and industrial dynamics 382
and competition 609 and industry life-cycle 383
and diffusion 465 and market structure 382
and economic growth 504 and research and development intensity 382
English cannon production 600 sources o f innovation 384
and European Union 155» 610, 621-3 technological profile 409-10
evaluation o f 6 10 -11 and technological regimes 382-3
extent of support 609 and technology policy 607-8, 610
focus o f 607-8 technology suppliers and users 383-4
and industrial complexes 610, 619-20 semiconductor industry:
instruments o f 610 evolution o f 44—5
Japan 608, 609, 610, 620-1 and networks 67,69
and national interests 608 offshore research and development
objectives o f 609 investment 328-9
old and new industries 609 post-war development 367-8
public sector actors 610 responding to technological change 106
and public sector competence 609 and role o f military 393-4
and small and medium-sized sectoral structure 393-4
enterprises 520-1 United States 45-6
and state intervention 608-9 service sector:
and strategic technologies 608 and catch-up 535
United States 609, 610» 619-20 classification of:
variable meaning o f 608 human services 443
and universities 602, 605, 606, 609, 6io, 613, information services 443-4
614, 624 physical services 442-3
and US-Europe-Japan relationship 624-5 diversity o f 434-5
sectoral innovation systems 183-4, 200, 380-1, and employment in 433-4* 449* 45°
384—7* 400—1 growth o f 433-4* 451
and actors 385, 390-1 industrialization o f 437~9
aggregation issues 387, 400 and ineffective learning 446
and demand 391 innovation in 433,451-2
and acquired technology 448
and diffusion 381,384, 39h 392~3* 399
dynamics and transformation o f 395-8 and client relationships 441-2
and evolutionary theory 386-7 and diffusion 617
geographical dimensions 399, 400 and employment 580, 588
and institutions 385, 394-5 and firm size 447
and information sources 448-9
and knowledge 385,387-90
652 INDEX

service sector: (confd) and finance 255

and information technology 439-45, 448
nature o f 446
organizational 447-8
and quality control 438
research on 436-7
and Reverse Product Cycle 440-2
sectoral differences 449
sources o f 448
and innovation systems 445,449
innovations surveys 447-9
and investment 436-7
and knowledge-intensive business
services 449-51
market diversity 435
and modularization 438
and organization o f innovation 445-6
and organizational innovation 447-8
product characteristics 435-6
and research and development 436-7, 448
and technology use 434-5
Shapiro, C 472, 473
Share (IBM user group) 73
Shaw, K 474
Shell 365
Shell, K 491
Shepard, A 471
Shin, Jang-Sup 517, 524
Shuen, Am y 33
Siemens 323, 361, 363
Siemens, Werner von 361,362
Sierra, C 329
Silicon Valley 46, 48, 400, 624
and adhocracy 127,129
and informal networks 71
and organizational learning model 130
origins o f 227
and regional economic development 225
success o f 307
and venture capital 251
Silverberg, G 497, 500
Simon, Herbert 17
Simonetti, R 582
Singapore:
and catch-up 518, 520-1, 525-34, 553-4
and foreign direct investment 521
and universities 216
skills:
and diffusion 624
and employment 584-7, 590
and organizational innovation 588
Skoda 418
small and medium-sized enterprises (SMEs):
and competencies 335
and government policy 428
and regional innovation systems 300-2
and research and development 333
responsiveness o f 420
and technological change 420
Small Business Adm inistration database
(USA) 161
Smith, Adam 89,107, 489, 534> 543>57*
social capability, and catch-up 523, 524
social conditions o f innovative enterprise 29,
30-4 ,3 4 9
and industrial districts 35-7
and Japanese model 41-5
the managerial corporation 38-40
New Econom y m odel 45-9
and organizational innovation 132-3
social exclusion, and services
industrialization 438
Soete, Luc 603
software industry 247, 278, 369
and coevolutionary processes 397-8
and diffusion 398
and institutions 395
sectoral structure 394
Solectron 49
Solow, R M 489, 500, 501, 505, 547
Sony 45
Soskice, D 305-6
Soule, S A 478
South Korea:
and catch-up 518, 520-1, 522, 525-34, 553-4
economic development 366, 373, 375
financial crisis in 522
and research and development 193
and technology policy 608
and universities 216, 218
and vocational and technical education 195
spaghetti organization 131
Spain:
and catch-up 525-34
and collaborative networks 197
and research and development 193
spillover effects:
and biotechnology industry 297
and economic growth 490-1
and foreign direct investment 338, 508
and innovative capacity 552-3
and knowledge 297-8
offshore research and development
investment 337-8
pharmaceutical industry 297
research and development 491-2, 502-3, 507-8
and universities 225
 INDEX 653

Stalker, G M, and organizational types 118 ,119 and path dependency 13

Standard Oil 365 spatial 12-13
standards: see also national systems o f innovation;
battles over 472 regional innovation systems; sectoral
and diffusion 471-2 innovation systems; systems of
technical 68, 73,191, 395 innovation
Stanford University 46 systems o f innovation (SI) 181,182
Starbuck, W C, and adhocracies 128-9 activities in 189-91
startups, and networks 66-7 boundaries o f 198-200
status, and diffusion rates 470 and competence building 191-2, 617
Statute of Monopolies (1623) 269 components o f 188-9
steam power, and First Industrial Revolution 353 and definition of system 187-8
steel industry, and networks 69, 71, 78 and diffusion 183,184,186,189,190, 200, 201
Stern, S, and national innovative capacity 197 emergence and development o f 182-4
Steuart, James 571 functions o f 189-90
Stevenson, Howard 248 and innovation policy 617
stock options, and New Economy model 46, 47, and institutions 188-9
49 and activities 196
Storper, M 298 relations with organizations 197-8
Strang, D 478 and internationalization of research and
strategic technology partnering (STP) 320-2, development 332, 337
334-6,339 пз and learning 184-5,191-2
structural holes, and networks 61, 62, 66, 69 competence building 194-5
subcontracting, and networks 62, 69-70 research and development 192-4
Sun Microsystems 49 national systems 183,198-9, 200
Sutton, J, and industrial classification 412-14 and organizations 188
Svennilsson, Ingvar 603, 626 and activities 196
Swann, G M P 447, 448, 450 interactions amongst 196-7
Sweden: relations with institutions 197-8
and collaborative networks 197 regional systems 184,199-200
and national system o f innovation 199 research gaps and opportunities 201-3
and patent laws 188 and resource allocation 248-9
and research and development 188,193 sectoral systems 183-4
and technological specialization 557 strengths o f 184-6
and universities 193, 217, 219, 232 technological systems 183, 200
and vocational and technical education 195 weaknesses o f 186-7
Switzerland 428, 429 see also national systems o f innovation;
patent laws 270-1 regional innovation systems; sectoral
and research and development 193 innovation systems; systems
and technological specialization 556-7
and university funding 193 Taiwan:
and use o f patents 282 and catch-up 518, 520-1, 525-34, 553-4
and vocational and technical education 194, economic development 366,373
195 and research and development 193
Syrquin, M 488 and technology policy 608
system integration 422-3, 451 Takahashi, K 272
systems: Takeuchi, FI 125,127,295
bottlenecks in 13-14 Tancioni, M 582
categorisation o f 12 Taylor, K 582
and complementarities 13-14 ‘Technical Change and Economic Policy
definition 187-8 (OECD) 603
and innovation 20 technological change:
nature o f 13 and ambidextrous organisations 119
openness o f 13 disequilibrating nature o f 569
654 INDEX

technological change: (confd) Tether, B S 16 9 ,447,448,450

and economic growth 255, 488-92 Teubal, M 189
and employment 569-71* 580* 582* 588,589, Texas Instruments 46,274
590-1 textile industry 421
and globalization 318 and First Industrial Revolution 352-3
and institutions 394 innovation in 425-6
and organizational change 133* 135* 136,590 and vertical alignment 423
and productivity 474* 489* 504 Thailand 441
and small and medium-sized enterprises 420 tire industry 413
and uncertainty 494 tobacco industry 425
technological congruence 523* 524, 535 Toivonen, M 451
technological regimes 17, 382-3, 579 Tordoir, P P 441
technological search 9 6 ,9 7,10 4 Toshiba 43
technology: total factor productivity 4 8 9 ,50 0 ,50 6 ,50 7,6 12
access to foreign 533 Toyota 43,45
acquired 156-8 and lean production 420
as change driver 4 16-17 and networks 70
and convergence 91-2, 99 and vertical alignment 422
critical 349-50 Trade and Industry, Department o f (UK) 437
cross-border commercialization o f 319-20 Trade-Related aspects o f Intellectual Property
cycle o f progress 136 Rights (TRIPS) 273, 274, 275, 284
and economic growth 350,492 trade unionism 36, 40, 42-3
general purpose 416-17, 504-5 training:
and managerial corporation 39-40 and competence building 19 2 ,19 4 -5
and organizational change 139-40 and competitiveness 553
and organizational practices 102-3 internal 390
as private good 489 and labor quality 589
as a public good 489 see also education; universities
and radical change 104-7 "Triple Helix’ 221
suppliers and users o f 383-4 and regional innovation systems 300
see also science, technology, and innovation and role o f universities 214, 233
policy trust;
technology gap: and innovation systems 332
and economic growth 19, 547 and networks 58, 65,70, 72
and technology transfer 338 and tacit knowledge 293
technology policy, see science, technology, and Tunzelmann, Nick von 353
innovation policy Tushman, M L 11, 68, 7 2 ,1 3 6 ,1 4 1 П4
technology transfer 8, 228
and Bayh-Dole Act 229-31 Uchupalanan, K 441
global 319-20 Udell, G 254
and productivity 338 uncertainty:
and university-industry collaboration 210,399 and diffusion 466,475-6
see also diffusion and innovation 9 ,10 , 30, 88, 493, 613
technopoles 303 and innovative investment 257-8
tech no structure 40 management o f 10 0 -1
Teece, David 33 ,12 2 ,12 6 , 336, 536 and networks 68
telecommunications industry: strong 493
and re v o lu tio n a ry processes 397 and technological change 494
Japan 621 weak 493, 502, 504, 509
liberalization 395 unemployment;
United States 370, 477 and innovation 571
Tellis, G J 465, 467, 470, 478 and technology 571-2, 581-3,590
Terman, Frederick 226 see also employment, impact o f innovation
Tesco 107 Union Carbide 365

United Kingdom:
and catch-up 525-34
and education 194, 554-5
and industrial districts 35-7
and patent laws 269
and research and development 323
and science policy 604, 607
and service sector research and
development 436-7
and technological specialization 555
and technology policy 608
and universities 210, 216, 218, 219, 220, 232
and vocational and technical education 195
see also Industrial Revolution, First
United States:
and agricultural innovation 602
and catch-up 515,525-34
chemical industry 364-6
and comparative advantage 555-6
Court of Appeals for the Federal Circuit 228-9,
274, 277
Department o f Justice 274
diffusion o f major innovations 468
and education 554-5
and employment 581
and government-funded programs 98
and industrial research and development 90
and innovation policy 624
and intellectual property rights 274-5, 276-7,
278
the managerial corporation 38-40
the managerial revolution 38
and the New Economy 45-9, 277
and organizational innovation 7
patent laws 188-9, 269,364
patenting activity 48
and research and development 188,193
and science policy 604-5
and Second Industrial Revolution 358,359-60,
363-4
and service sector research and
development 436
Small Business Administration database 161
and technological specialization 555-6
and technology policy 609, 610, 619-21
and Third Industrial Revolution 366-73
and universities:
and Bayh-Dole Act 228-32
and co-authorship 220
degreeholders 215-16
and employed researchers 217-18
enrollment 215
funding o f 210
impact on industrial innovation 221-4
INDEX 655
industrial funding 218
industrial links 209-10, 219
and regional economic development 225—7
and research and development 216, 217
and venture capital 251-3
and wage polarization 587
United States Patent Office 269
United States Supreme Court, Diamond
v Chakrabarty 229
universities:
analysis difficulties 233-4
and Bayh-Dole Act 228-32
and catch-up 518
cross-national comparisons:
degreeholders 215-16
and employed researchers 217-18
enrollment 215
industrial funding 218
industrial links 219
and research and development 216, 217
and diffusion 211-12
and economic development 209, 210
and regional development 225-7
and economic growth 211, 212, 232
impact on industrial innovation 221-4
industry links 93-5,110 П14,194, 209-10, 2U,
224-5, 247
and co-authorship 219-21
lack o f data on 234
and intellectual property rights 606
and inter-university competition 233
and linear model o f innovation 212
and Mode 2 research 213-14, 221, 233
and national innovation systems 210, 211-15
origins o f 235 n8
and origins o f industrial research 360-1
outputs o f 212
and patents 80 пз, 222, 225, 228, 229-31
and regional economic development 225-7
and regional innovation systems 210
and research and development 110 m 3 ,194-5
and research disclosure norms 213, 230-1
role o f 209, 210, 211-12, 233, 234
and science parks 226-7
and science, technology, and innovation
policy 602,605,606,609,610,613,614,624
and scientific autonomy 606
and Second Industrial Revolution 360-1
and technology transfer 210
and 'Triple Helix’ perspective 214, 221, 233
Uzawa, H 491
Valentine, Don 251
Van de Ven, A 12
656 INDEX

Van Pottelsberghe, B 508 Weick, K E 137

Varian, H 473 Wellcome Trust 56
Veblen, Thorstein 516-18 Western Union 360
Vega, M 329 White, Steven 189
Venables, A J 298 Whitehead, Alfred North 350
Venice, and patent code 267 Whitley, J D 582
Venkataraman, S 474-5 Wilson, R A 582
venture capital 135, 248, 250, 251-3 Winter, Sidney 11,17 ,2 0 ,3 8 2
and New Econom y model 46-7 and evolutionary theory 497, 499-500
Verdoom , P J 491, 523 and organizational capabilities 33
Vernon, R 15,32 6 -7 and organizational routines 126
Verspagen, Bart 19, 500,508 Wolff, E 581, 584
and catch-up 524, 535 Womack, J P 127
vertical disintegration 90, 91-3, 99,421 World Bank 534, 535
video recorders, and network effects on diffusion World Intellectual Property Organization
o f 471 (W IPO) 270, 271
Vivarelli, M 581-2 World Trade Organization 535
Vivendi 95 and Trade-Related aspects o f Intellectual
Von Hippel, E 71 Property Rights 275
Wright, Gavin 259,372
wage polarization 583,587,590, 592
W al-Mart 427 Young, A 491
Watson, Thom as J Sr 494
Watt, James 353,494 Zander, Udo 33
wave theory 350, 499 Zeitlin, Jonathan 37
Weber, M ax, and organizational design 117 Zingales, Luigi 255-6, 259
Wedgwood, Josiah 357-8 Zucker, L 297,607
 OXFORD
U N IV E R S IT Y P R E SS
Great Clarendon Street, Oxford 0 x2 6 d p
Oxford University Press is a department o f the University o f Oxford.
It furthers the University's objective o f excellence in research, scholarship,
and education by publishing worldwide In
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
Oxford is a registered trade mark o f Oxford University Press
in the UK and in certain other countries
Published in the United States
by Oxford University Press Inc., New York
© Oxford University Press 2005
The moral rights o f the authors have been asserted
Database right Oxford University Press (maker)
First published 2005
First published in paperback 2006
Reprinted 2006 (Twice), 2007
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
without the prior permission in writing o f Oxford University Press,
or as expressly permitted by law, or under terms agreed with the appropriate
reprographics rights organization. Enquiries concerning reproduction
outside the scope of the above should be sent to the Rights Department,
Oxford University Press, at the address above
You must not circulate this book in any other binding or cover
and you must impose this same condition on any acquirer
British Library Cataloguing in Publication Data
Data available
Library of Congress Cataloging in Publication Data
Data available
Typeset by SPI Publisher Services, Pondicherry, India
Printed in Great Britain
on acid-free paper by
Antony Rowe Ltd, Chippenham, Wiltshire
ISBN 978-0-19-926455-1 (Hbk.)
ISBN 978-0-19-928680-5 (Pbk.)
5 7 9 10 8 6 4