Research Policy 35 (2006) 609–625

Innovation processes in large technical systems: Market

liberalization as a driver for radical change?
Jochen Markard ∗ , Bernhard Truffer
Cirus—Innovation Research in Utility Sectors, Swiss Federal Institute for Aquatic Science and Technology, Eawag,
6047 Kastanienbaum, Switzerland
Received 1 July 2004; accepted 1 February 2006
Available online 11 May 2006

Abstract
Electricity supply is a large technical system, which exhibits strong path dependencies and high barriers for radical innovations.
Recent market liberalization, however, has initiated a fundamental restructuring. The paper analyzes how liberalization has altered
innovation processes in the field of electricity supply. We examine three radical innovations under monopoly conditions and contrast
the findings with the results from a survey on the innovation behavior of electric utilities in liberalized markets. We argue that the
selection environment for innovations has changed in various respects. In our sample, new options open to innovation activities
have emerged at the level of the firm and utility strategies turn out to be more heterogeneous. This may be interpreted as an increase
in the variety of search processes, which may lead to a mid-term decrease of path-dependencies.
© 2006 Elsevier B.V. All rights reserved.

Keywords: Radical innovation; Large technical systems; Market liberalization; Electricity supply

1. Introduction Radical innovations are confronted with consider-

Large technical systems like the electricity supply
system encompass a capital-intensive infrastructure, a
broad range of technical components and technologies
and a variety of actors and institutions. Most system
components are closely interrelated and various kinds
of technical norms, organizational practices and institu-
tional procedures have emerged to guarantee a smooth
joint operation of all these components. As a conse-
quence the development of the electricity supply system
tends to be strongly path-dependent. Innovation pro-
cesses are more of the incremental than of the radical
type (e.g. Hughes, 1987).
∗ Corresponding author. Tel.: +41 41 3492 161;
fax: +41 41 3492 162.
E-mail address: jochen.markard@eawag.ch (J. Markard).
able barriers as they have both to overcome prevailing
standards and to compete against the network external-
ities of established products or technologies. Nonethe-
less, radical innovations sometimes take place in large
technical systems. Such innovation processes may be
triggered by factors or ‘events’ from outside the sys-
tem including new preferences on the political agenda
or technological breakthroughs in other sectors. Radical
innovations may also emerge in the course of contin-
uous expansion and growth of a technical system. If
reverse salients, i.e. frictions in specific parts of the sys-
tem hinder its further expansion, innovation activities
will be concentrated at these points (Hughes, 1987).
Reverse salients that cannot be overcome within the
context of the existing system may bring about radi-
cal innovations and a shift in prevailing technological
regimes.

0048-7333/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.respol.2006.02.008
610 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625
In the field of electricity generation, we can in fact
observe several cases, in which radically new technolo-
gies were driven by external stimuli and government
action in order to overcome reverse salients. For instance,
nuclear or wind power were supported by policy instru-
ments in the light of public debates about the security of
supply in the aftermath of the oil crises and a reduction
of environmental pollution (Hadjilambrinos, 2000). The
diffusion of combined cycle gas turbines as baseload
power plants represents another example for a radical
technological change in electricity supply. Here, sig-
nificant innovation impulses can be traced back to the
advances in gas turbine technology in the aircraft indus-
try (Islas, 1997; Patterson, 1999, p. 73). The main driving
factors can be classified as external in all three cases
because the internal resistance of the electricity supply
system confronted with these innovations proved to be
considerable (Jacques et al., 2001; Patterson, 1999; Sine
and David, 2003).
The recent introduction of market liberalization in
the electricity sector may also be interpreted as an exter-
nal stimulus to the electricity system. Electric utilities,
which used to operate on the basis of regulated, territorial
monopolies over many decades, now have to compete for
customers in the fields of power generation, trading and
sales. However, unlike the before mentioned driving fac-
tors, the impact of market liberalization has no specific
impact on a certain technology but changes the selec-
tion criteria in a general way. More importantly, market
liberalization has a considerable impact on the organi-
zational structures of electricity supply (Joskow, 1998;
Sioshansi, 2001a). As a consequence, we may expect
that forms of resistance towards radical innovations are
likely to change under the new market rules: new play-
ers enter the market, new roles are assigned to electricity
consumers and new rules for investment decisions and
technology choices emerge.
This paper addresses the question how market liberal-
ization has altered the way innovations are handled in the
electricity supply system. Our focus is on radical innova-
tions and we ask whether recent changes have increased
the options open to innovative activity, thus broadening
the range of development paths pursued and decreasing
the path dependency of the sector. On a conceptual level,
the article contributes to the analysis of particularities of
radical innovation processes in large technical systems.
In the field of science and technology studies there
are several studies dealing with the basic attributes of
large technical systems (Gökalp, 1992; Hughes, 1987;
Joerges, 1998) but only few scholars have explicitly
addressed the characteristics of innovation processes
and radical innovations (Davies, 1996; Godoe, 2000).
With our analysis, we want to enrich this body of lit-
erature. We furthermore relate our study to economic
analyses of technological change and radical innova-
tions (Mascitelli, 2000; Teece, 1996; Tushman and
Anderson, 1986). We particularly refer to lock-in phe-
nomena, the path-dependency of innovation and the
notion of technological regimes, which guide innova-
tion processes (David, 1985; Dosi, 1988; Rip and Kemp,
1998; Rosenberg, 1982).
The paper is structured as follows. Section 2 shortly
presents the theoretical concepts of large technical sys-
tems and technological regimes on which our analysis is
based. In Section 3 we introduce the electricity supply
system and discuss how various kinds of interdependen-
cies and standards impede radical innovations. Section
4 elaborates on three radical innovations in the field
of power generation. On this basis, we examine key
drivers for radical innovation under monopoly condi-
tions. Section 5 deals with the consequences of market
liberalization for innovation processes. It starts with an
analysis of the transformation of the selection environ-
ment for innovations. Then, electric utility innovation
strategies and the underlying motivations are illustrated
on the basis of a case study on stationary fuel cells. Sec-
tion 6 concludes that market liberalization has increased
the scope of variation for innovation activities at the level
of the firm and of the sector. However, it remains to be
seen whether these opportunities, in fact, lead to radical
technological changes or not.
2. Particularities of innovation processes in large
technical systems
2.1. Key characteristics of large technical systems
Technical systems like infrastructures for the supply
of energy and water, railway systems, telecommu-
nication networks or military defense systems show
similarities in their structure and in their innovation
characteristics. In the literature on science and technol-
ogy studies, the notion of large technical systems has
been developed together with a conceptual framework to
describe the general features and the dynamics of such
systems (e.g. Gökalp, 1992; Hughes, 1987; Joerges,
1998). Large technical systems can be defined as
“. . . complex and heterogenous systems of physi-
cal structure and complex machinery which (1) are
materially integrated, or ‘coupled’ over large spans
of space and time . . . and (2) support or sustain the
functioning of very large numbers of other technical
systems . . .” (Joerges, 1998, p. 24)
 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625
Although this definition emphasizes physical com-
ponents and machinery, large technical systems also
encompass non-physical artifacts like specific regula-
tions and norms as well as various kinds of actors such
as utility companies, network operators, manufacturing
firms or investors. One of the key attributes of a large
technical system is the high degree of interdependen-
cies between the various components. If one element
is altered or even removed from the system, other ele-
ments in the system and the overall system characteristics
will change accordingly (Hughes, 1987). Similar but
weaker dependencies also exist between the system and
its ‘outside’ world as large technical systems often rep-
resent the basis for other technical and social systems.
Large technical systems are also characterized by dif-
ferent kinds of standards including technical and social
norms, organizational practices or patterns of use. These
have co-evolved with the system and assure the compati-
bility and inter-operability of its numerous components.
The high degree of interdependence and the existence
of powerful standards represent obstacles for a trans-
formation of the system. A novel product, for instance,
must not only meet the consumer needs but it must also
be compatible with the existing infrastructure and with
technical norms. An innovative technology may be an
improvement in one part of the system but may have
a negative impact on other parts. Innovations in large
technical systems, therefore, tend to be incremental in
nature and existing products and technologies undergo
processes of slight, continuous improvement rather than
radical change. In other words, large technological sys-
tems are characterized by a high degree of stability and
inertia.
2.2. The influence of regimes on innovation
processes
Such situations, in which innovation processes are
mainly incremental and path-dependent, have received
much attention in the economic literature on innova-
tion (David, 1985; Dosi, 1988; Rosenberg, 1982). Dosi
explains the path-dependency of innovations with the
influence of technological paradigms, which are pre-
vailing models for the solution of techno-economic
problems. A paradigm is based on a technological arti-
fact that is to be developed and improved and a set
of search heuristics or engineering practices shaped
by selected principles from natural sciences, prevail-
ing beliefs of technicians and customer expectations
(Dosi, 1988, p. 224; Kemp et al., 1998). Technological
paradigms guide the direction of search, thus keeping the
degree of technological variation within certain bound-
611
aries. Prevailing paradigms represent a barrier for radi-
cally new solutions in a very early phase of the innovation
process.
Similar to the concept of paradigms, other schol-
ars have proposed the notion of technological regimes
(Nelson and Winter, 1982; Rip and Kemp, 1998; van
den Ende and Kemp, 1999). This concept also takes into
account the innovation guiding effects of established rou-
tines and skills on the user side and of the broader social
and institutional context a technology is embedded in.
“. . . a technological regime is defined as the complex
of scientific knowledge, engineering practices, pro-
duction process technologies, product characteristics,
user practices, skills and procedures, and institutions
and infrastructures that make up the totality of a tech-
nology.” (van den Ende and Kemp, 1999, p. 835)
In evolutionary economic terms, the regime addresses
the boundaries of variation at the level of innovating
firms or research institutes as well as the prevailing selec-
tion rules in the market and in the broader social context
(Kemp et al., 1998).
Within a large technical system different regimes
may exist. In the field of electricity supply, technolog-
ical regimes have developed, for example, all around
hydropower, nuclear power or fossil fuel-based power
generation. On a more aggregated level, the dominant
structure of centralized electricity generation in large
power plants together with long-distance transmission
lines also represents a regime, which may be referred
to as sector regime. Technological regimes may either
equally co-exist within a large technical system, e.g.
with similar market shares, or there may be a dominant
regime and niches, respectively. Regimes co-evolve with
the corresponding technology, i.e. they become more
powerful and rigid as a technology matures and diffuses.
In the case of wind energy, for instance, the underly-
ing technological regime is still of minor importance but
growing. Wind power is all the more interesting as it
leads to a competition of regimes in the field of power
generation and on the level of the sector because, like
other distributed energy sources, wind power is largely
incompatible with the dominant regime of centralized
generation.
2.3. Origins of radical innovations
Facing the broad set of factors that favor estab-
lished technologies and products, radical innovations
can only become competitive if they get the opportu-
nity to develop and improve in a niche. Such a niche is
an application context, in which the new product or tech-
612 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625
nology is temporarily protected from the standards and
selection rules of the prevailing regime (Hoogma et al.,
2002; Kemp et al., 1998). Niches can be created by well-
resourced economic actors but also by policy makers that
have the goal to introduce a novel technology. Further-
more, market niches may emerge due to the demand of
specific customer segments or particular application con-
texts in which a novel technology might be superior to
the established technology.
Apart from the general resistance to change, inno-
vation processes in large technical systems also have
certain particularities with regard to the focus of inno-
vation. In his analysis of the development of the elec-
tricity sector, Hughes observed that the system showed
a tendency to continuously expand in space and scope
(Hughes, 1987). Whenever this expansion, or the steady
increase of system performance, was hindered by a lim-
itation of certain (technical or non-technical) system
components, innovation activities concentrated on these
bottlenecks or ‘reverse salients’ in order to overcome
the limitations. Reverse salients may lead to incremen-
tal as well as radical innovations. If no solution can be
found within a prevailing technological regime, reverse
salients may even trigger the emergence of a new regime.
This was the case in the early days of electricity sup-
ply when the prevailing direct current (DC) technology
was not able to transmit ever more growing energy
flows efficiently over long distances (Hughes, 1987).
The result was a regime shift to alternating current (AC)
voltage.
In many other cases, however, radical innovations
in large technical systems were triggered by external
factors or contingent outside events. These include tech-
nological developments or scientific breakthroughs in
related fields of knowledge. The transfer of computer
technology and digitalization to the field of telecommu-
nication, for example, not only led to a radical change of
switching technology but also affected the topology of
telecommunication networks and the range of products
and services (Davies, 1996). Similarly, a fundamental
change of preferences in the public, and in the political
domain, respectively, can trigger far-reaching innova-
tions in large technical systems (cf. Section 4).
However, the key role of external events for radical
innovations should not be interpreted as an anything-
goes claim for radical transformations of large techni-
cal systems. Prevailing technological regimes are strong
and well-established firms tend to oppose far-reaching
innovations as these might endanger their market posi-
tion. When assessing radical innovation processes, we
therefore have to analyze the interplay between external
impulses and internal structures carefully.
2.4. Proposition of a concept to assess the
“radicality” of innovations
In order to carry out the intended analysis on radical
innovations, we have to explicitly define the concept of
“radicality”. There are a number of different classifica-
tion schemes of innovations in the literature (Abernathy
and Utterback, 1978; Christensen, 1997; Henderson and
Clark, 1990; Tushman and Nadler, 1986). A widely
used approach is the distinction between incremental and
radical innovations. Incremental innovations are minor
changes in existing products, whereas radical innova-
tions represent an entirely new class of products or tech-
nological devices based on a novel set of engineering
and scientific principles.
Such a classification, however, seems to be over-
simplified for several reasons. First, innovations can
be distinguished along different dimensions like price
or performance attributes, competences required on the
producer or the consumer side, or their potential to
substitute existing products or technologies. Second, in
practice we find rather a continuum of differences than a
clear-cut dividing line. Third, any classification depends
on the perspective of the one who undertakes it, i.e. a
novel process technology may barely affect the end user
of the respective product, while representing a funda-
mental challenge for the producers. In a more general
way, we may say that the degree of “radicality” can vary
along the value chain. This effect is particularly impor-
tant for highly interrelated technical systems.
We therefore propose a novel concept, which allows
a gradual, qualitative assessment of radical innovations
along two dimensions: the degree of change along the
value chain (vertical novelty) and the degree of change in
a single element of the value chain (horizontal novelty).
Radical innovations may be considered to score high on
either of the two dimensions or even on both.
In the field of electricity supply the value chain can
be conceptualized on the basis of five major functions:
the exploitation of primary energy carriers, their trans-
port, their conversion into electricity, power transmis-
sion and distribution as well as trading and marketing.
Furthermore, we will assess how innovations affect the
consumer side in terms of power consumption behav-
ior, cf. Fig. 1. For each field we ask how far-reaching
the innovation effects are and in what respect the inno-
vation differs in comparison to established technologies
and products.1
1 As the innovation effects vary with the market share of the new
product or technology, we have to assume a similar degree of diffusion
in order to compare innovations in different development stages.
 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625 613

Fig. 1. Realized and potential changes in the traditional value chain of power supply due to the introduction of new generation technologies.

In the following, we will apply this approach to case
studies in the field of power generation. With respect
to monopoly conditions, we will present nuclear power,
combined cycle gas turbines and wind power, for which
we review the innovation processes and the respective
drivers on an ex-post basis. For the liberalized market
environment, changes are too recent for ex-post stud-
ies and we cannot refer to the former cases because
they already established stable technological regimes in
monopoly times. We therefore review the general dif-
ferences in the selection of innovations. In addition,
innovation strategy patterns are presented in a fourth
case, stationary fuel cells. This innovation is still in a
very immature state of development but it can, nonethe-
less, illustrate the changes in strategy making related to
radical innovations.
However, before we turn to these analyses we have
to introduce the basic characteristics of the electricity
supply system.
3. Electricity supply as a large technical system
The large technical system of electricity supply can
be regarded as a set of different actors, institutions and
technical components and their relationships that serve
the purpose to supply consumers with electrical power.
The actors of the electricity supply system include
power producers, electricity trading companies, power
marketers, grid operators, primary energy suppliers,
equipment manufacturers and the like. Institutions
encompass sector specific regulations, technical norms
but also power exchanges, research laboratories and
614 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625
third-party organizations like independent system opera-
tors or price regulators. Technical components comprise
generators, transformers, power lines, switches as well
as control and metering devices. Finally, the broad
range of power consuming appliances including motors,
lighting equipment or heaters may be regarded as the
boundary elements of the system.
The system is embedded into an environment, which
includes various types of electricity consumers, stake-
holders, NGOs and policy makers but also societal norms
and expectations, political preferences, macro-economic
conditions, geographical particularities and the like.
The elements of the system are interdependent in
many ways. Most technical components are physically
connected with one another. Although the specific tex-
ture of these interconnections varies over time, e.g. due
to the switching of loads and power plants, a proper
function of this huge number of components at every sin-
gle moment in time is required to assure the stability of
the system. Due to the close physical interrelation local
failures might have far-reaching consequences: Recent
experiences with power blackouts in the US North West
(Burns et al., 2004; White et al., 2003) or in Italy showed
that local failure events may cause a chain reaction that
soon affects a large geographical area.2 Furthermore, the
negative impacts go far beyond the system itself as most
loads are not designed to work in a temporary stand-alone
mode. A breakdown of the electricity system thus leads
to a breakdown of dependent transportation systems, IT
systems, industry production lines, public lighting and
even water supply.
The elements of the electricity supply system are
also technologically interrelated. Within the system,
most technologies have been adapted to one another.
Manufacturers of power plant and metering equipment,
for instance, have designed their products in a way
2 The nationwide blackout in Italy on 28 September 2003 may ulti-
mately be traced back to a tree flashover and the subsequent breakdown
of the Lukmanier transmission line in the Swiss Alps. Thus, a second
power line close by had to carry a significant overload. The whole
situation was even aggravated by the fact that Italy imported 300 MW
more than the agreed upon schedule of 6400 MW at that time. Although
these 300 MW could be reduced 20 min after the first event, the addi-
tional load in the second line had caused a sag (due to overheating).
This caused another tree flashover and a breakdown of the second
line. Now the overloads became intolerable for the remaining lines
and several emergency switches disconnected Northern Italy from the
European network. This again led to instabilities and low voltage lev-
els. Several power plants were automatically disconnected. Although
several countermeasures like load reductions were taken, the frequency
drop of the network could not be stopped. 2.5 min after the separation
from the European network the complete power supply system in Italy
collapsed (UCTE, 2004).
that they meet the requirements of power producers,
marketers and grid operating companies. Similarly,
electric equipment on the demand side has to match
the (national) particularities of electricity supply
including the voltage level, frequency or the type of the
socket.
Apart from physical and technological interdepen-
dencies the electricity system is based upon contrac-
tual, legal, economic or informational linkages between
its actors. Firms may be tied to one another by long-
term delivery contracts, shares of financial capital,
joint ventures or informal producer–supplier relation-
ships. Public authorities or grid-operators, for exam-
ple, may impose certain rules on electric utilities but
may as well be subject to the strategic actions of
the utilities (Hirsh, 1999). The players of the system
also interact in different arenas—be it on the market,
in innovation specific working groups, in public dis-
course or in the field of policy decision making and
lobbying.
In the electricity supply system, different kinds of
standards have been established including technical,
operational, organizational and social standards. Tech-
nical standards play an important role for the physical
interconnection of technical components and machinery.
The fixed AC frequency of the grid together with a few
determined voltage levels may serve as a prominent
example of such a technical standard. Operational
standards assure the inter-operability of the technical
system components. This includes the principle that
power generation always has to follow demand as
well as sophisticated procedures for the management
of grid-wide power flows or the handling of local
breakdowns.
Organizational standards, for instance, determine the
roles or tasks of the various actors in the electricity sup-
ply system. Together with regulatory standards, they
also assign obligations, liabilities as well as specific
rights like for instance exclusive supply under monopoly
conditions. Social standards, finally, have emerged as a
consequence of societal embedding of electricity supply.
They shape patterns of use as well as customer expec-
tations, e.g. that electricity has to be available at every
place, all the time, at a high quality and low costs. Social
standards are closely related to main stream convictions
or beliefs.
Apart from its high degree of interdependence and
standardization, electricity supply is characterized by an
infrastructure that requires sizeable capital investments.
Over the calculated life span, the investments into power
plants or the transmission network represent sunk costs.
In the case of large, central thermal power stations, for
 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625
instance, this life span may range between 20 and 40
years. Thus, incumbent actors like electric utilities tra-
ditionally had no or little interest to get involved with
radical innovations, which are likely to jeopardize their
long-term investments. Instead, the companies tended to
prevent any change in institutional frame conditions and
to stabilize prevailing technological paths by incremen-
tal improvements.
4. Radical innovations in electricity supply
under monopoly conditions
Despite the obstacles discussed above, a number of
radical innovations could be observed in the electric-
ity supply system under monopoly conditions. Here we
will present three technological innovations in the field
of power generation. The first section elaborates on the
development and diffusion processes. It also includes an
analysis of the degree of radicality, cf. Section 2.4. The
second part concentrates on the identification of the var-
ious drivers of innovation.
4.1. Examples of radical innovations in electricity
generation
4.1.1. Nuclear power
The use of nuclear power for electricity generation
became technologically feasible in the 1950s and was
first pushed by politicians and national government agen-
cies in the US, the UK and France (Hadjilambrinos,
2000; Patterson, 1999). Nuclear power promised to
become a viable alternative to electricity generation on
the basis of coal and oil. Moreover, certain types of
nuclear power plants could be used to produce plu-
tonium, which provided a basis for nuclear weapons.
Electric utilities initially opposed the introduction of the
new technology as they were familiar with the fossil fuel-
based alternatives and concerned about the high capital
costs of nuclear power (Patterson, 1999, p. 62). By the
late 1960s, however, purchase orders for new nuclear
power plants had reached some significance and after
the first oil price crisis in 1973 policy makers eagerly
increased the support for the new technology in many
countries, cf. Section 4.2.
In the context of nuclear power, a new regime
emerged, which was fundamentally different compared
to existing technological regimes. The technological
principle to generate steam from nuclear fission had not
been applied so far and required a broad range of auxil-
iary technologies, e.g. for safety issues, cooling purposes
or process control. Moreover, a novel concept of fuel pro-
vision including exploitation, transport, pre-processing
615
and waste handling had to be established.3 All these tasks
called for a large body of new scientific and engineering
knowledge, which was developed in specially founded
research institutes and in the R&D laboratories of uni-
versities and private firms. Furthermore, new institutions
were needed for the supervision of the emerging nuclear
industry including the regulation of safety and technol-
ogy transfer issues.
Nuclear power can be considered as a radical innova-
tion with a medium degree of vertical and a high degree
of horizontal novelty. It changed several elements in the
primary part of the value chain and even brought about
new elements, cf. Fig. 1. Moreover, the core technol-
ogy and the auxiliary technologies were fundamentally
different from existing technologies. In many countries,
finally, nuclear caused a significant substitution of oil-
and coal-fired power plants including the corresponding
support products and technologies.4
However, nuclear power also had some similarities
with existing modes of power generation. It incorpo-
rated the established steam turbine technology as well
as generators, transformers and other machinery on the
‘electricity’ side of the power plant. Moreover, nuclear
power plants had a similar size (though they tended to
be larger) as conventional power plants. As a matter
of fact, the emerging nuclear power regime was highly
compatible with the sector regime of centralized power
generation.
4.1.2. Combined cycle gas turbines (CCGTs)
Although the gas turbine was invented at the begin-
ning of the 20th century, it did not play a role in the
field of electricity generation until the 1950s. From then
on, gas turbines could establish a small, but stable market
niche for peak load power generation (Winskel, 2002), in
which power plants are favored that are able to reach full
output within a few minutes from a cold start. The much
larger market share for medium and base load generation,
however, was dominated by the steam turbine, which was
at the heart of a strong and stable technological regime
(Islas, 1999). In the mid 1970s, a new type of power
plants revealed highly promising techno-economic char-
acteristics (Islas, 1997). It was based on a novel combi-
nation of a steam and a gas turbine—whereby the latter
produced the major part of the electricity. CCGT tech-
3 As pre-processing and waste handling became issues of key impor-
tance they can even be considered as separate steps in the value chain.
4 France may serve as an impressive example in this respect. In the
1970, 40% of the electricity was generated on the basis of oil combus-
tion. About a decade later this share had been completely substituted
by nuclear power (Hadjilambrinos, 2000).
616 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625
nology achieved favorable construction times together
with low O&M costs and, in terms of energy efficiency,
it was comparable with contemporary steam turbines.
Moreover, the two turbines could operate independently,
i.e. the full load of the gas turbine was available on short
notice and in case of a breakdown of a single turbine the
other could still be used for backup.
However, technology diffusion was still hampered:
Natural gas was not yet widely available and, as a reac-
tion to the oil price shocks, policy makers in Europe and
the US restricted the use of natural gas for electricity
generation (Winskel, 2002).5 Furthermore, technology
adoption by incumbent electric utilities was not very
enthusiastic—partly because they were afraid to turn
long-term investments into coal and nuclear power plants
into sunk costs, but mainly because energy experts and
decision makers did not consider CCGT to be a viable
alternative for large-scale generation (Winskel, 2002).
So it was not until the mid 1980s that the orders for
CCGT power plants increased significantly. Meanwhile,
the new technology had become more energy efficient
than conventional power plants. In the following years,
the technology developed further and nowadays achieves
a fuel conversion efficiency of nearly 60%—compared
to some 40 or 45% of the best coal-fired power plants.
CCGT power plants also offer favorable investment con-
ditions since they only account for 40% of the specific
capital costs of coal-fired power plants (Sondreal et al.,
2001).
CCGT power plants represent a radical innovation
as they achieve a high degree of vertical and a medium
degree of horizontal novelty, cf. Fig. 1. In technological
terms, CCGTs were not fundamentally different from
existing technologies but rather represented an architec-
tural innovation, i.e. a novel combination of basically
known technical components (Henderson and Clark,
1990). With regard to vertical novelty, the new technol-
ogy has the potential to affect all elements of the value
chain. CCGTs already led to a quite far-reaching sub-
stitution of oil- and coal-fired power plants in several
countries, together with a partial substitution of oil and
coal as primary energy carriers and an enlargement of
the gas supply network. This ‘dash for gas’ has been
particularly striking in the UK (IEA, 2002). Moreover,
the technology may change the upper parts of the value
chain as well. Economies of series production permit
to build ever smaller power generation units, which are
still cost and energy efficient, and natural gas is widely
5 These restrictions were partly or completely removed in the late
1980s.
available due to a sophisticated pipeline infrastructure in
many industrialized countries. As a consequence, small
or medium-sized gas-fired power plants can be operated
close to electricity consumers, which does not only lead
to a decentralization of power supply but also allows for
a local use of the excess heat.
Until today, a technological regime has emerged in
the context of CCGTs, which competes with the regime
of oil- and coal-fired power plants but is largely com-
patible with the dominant sector regime of centralized
power generation. The gas turbine technology, however,
has the potential to weaken or even to break up the latter.
Developments are under way to build very small power
generation units like micro-turbines with a few MW of
electrical power output (Pilavachi, 2002). If such tur-
bines become economically and technically feasible they
may be at the heart of a radically decentralized electricity
supply system.
4.1.3. Wind power
By the beginning of the 20th century, first windmills
were designed to generate electricity but the technology
was soon abandoned as other options like steam turbines
matured and became widespread. It was not before the
1980s that wind power again started to play a role in the
electricity sector. At first, several public R&D programs,
which were targeted at the development of windmills
with a power output of some megawatts, ended in fail-
ures (Street and Miles, 1996). Instead small wind power
plants of around 50 kW turned out to be technologically
feasible and reliable. The US and California in particular
and later Denmark were key regions, in which hundreds
of windmills were installed in the early years of the tech-
nology development (Loiter and Norberg-Blohm, 1999).
In many cases, wind power was heavily opposed by
major electric utility companies, who argued that the
technology is too costly, not reliable and insignificant.
However, in countries like Germany, Spain and Denmark
governmental programs supported the development of
impressive niche markets (Lako, 2002, p. 10). So the size
and the performance of wind energy converters steadily
increased and new materials also allowed constructing
ever larger wings. Today, modern windmills achieve an
output of 1.5–5 MW and developments are under way to
construct large offshore wind parks in shallow sea waters
(Zwolinksi, 2003).
Wind energy converters entail a high degree of verti-
cal and horizontal novelty, cf. Fig. 1. The technology
core of wind turbines is completely new, i.e. neither
the technical principle nor these types of generators or
gear units have been applied in the electricity system
before. The technical characteristics of wind power are
 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625 617

fundamentally different as well. The power output can- the security of supply had become an issue of utmost
not be adapted to electricity demand but is determined importance.6 The oil crises of the 1970s made appar-
by the natural availability of the wind, i.e. it is some- ent that a too strong dependence on fossil fuel could
how predictable but not controllable. Moreover, wind severely hamper economic development in the indus-
power plants are very small compared to conventional trialized world. Thus, an almost exclusive reliance on
power stations—the difference in electricity output is oil-based power generation became a reverse salient for
about three orders of magnitude. the further development of the electricity system.7
Due to the latter two characteristics wind power is
incompatible with the dominant sector regime of central-
4.2.2. Combined cycle gas turbines
ized power generation and with prevailing operational
The development of CCGT power plants received
standards. Thus, a widespread diffusion of wind power
major impulses from scientific and engineering advances
technology would deeply affect every part of the elec-
of the underlying gas turbine technology in aeronau-
tricity supply value chain. As long as wind power covers
tics and other industrial applications (Islas, 1997). In
just some percent of electricity supply, its intermittent
the 1950s and 1960s, R&D on jet engines received large
nature can be balanced with other power plants, i.e. a
amounts of funding from military budgets and gas tur-
co-existence of the different regimes is technologically
bines soon became the dominant propulsion technology
feasible. With a higher degree of diffusion, however,
for aeroplanes. Military R&D, together with an increas-
incompatibilities become increasingly costly, thus mak-
ing demand for jet engines in civil aeronautics and the
ing the struggle of the regimes more and more virulent.
opportunity to transfer the technical progress to other
Wind power, in other words, has a potential to foster a
application contexts, stimulated incumbent manufactur-
regime shift in the electricity supply system.
ers of electromechanical equipment like General Elec-
tric, Alstom, Kraftwerk Union, Westinghouse, Brown
4.2. Factors driving radical innovations
Boveri or Sulzer to expand their involvement in gas
turbine R&D and manufacturing (Islas, 1997). This pro-
4.2.1. Nuclear power
vided the basis to successfully occupy the market niche
In the case of nuclear power, the technological devel-
for peak load electricity generation. Ongoing military
opment from early reactor prototypes to commercial
R&D and technological advances in aeronautics created
power plants was strongly supported by policy makers.
a continuous knowledge transfer, which continued well
This included the provision of large amounts of finan-
into the 1990s.
cial resources for R&D programs and pilot installations
The introduction and diffusion of CCGTs, how-
as well as the implementation of supportive institutions
ever, was not only technologically driven but comple-
and regulatory frame conditions. These measures were
mented by socio-economic and regulatory influences
all the more important because nuclear power was at first
(Islas, 1999). Public protest movements against nuclear
a non-competitive, high-risk endeavor resisted by many
power and environmental pollution in the 1970s and
electric utilities. Major obstacles for the new technology
1980s led to an increasing discrimination of the dom-
were not only high costs and its enormous capital inten-
inant generation technologies and ever more restrictive
sity but also well established combustion technologies on
environmental regulations. Moreover, new and large nat-
the basis of fossil fuels. In addition, more and more con-
ural gas reserves had been discovered and the collapse
cerns about safety and environmental hazards emerged in
of oil and gas prices since the mid 1980s set clear eco-
the broader public (Patterson, 1999, p. 64). Again policy
nomic incentives in favor of CCGT technology (Winskel,
makers – later supported by electric utilities, which had
2002).
adopted the innovation – intervened in favor of the new
technology as they launched public information cam-
paigns and implemented specific regulations to protect 6 Although the argument is widespread, some authors doubt the rea-
the new technology and the economic interests of the soning that nuclear power was politically supported in order to increase
firms involved (e.g. Walker, 2000). the security of electricity supply (Proops, 2001). Nuclear energy, so the
The major driving forces behind the introduction of alternative argument, rather fit well into the prevailing public discourse
nuclear power had a political background: the general and the picture of a centralized, interventionist and modernizing state.
7 Nuclear power, however, was not the only alternative to reduce oil
aim to increase the diversity of primary energy carriers
dependency. Denmark, which was confronted with a similar situation
and to become less dependent on oil as well as the pos- as France (see footnote 6), developed a completely different pathway
sibility to produce plutonium for nuclear weapons. With based on small, decentralized combined heat and power plants and
the ever increasing expansion of the electricity system wind energy (Hadjilambrinos, 2000).
618 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625

Table 1
Selection and classification of factors that triggered radical innovation processes in the electricity supply system
External developments Internal developments Reverse salients Government policies

Oil crises in 1973 and 1979
Demand for nuclear weapons
Success and development of
gas turbines in aeronautics
Increasing awareness of
environmental issues
System expansion
. . .made the security of
supply more important
. . .rendered environmental
impacts more striking
Large share of oil-fired power plants R&D programs
Inflexibility of the “supply system” Investment subsidies
to respond to new challenges Information campaigns
Power plants with particular high air Production incentives
emissions (SOx , NOx , CO2 ) Institutional support: licensing
procedures, legal frame
conditions
Grid access for independent
power producers

In the US, the Public Utilities Regulatory Policy
Act from 1978, which encouraged power generation by
independent power producers on the basis of renewable
energies or the cogeneration of heat and power, initi-
ated far-reaching institutional changes in the electricity
industry (Patterson, 1999, p. 72).
This legislation was at the beginning of a funda-
mental shift in the evaluation criteria for capital invest-
ments in power plants—in the US and elsewhere: cost
and risk considerations received increasing attention
and capital intense technologies were put at a dis-
advantage. Furthermore, CCGT technology proved its
competitive strengths with regard to short construction
time—especially in situations where a rapid growth in
electricity demand had to be covered. Following an
LTS interpretation, we may say that an ever increasing
size of power plants together with highly paradigmatic
investment decision in favor of established technolo-
gies caused a fundamental inflexibility to adapt to vari-
ous challenges—be it major ups and downs of primary
energy prices, environmental problems or rapid fluctua-
tions of demand. Thus, gas turbines as well as grid access
for independent producers may be regarded as examples
of very different counter developments to increase flex-
ibility.
benefit from the financial incentives. In many countries,
however, the key actors and business firms that pushed
the diffusion of wind power were not established com-
panies in the electricity industry but newcomers.
A major driver for the political support of wind power
was the concern about environmental issues. The con-
tinuous growth in the electricity demand led to ever
increasing ecological impacts and environmental haz-
ards, which can be interpreted as reverse salients for the
expansion of the system. Whereas some environmental
problems like acid rain have been overcome by end-of-
pipe technologies, others like the depletion of fossil fuels
or the emission of greenhouse gases call for renewable
energy sources.
Summing up the lessons learned from these exam-
ples, we can identify an interplay of external and internal
developments causing the emergence of reverse salients
in the electricity supply system. As the reverse salients
caught increasing attention, specific energy policies were
formulated to foster innovations including radically new
technologies. In every case, the initial resistance against
these new technologies was quite considerable and sev-
eral drivers were needed simultaneously to push the inno-
vation. Table 1 resumes the various factors that triggered
radical innovations in the examples discussed above.

4.2.3. Wind power 5. The influences of market liberalization on

The development and market diffusion of wind tur-
bines was strongly supported by a variety of government
policies including specific R&D programs, construc-
tion incentives like investment subsidies and produc-
tion incentives such as feed-in tariffs or tax exemptions
(Enzensberger et al., 2002; Street and Miles, 1996).
Unlike policy makers, many of the incumbent industrial
actors in the electricity supply system tended to oppose
and lobby against wind power. This resistance could be
lowered in some cases as support policies were designed
in a way that, for instance, electric utilities could also
innovation processes
Market liberalization has altered the electricity supply
sector in various respects (e.g. Rider, 1999; Sioshansi,
2001a). Customers have a choice which kind of elec-
tricity product or contract to buy from which supplier
and utility companies developed a variety of new prod-
ucts and energy services accordingly. Incumbent electric
utilities went through a phase of organizational restruc-
turing including far-reaching changes in the structure of
personnel. New players entered the market in traditional
 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625
fields like power generation and sales but also in new
areas like electricity trading, risk management, demand
aggregation, green power marketing, etc. Market liber-
alization is furthermore accompanied with a variety of
institutional changes in fields like grid access and net-
work pricing, power trading, risk management, market
transparency and consumer protection, etc.
Following the findings from Section 4 one might
argue that market liberalization can be regarded as
another external driver, which brings about radical orga-
nizational innovation in the electricity supply system.
The impacts, however, are likely to be more far-reaching
because liberalization encompasses a change of the most
basic working principles of the system as such. This may
result in a fundamentally different way of handling all
kinds of innovations.8
Given that market liberalization is a rather recent
process, it is not an option to analyze successful inno-
vation processes and the emergence of new regimes
on an ex-post basis like we did before with regard to
monopoly conditions. We also have to explore a fur-
ther case because the innovations studied in Section
4 could already establish stable technological regimes
under monopoly times—not everywhere, but generally.
As a new case we chose stationary fuel cells, which is
a radical, technological innovation for power generation
and has only reached an early stage of development. In
addition to that case, we will provide a general review of
the changes of innovation processes and selection prin-
ciples due to market liberalization.
The following results are based on an empirical sur-
vey, which had the aim to study the impact of market
liberalization on the innovation behavior of electric util-
ity companies. Data were obtained in three series of
expert interviews including altogether 44 interviews in
more than 30 utilities from Germany, the Netherlands
and Switzerland.9
8 Approaching liberalization and its potential effects in such a gen-
eral way, we have to keep in mind that market liberalization is not
a consistent phenomenon but varies according to the specific regula-
tory frame conditions in each state or country. As our data basis (see
below) does not support a qualified comparison of different regulatory
frameworks, we will again refer to this issue in our conclusion.
9 An initial set of interviews concentrated on the general changes
in the handling of innovation (cf. Section 5.1). This was followed by
two innovation case studies, one on green electricity and the other one
on fuel cells (cf. Section 5.2). The results on green power innovation
strategies have much in common with the fuel cell case, e.g. with regard
to different types of strategies. However, these findings are not reported
here because we do not consider green electricity products as a radical
innovation. For further details see Markard (2003) or Markard et al.
(2004).
619
5.1. Changes in the selection environment for
innovations
Innovation processes may be described as a sequence
of variation and selection decisions. Variation opens
up the alternatives, which should be further developed;
selection reduces the set again in order to focus scarce
resources on the most promising options. The selec-
tion process takes place at different levels. At the level
of the sector, the selection of innovations depends on
consumer interests, regulatory frame conditions and the
possibilities for innovation diffusion, i.e. replication and
imitation (Nelson and Winter, 1982, p. 226). At the level
of the innovating firm, decision makers have to identify
and select, for example, fields of technological progress
as well as particular innovation projects that are most
promising for the business areas the firm wants to be
involved in.
5.1.1. Selection of innovations at the level of the
firm
With regard to the situation under monopoly condi-
tions, several interview partners reported that innovation
processes in electric utilities mainly concentrated on
technical aspects since a high quality of power supply
was the guiding principle for decision making. It was said
that there was a focus on incremental improvements of
established technologies rather than on the development
of new ones. In other words, innovations had to be com-
patible with the established system of supply and the
corresponding engineering practices. As many electric
utilities were vertically integrated, innovation activities
in one part of the firm also had to take the interests of the
other parts into account. The room for horizontal vari-
ation and experimentation was thus limited by vertical
constraints.
Moreover, it was reported that innovations had to fit
into the existing organizational structures and that they
were developed mainly on the basis of existing resources
and competences. As a matter of fact, standard operating
procedures as well as investment and search routines did
not change very much over time.
In the liberalized market environment, the entre-
preneurial goals and decision making principles have
changed—together with the selection guidelines for
innovations, cf. Table 2 (upper part). Cost efficiency and
customer service orientation were said to receive highest
attention now. In a similar vein, many interviewees stated
that product differentiation and the diversification of
business fields have gained increasing importance. The
study also indicated that innovation projects in several
cases were designed in a way to strengthen co-operations
620 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625
Table 2
Changes in the selection environment at the firm and sector level—monopoly vs. liberalized market
Monopoly
Selection criteria (firm level)
Costs and quality
Does the innovation contribute to the technical quality of
power supply?
Does the innovation improve established technologies?
Innovation characteristics
Is the innovation compatible with existing regimes?
Does the innovation comply with the interests of other
business units of the (vertically integrated) firm?
Can the innovation be integrated into the organization
structures of the firm?
Competences, learning
Can the innovation be realized with the existing resources and
competences of the firm?
Selection mechanisms (sector level)
Role of consumers
No or little influence on innovation process, consumer
interests were indirectly mediated by public authorities
Regulation, policy making
Innovation processes influenced by the price and tariff system
regulation, various environmental policies, sometimes even
particular interests at the local-regional level
Innovation diffusion
Innovation management, influenced and often coordinated by
associations at the second level, widespread and deliberate
knowledge dissemination among firms, replication spatially
restricted
and strategic alliances with other utilities and manufac-
turing companies.
Furthermore, it was reported that the activities of com-
petitors as well as specific policies affected the decision
to invest in a specific innovation field. As uncertainties
are much higher than under monopoly conditions, some
interview partners also highlighted that the investment
priorities have shifted towards shorter time horizons
and less capital intense technologies. It could also be
observed that there was more room for experimenting
with radically new technologies—a fact that might be
attributed to the unbundling of formerly integrated busi-
ness units.
5.1.2. Selection of innovations at the level of the
sector
In monopoly times, electricity consumers had
almost no influence on innovation processes. At the
supplier–consumer interface there was no choice among
different electricity offerings and the majority of inno-
Liberalized market
Costs and quality
Does the innovation contribute to cost reductions?
Is the new product of interest for the customers?
Is there a potential to develop a new business field?
Innovation characteristics
Do competitions pursue similar innovation activities?
Are the policy incentives to invest in a novel technology?
Has the new technology a rather low degree of capital intensity?
Competences, learning
Does the innovation support strategic co-operations?
Does it contribute to the development of new competences?
Role of consumers
Influence on innovation on the basis of the electricity purchase decision
More intense communication between suppliers and consumers
Regulation, policy making
Price and tariff regulation mostly removed, environmental policies
remain, new regulations with regard to consumer protection,
market transparency and competition enhancement
Innovation diffusion
Not coordinated any more, reduced knowledge exchange among
competing innovators, no restrictions for replication
vations concentrated on parts of the value chain
(power generation, transmission, distribution), which
were assumed to be of no interest to the consumer.
Instead, public authorities had the task to represent the
interests of consumers and citizens, e.g. with regard
to electricity prices, technology “choices”, security of
supply or environmental issues.10 Sometimes, public
authorities also pursued local political interests in their
role as owners of electric utilities.
Innovation management and the diffusion of innova-
tions were often coordinated by associations at the level
of the sector. These institutions organized the dissemina-
tion of knowledge and most utilities jointly contributed
to these networks. Thus, the possibilities of innovation
10 It has already been pointed out that policy interventions were a
powerful driver for innovations in the electricity sector—be it radical
ones like the introduction of nuclear power or the emergence of wind
turbines or incremental ones like the desulphurization of coal-fired
power plants.
 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625
diffusion and imitation were quite significant, whereas
replication was restricted to the regional monopoly area
of each utility.
In a liberalized market, however, consumers have
the opportunity to influence decision making – and the
innovation processes – in the electricity supply system
more directly. Residential customers can chose among
different supply offerings and commercial customers
may even negotiate the details of their delivery con-
tracts individually.11 The influences of regulatory frame
conditions have rather changed gradually. Electricity
marketers, for instance, have more freedom to design
products and prices. They also face new regulations, e.g.
with regard to consumer protection and market trans-
parency.
Finally, knowledge dissemination and innovation
management are no longer coordinated by central associ-
ations and firms are more reluctant in co-operating with
one another. Thus, imitation is much harder now and
the corresponding information channels have changed.
Table 2 (lower part) summarizes the changes in the
selection mechanisms for innovations at the level of the
electricity sector.
5.2. Innovation strategies of electric utilities in the
case of the fuel cell
In the following we will illustrate on the basis of a case
study on fuel cell innovation strategies, how the changes
in the selection environment led to a diversified behavior
of electric utilities dealing with new ideas, products and
technologies. In a competitive environment, utilities are
for the first time called to develop explicit innovation
strategies, to identify and evaluate alternative innovation
projects and to pursue and communicate their activities
in a strategic way.
The fuel cell is a conversion technology that trans-
forms hydrogen (e.g. from natural gas or biogas) and
oxygen into electricity, heat and water. This conversion is
highly energy efficient, has a favorable electricity to heat
ratio and is characterized by low air emissions. Fuel cells
are in an early stage of the innovation process and due to
its high costs the technology is still limited to niche appli-
cations and pilot projects (Adamson and Jollie, 2004;
Baker and Jollie, 2004). The range of potential fuel cell
applications is very broad as the technology is based on
11 Yet, actual consumer choice, e.g. in terms of switching rates, has
been rather limited in several liberalized markets, especially in the
smaller customer segments (EU, 2002; Sioshansi, 2001b). It therefore
remains to be seen how much selection pressure customers will exert
on a mid-term basis.
621
a modular design. Fuel cell power plants can vary in the
size of their power output from several watts up to some
megawatts.
Fuel cells represent a radical innovation with a high
degree of vertical and horizontal novelty. A widespread
application of fuel cells would deeply affect every part
of the value chain, cf. Fig. 1. A diffusion of fuel cells will
contribute to a decentralization of electricity supply and
might also foster a local use of heat. Like the gas turbine,
fuel cells substitute oil, coal and uranium as primary
energy carriers. The new technology may also lead to a
further enlargement of the gas supply infrastructure. On
a mid-term perspective also non-fossil energy sources
may be used to provide the hydrogen needed.
5.2.1. Identification of different innovation
strategies
In order to analyze and compare the strategies of elec-
tric utilities in the field of the fuel cell, we asked for the
goals they pursue, for the priority of different goals and
for the resources allocated.12 Regarding the latter we
made a rather rough distinction of high, medium and
low resources in relation to the size of the company.
In a few utilities, the interviewees expressed the inten-
tion to create a market for fuel cell power plants and to
become a leader in this field. These firms also had the
goal to gather experiences with the new technology and
to improve their corporate image. The fuel cell projects
had a high or very high priority and received a high
amount of resources accordingly. In the public relation
activities the fuel cell played a key role. Furthermore,
the companies concentrated on a particular technologi-
cal design—small-size fuel cell systems for the heating
of residential buildings (one-family or multiple-family
homes).
Other interviewees said that the main goal is to gather
experiences with the new technology and its perfor-
mance in different applications. These firms were also
interested in learning about customer preferences and
wanted to establish contacts with manufacturers and
other actors in the fuel cell business. The resources were
on a medium level and the importance of fuel cell projects
was said to be similar to that of other innovations. Fur-
thermore, the companies had quite a broad portfolio of
different fuel cell systems and applications.
12 Between 2002 and 2003 we conducted 13 interviews with employ-
ees of 8 utilities, mostly in Germany (Markard, 2004). The interviews
were organized in two sequences—an initial study of innovation strate-
gies and influence factors on strategies (6 interviews) and a second
survey on the changes of innovation strategies and the learning pro-
cesses within the firms (7 interviews).
622 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625
Table 3
Basic types of fuel cell innovation strategies of electric utilities
Strategy type Resources Goals
Leading High Make profit (+); improve corporate image, gather experiences
Learning Medium Gather experiences (+)
Image shaping Low Improve and shape corporate image (+), gather experiences
+ high priority, all others: medium priority, goals that received a low priority are not mentioned.
A third group of utilities stressed the innovative image
of the fuel cell: their projects were primarily aimed at
shaping the corporate image of the company. But expe-
riences and the creation of knowledge also played a role.
The projects were regarded as experiments, on which
to base further innovation decisions in the future. They
received a medium – in one case even low – priority
and the investments were strictly limited. These com-
panies just operated a small number of fuel cell power
plants—for residential applications as well as for small
business customers or municipalities.
Although our empirical sample was rather limited
(eight firms) we observed some characteristic combina-
tions of goals and resources, to which we assign three
types of innovation strategies, cf. Table 3. The leading-
strategy is characterized by a relatively high amount of
resources and the goal to become a market leader in
future and to make profits with the new technology. The
learning-strategy is characterized by a smaller amount
of resources and the goal to gather experiences. And the
image-strategy is characterized by a rather small amount
of resources and the main goal to shape the corporate
image with a fuel cell project.
5.2.2. Motivations to initiate and pursue innovation
activities
Similar to the innovation case studies presented
before, we sought to explore the factors driving radical
innovations. Here, the analysis had to concentrate on the
decisions at the firm level to become active in the field
of fuel cells and to set up a specific innovation strategy.
Basically, we can distinguish external drivers and
internal influences that depend on the particularities of
each utility company. The interviewees reported that they
felt a need to position the company as innovative and to
develop new competences in fields like product devel-
opment, customer oriented services or contracting, cf.
Table 2. Furthermore, the potential threat of being swept
away by a new technology in the traditional field of elec-
tricity supply was mentioned as a driver. This aspect,
however, was also expressed in firms with a leading-
strategy. Here, the mid-term profit potential was empha-
sized.
No. of firms in the survey
2
4
2
In another case, it was reported that the initiative
to launch the fuel cell project came from the fuel cell
manufacturer and that the availability of public funding
also had a positive influence. External influences, in fact,
determined in most cases the decision to become active.
And we can even relate many of the external driving
factors to the introduction of market liberalization.
Internal factors rather had an impact on the type of
innovation strategy. Only large companies could afford
to carry out a variety of fuel cell projects and to launch
far-reaching advertising campaigns on this subject. Indi-
vidual strategies and certain visions for the company as
a whole also had an important influence on the innova-
tion strategy chosen. Furthermore, a general openness for
novel technologies and products facilitated the launch of
fuel cell projects and highly dedicated employees were
identified to be a key driver for a learning-oriented or
even a leading-strategy.
In sum, the analysis showed that incumbent electric
utilities perceived the need to develop new markets and
to actively engage in a broad range of learning processes.
A few firms even turned out to become prime movers, i.e.
they had the intent and the resources to significantly push
the innovation and guide the direction of development.
This is all the more striking as the fuel cell is not only
a radical innovation but also requires a broad range of
novel competencies, which the utilities did not possess.
6. Conclusions
Large technical systems like electricity supply are
characterized by significant barriers for the devel-
opment and diffusion of radical innovations. Still,
far-reaching technological changes have occurred in
the electricity sector over the last decades. A review
of three cases, in which radical innovations developed
and finally became established in the electricity supply
system under monopoly conditions, showed that such
processes were driven by internal as well as external
developments, which caused frictions (reverse salients)
and thus provoked policy interventions to support new,
promising technologies as potential solutions. In all
these cases, the initial resistance of incumbent actors
 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625
like electric utilities was quite considerable and all the
more influential as the firms had common interests
and pursued similar strategies. On this basis, powerful
associations could be established to lobby policy makers
but also to coordinate innovation processes focusing on
incremental improvements. As a consequence, radical
innovations depended on strong and enduring support
by government policies in order to penetrate the sector.
Market liberalization has changed the handling of
innovations in at least two respects, the scope of vari-
ation and the focus of innovation management. With
regard to the former, our comparison of innovation pro-
cesses under monopoly conditions with the situation in
a liberalized market environment revealed considerable
changes at the level of the firm. Organizational routines
including daily business practices as well as investment
principles and strategic goals have been modified. Orga-
nizational structures and firm culture were transformed
as well. In our sample, we could identify an organiza-
tional environment for innovation managers, which is
more creative and risk oriented with regard to uncon-
ventional ideas and solutions. As product differentiation
and diversification have gained importance and room to
maneuver has enlarged, these changes can be interpreted
as an increase in the scope of variation at the level of
the firm. At the level of the sector, we could observe a
broader diversity of innovation strategies of electric util-
ities, which also indicates that the scope of variation has
increased. In the long run, such a gain in variation may
well lead to a decrease in path-dependency, thus paving
the ground for radical innovations to develop.
One may argue, however, that what is reported in
interviews can be different from what firms actually do.
In the case of stationary fuel cells, findings from the
interviews were supported by actual innovation activities
documented in internal reports, annual reports or on the
internet. Still, further analyses will be needed to clarify
whether, for example, a similar degree of strategic vari-
ation can be observed in other innovation fields as well.
With regard to the focus of innovation management
at the firm level, our findings suggest that liberalization
induced a shift from incremental, technology-oriented
innovation to more radical, customer-oriented product
innovations and organizational innovations. Innovation
opportunities are selected according to whether they
might reduce costs, fulfill customer needs, open up new
markets, etc.
In addition, we may also infer that market liberaliza-
tion can serve as a driver for the overall level of inno-
vation activity as competition represents a significant
challenge for incumbent electric utilities as well as new-
comers. Firms face the need to develop a broad range of
623
market specific competencies, which explains the signif-
icance of learning oriented innovation strategies. Electric
utilities also started to systematically observe each other
and to imitate innovations as they perceived an “ignoring
type” of strategy as too risky.
However, there are also indications that market
liberalization may as well lead to a decline of the R&D
expenditures of electric utilities and to a shift towards
short-term oriented R&D projects (Dooley, 1998). Our
findings on the investments into stationary fuel cell
development seem to be at odds with the suggested
short-term orientation of R&D as the technology
presents a mid- to long-term endeavor. A possible expla-
nation is that there is a general difference in handling
incremental and radical innovation. In this case, one
may well observe cutbacks in incremental innovations
while, at the same time, selected radical technologies
are pursued or even extended. Our analyses were far too
limited to provide representative evidence on that issue.
Thus, it remains a topic for further research to assess
the shifts in overall innovation activity at the level of
firms and the sector both with regard to incremental
and radical innovation. There is also a need to analyze
the corresponding effects on established regimes and
development paths on a mid-term basis.
In sum we may argue that market liberalization has
transformed variation and selection principles related
to radical innovations. One possible consequence at
the sector level is a general broadening of innovation
routes. We may also postulate that the widespread and
coordinated resistance of established utility companies
against radical innovations, which was characteristic
under monopoly conditions, gives way to much more
diversified strategic responses. New strategic alliances
and innovation networks are formed, many of them pur-
suing heterogeneous interests and some even obtaining
a benefit from the promotion and diffusion of radically
new technologies.
Still, we have to qualify the possibly over-optimistic
tune of these findings in two respects. Firstly, the empiri-
cal basis of our studies on the liberalized electricity mar-
ket environment was restricted to a set of three countries.
So the potential for generalization is limited; even more,
as our empirical basis was not sufficient to study how dif-
ferent regulatory frameworks affect innovation processes
in different ways.13 Other policies, e.g. environmental
13 One might, for instance, imagine that market conditions, in which
incumbent utilities are still able to cross-subsidize their competitive
activities by excessive charges for the remaining monopolistic bot-
tlenecks, are less favorable for radical innovations than situations
characterized by a high degree of competition.
624 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625
regulations, are also likely to co-affect innovation pro-
cesses. The case on wind power is a typical example
in this respect. Further studies on the drivers for radi-
cal innovation therefore will surely benefit from a cross
comparison of the regulatory framework conditions in
different countries.
Secondly, our empirical analysis concentrated on
innovation activities, which might be well different from
actual investment behavior of electric utilities. In fact,
generation technologies, network structures, power con-
suming appliances, etc. have not changed much since
the beginning of liberalization and the established tech-
nological regimes still exist. Market liberalization cannot
be interpreted as an external factor that leads to a shift
in a particular technological regime or supports a spe-
cific set of new technologies. It is rather a driver that
transformed the basics of search and innovation pro-
cesses and may thus weaken prevailing technological
regimes.
Finally, the current transformations in electricity mar-
kets and organizational structures rather seem to repre-
sent a window of opportunity than a steady state. Re-
stabilization processes of the system are already under
way. The ongoing concentration processes due to utility
mergers and acquisitions may serve as an indicator. One
might well foresee a future situation, in which the former
regional monopolies will be substituted by international
oligopolies. In an oligopolistic market the question rises
anew whether such developments will again reduce the
scope of variation in innovation processes.
While the question of how to judge the effects of mar-
ket liberalization on innovation processes still needs fur-
ther investigation, variety and focus of innovation have
been revealed as key issues to pay attention to—with
regard to the assessment of market liberalization, but
even more in terms of policy making, especially when
regime transformations (e.g. towards sustainability) are
intended.
References
Abernathy, W.J., Utterback, J.M., 1978. Patterns of industrial innova-
tion. Technology Review 80, 40–47.
Adamson, K.-A., Jollie, D., 2004. Fuel cell market survey: small sta-
tionary applications. Fuel Cell Today.
Baker, A., Jollie, D., 2004. Fuel cell market survey: large stationary
applications. Fuel Cell Today.
Burns, R.E., Potter, S., Witkind-Davis, V., 2004. After the lights went
out. The Electricity Journal 17 (1), 11–15.
Christensen, M.C., 1997. The Innovator’s Dilemma. Harvard Business
School Press, Boston.
David, P.A., 1985. Clio and the economics of QWERTY. American
Economic Review 75, 332–337.
Davies, A., 1996. Innovation in large technical systems: the case
of telecommunications. Industrial and Corporate Change 5 (4),
1143–1180.
Dooley, J.J., 1998. Unintended consequences: energy R&D in a dereg-
ulated energy market. Energy Policy 26 (7), 547–555.
Dosi, G., 1988. The nature of the innovative process. In: Dosi, G.,
Freeman, C., Nelson, R., Silverberg, G., Soete, L. (Eds.), Technical
Change and Economic Theory. Pinter, London, pp. 221–238.
Enzensberger, N., Wietschel, M., Rentz, O., 2002. Policy instruments
fostering wind energy projects—a multi-perspective evaluation
approach. Energy Policy 30 (9), 793–801.
EU, 2002. Second Benchmarking Report on the Implementation of the
Internal Electricity and Gas Market. EU Kommission, Brussels.
Godoe, H., 2000. Innovation regimes R&D and radical innovations in
telecommunications. Research Policy 29 (9), 1033–1046.
Gökalp, I., 1992. On the analysis of large technical systems. Science,
Technology and Human Values 17 (1), 57–78.
Hadjilambrinos, C., 2000. Understanding technology choice in elec-
tricity industries: a comparative study of France and Denmark.
Energy Policy 28, 1111–1126.
Henderson, R.M., Clark, K.B., 1990. Architectural innovation: the
reconfiguration of existing product technologies and the fail-
ure of established firms. Administrative Science Quarterly 35,
9–30.
Hirsh, R.F., 1999. Power Loss: The Origins of Deregulation and
Restructuring in the American Electric Utility System. MIT Press,
Cambridge.
Hoogma, R., Kemp, R., Schot, J., Truffer, B., 2002. Experimenting for
Sustainable Transport. The Approach of Strategic Niche Manage-
ment. Spon Press, London/New York.
Hughes, T.P., 1987. The evolution of large technological systems. In:
Bijker, W., Hughes, T.P., Pinch, T. (Eds.), The Social Construction
of Technological Systems. Cambridge, MA, pp. 51–82.
IEA, 2002. Energy Policies of IEA Countries—United Kingdom 2002
Review. International Energy Agency (IEA), Paris.
Islas, J., 1997. Getting round the lock-in in electricity generating sys-
tems: the example of the gas turbine. Research Policy 26 (1),
49–66.
Islas, J., 1999. The gas turbine: a new technological paradigm in elec-
tricity generation. Technological Forecasting and Social Change
60 (2), 129–148.
Jacques, C., Lafrance, G., Doucet, J.A., 2001. Inertia in the North
American electricity industry: is it realistic to think that the Kyoto
Protocol objectives can be met? Energy Policy 29, 453–463.
Joerges, B., 1998. Large technical systems: concepts and issues. In:
Mayntz, R., Hughes, T. (Eds.), The Development of Large Techni-
cal Systems. Westview Press, Boulder, pp. 9–36.
Joskow, P.L., 1998. Electricity sectors in transition. The Energy Journal
19 (2), 25–52.
Kemp, R., Schot, J., Hoogma, R., 1998. Regime shifts to sustainability
through processes of niche formation: the approach of strategic
niche management. Technology Analysis & Strategic Management
10 (2), 175–195.
Lako, P., 2002. Learning and Diffusion for Wind and Solar Power
Technologies. Energy Research Centre of the Netherlands (ECN),
Petten.
Loiter, J.M., Norberg-Blohm, V., 1999. Technology policy and renew-
able energy: public roles in the development of new energy tech-
nologies. Energy Policy 27 (2), 85–97.
Markard, J., 2003. Liberalisierung des Elektrizitätsmarkts und
ökologische Innovationen. Swiss Federal Institute of Technology,
Zurich.
 J. Markard, B. Truffer / Research Policy 35 (2006) 609–625
Markard, J., 2004. Strommarkt im Wandel—Veränderung von Inno-
vationsprozessen am Beispiel von Ökostrom und Brennstoffzelle.
vdf Hochschulverlag, Zürich.
Markard, J., Truffer, B., Imboden, D.M., 2004. The impacts of mar-
ket liberalization on innovation processes in the electricity sector.
Energy & Environment 15 (2), 201–214.
Mascitelli, R., 2000. From experience: harnessing tacit knowledge to
achieve breakthrough innovation. Journal of Product Innovation
Management 17 (3), 179–193.
Nelson, R.R., Winter, S.G., 1982. An Evolutionary Theory of Eco-
nomic Change. Harvard University Press, Cambridge.
Patterson, W., 1999. Transforming Electricity. The Royal Institute of
International Affaires, London.
Pilavachi, P.A., 2002. Mini- and micro-gas turbines for combined
heat and power. Applied Thermal Engineering 22 (18), 2003–
2014.
Proops, J., 2001. The (non-) economics of the nuclear fuel cycle: an
historical and discourse analysis. Ecological Economics 39 (1),
13–19.
Rider, G., 1999. Ten lessons for the changing European electricity
landscape. The Electricity Journal 12 (3), 13–19.
Rip, A., Kemp, R., 1998. Technological change. In: Rayner, S., Malone,
E.L. (Eds.), Human Choice and Climate Change—Resources and
Technology. Battelle Press, Columbus, pp. 327–399.
Rosenberg, N., 1982. Inside the Black Box: Technology and Eco-
nomics. Cambridge University Press, Cambridge.
Sine, W.D., David, R.J., 2003. Environmental jolts, institutional
change, and the creation of entrepreneurial opportunity in the
US electric power industry. Research Policy 32 (2), 185–
207.
Sioshansi, F.P., 2001a. Competition in liberalized European electricity
markets. The Electricity Journal 14 (2), 73–83.
625
Sioshansi, F.P., 2001b. US electricity restructuring experience: hits and
misses. Utilities Policy 10 (1), 1–6.
Sondreal, E.A., Jones, M.L., Groenewold, G.H., 2001. Tides and trends
in the world’s electric power industry. The Electricity Journal 14
(1), 61–79.
Street, P., Miles, I., 1996. Transition to alternative energy supply tech-
nologies: the case of windpower. Energy Policy 24 (5), 413–425.
Teece, D.J., 1996. Firm organization, industrial structure, and techno-
logical innovation. Journal of Economic Behavior & Organization
31 (2), 193–224.
Tushman, M.L., Anderson, P., 1986. Technological discontinuities and
organizational environments. Administrative Science Quarterly 31,
439–465.
Tushman, M.L., Nadler, D., 1986. Organizing for innovation. Califor-
nia Management Review 28 (3), 74–92.
UCTE, 2004. Final Report of the Investigation Committee on the 28
September 2003 blackout in Italy, Brussels.
van den Ende, J., Kemp, R., 1999. Technological transformations in
history: how the computer regime grew out of existing computing
regimes. Research Policy 28 (8), 833–851.
Walker, W., 2000. Entrapment in large technology systems: insti-
tutional commitment and power relations. Research Policy 29,
833–846.
White, D., Roschelle, A., Peterson, P., Schlissel, D., Biewald, B., Stein-
hurst, W., 2003. The 2003 blackout: solutions that won’t cost a
fortune. The Electricity Journal 16 (9), 43–53.
Winskel, M., 2002. When systems are overthrown: the ‘dash for gas’
in the British electricity supply industry. Social Studies of Science
32 (4), 563–598.
Zwolinksi, S., 2003. Offshore wind energy takes off at General Electric:
interview with Steve Zwolinksi, GE Wind Energy. Refocus 4 (6),
54–56.