Você está na página 1de 16

Cities and Networks. What territorial integration?

- La Rochelle, 20-21 October 2005

The Mediterranean Latin Arc: innovation and polycentrism in the urban structure
Alberto Vanolo
Dipartimento Interateneo Territorio Politecnico e Universit di Torino vanolo@econ.unito.it

Summary The paper proposes an analysis of the territorial structure of South-Western Europe (the so-called Mediterranean Latin Arc) organized around two main concepts: on the one hand, polycentric development, one of the policy aims identified by the European Commission in the European Spatial Development Programme, and on the other hand, technological innovation, explored with reference to the theory of industrial clusters. The empirical analysis has been carried out the geographical scale of the local systems, considering a wide amount of data in order to draw a general picture of the development dynamics of the area and to evaluate both the high-technology potential of the local systems, and the emerging territorial patterns in terms of monocentric or polycentric structures. Finally, the paper proposes some theoretical reflections concerning the concept of polycentric development and its implications with innovative strategies.

1. Mediterranean Latin Arc, polycentrism and innovation The Mediterranean Latin Arc is an area including coastal regions between Andalusia and Campania. The idea of such a geographical entity is strictly connected to some synthetic representations of the European territory proposed by geographers and urban scholars, evidencing the presence of a centreperiphery model, with a strong (in terms of socio-economic indicators) Western European core opposed to the weaker peripheral regions (see, among the others, Meijer, 1993; European Commission, 1999). This centre-periphery model, in particular, refers to the presence of a central backbone, an urban structure already identified by Brunet (1989), which includes the London region, Randstad Holland and Rhr area, continuing through the Rhine axis down to Lombardy. This area shows greater economic and urban performances vis--vis the more peripheral regions (Crpm, 2002), and over recent years several scholars have underlined the advisability of strengthening other geographical development axes in order to promote a more balanced structure. From this perspective the idea of promoting an alternative Mediterranean axis labelled by the European Commission (1994) Mediterranean Latin Arc was born. This is an area with uncertain geographical boundaries, and in this sense it has been defined by Boulifard (1994) a variable geometry area: a general picture, drawn upon a series of theoretical contributions and empirical analysis1, is presented in Figure 1. This
1

The idea of a Mediterran Latin Arc has been proposed in a number of different geographical images, shapes and boundaries: for a brief review see Pere and Toms (1998). Figure 1 has been drawn on the basis of the original idea proposed by the

area constitutes the geographical context of the analysis, concerning its dynamics and potentialities under two specific perspectives, particularly recurrent in theoretical and political debates: polycentric territorial development and technological innovation. Fig. 1 The Mediterranean Latin Arc

1.1 Polycentrism The concept of polycentrism has become particularly popular over recent years2. It refers, in an intuitive way, to the presence of multiple development nodes on the territory, and in this sense it can be considered to be in opposition to the traditional polarization theories that characterize post-war regional sciences. The concept of polycentrism is ideally the evolution of the decentralized concentration one, traditionally tied to Dutch spatial planning and referring to policies aimed at the spread of economic activities from major congested urban areas in order to reconcentrate them in the main poles of the less developed regions. Obviously, the aim does not simply consist in the reorganization of a fixed amount of economic activities in a sort of zero-sum game, but to encourage development processes in regions and cities characterized by economic stagnation, industrial decline, or depopulation (Hall, 2001). As a geographical concept, polycentrism can be consider from three different perspectives: as a spontaneous phenomenon, as a theoretical model, and as a political objective. From the first point of view, the current popularity of polycentrism among scholars reflects a general trend of the urban phenomenon: several monocentric and polarized structures, i.e. characterized by one big dominant centre, tend to reorganize themselves in networks and polycentric structures: this is, for example, the case of many Italian urban areas, as pointed out by Dematteis (1995). The increasing mobility of goods and people, together with the fast diffusion of information and communication technologies, renders many traditional monocentric urban theoretical models (i.e. Nortons city-life cycles model, 1979) unable to describe and to interpret the most complex current city dynamics.

European Commission (1994) with some integrations based on the works of Conti and Salone (2001) and CRPM (2002). For a full description of the qualitative logical path at the basis of the construction of Figure I, see Vanolo (2003). 2 See, for example, Camagni and Salone (1993), Datar (2000), Bailey and Turok (2001), Kloosterman and Musterd (2001). Considering polycentrism as a rather ambiguous term is coherent with the words of Musterd and van Zelm (2001, pp. 67980), stating that although () the term has become en vogue in the literature and in political documents, a clear and unambiguous definition of a polycentric system is still lacking.

This topic therefore overlaps with polycentrism as a theoretical concept, almost in opposition to the traditional idea of urban hierarchy in terms of dimensions and rank of the urban functions (for example, in terms of population in the famous Zips rank-size model, 1949). From this perspective, the recent debate on the polycentric urban region does not only concentrate on the different endowments of functions of the territorial systems, but on the variety and diversity of these functions, their distribution between the centres, and the consequent relations of integration and interdependence. From a political point of view, finally, the objective of polycentrism is to promote a fair and balanced spatial development and, for this reason, it has been included in the policy aims identified by the European Commission (1999). This polycentric strategy can be read on different geographical scales: in a wider European perspective, it refers to the development of new development axes in the peripheral regions outside of the European core (CRPM, 2002), while on the regional scale, the aim is to contrast the formation of monocentric urban systems in which the greater part of the functions are concentrated in one narrow area, in order to promote equipotential networks of competitive centres sharing different functions (Dematteis, 1995). At the centre of this second perspective lies the idea that polycentrism may represent a tool for enhancing regional competitiveness. This is an aspect of the debate that presents a certain ambiguity: arguments in favor of polycentrism, in fact, apparently contrast with the economic geography literature referring to the advantages of the major cities in terms of increasing returns to scale (Krugmann, 1991). Actually, this theoretical body is not strictly in contrast to the idea of polycentric development in terms of decentralized concentration. Polycentrism, in fact, never denies the benefits of spatial concentration, but underlines the necessity to promote networks and different development paths in those situations where an excessive geographical concentration (at different scales) leads to social and territorial imbalances, as in the case of the European core within the framework of the Community space, or in that of some regions characterized by a critical centre-periphery model3. Nevertheless, the literature concerning polycentrism calls attention to some specific ways to enhance the competitive advantage of urban systems (Bailey and Turok, 2001). First, cities can enjoy particular external economies deriving from sharing a common labour market, infrastructures like airports and freeways, or highly specialized services like universities. Second, cities can take advantage of their different complemental elements and specializations: basically, referring to the traditional economic theories, every city can specialize in economic sectors in which they enjoy specific comparative advantages. Third, it is reasonable to suppose that frequent interaction between the nodes of a polycentric structure, together with the sharing of problems, solutions and perspectives, will promote governance synergies, while the sharing of resources and programs allows the financing of bigger projects. Finally, considering the network dimension of polycentrism, it is possible to argue that, in some ways, it may favour the diffusion of knowledge and innovations. In order to clarify this hypothesis, it is necessary to refer to the spatial aspects of technological innovation processes and, more specifically, to the theory of geographical industrial clusters4.

Basically, theoretical models and empirical regularities have to be considered in conjunction with real territorial situations: quantitative laws and geometrical models (such as those concerning increasing returns, central localities, rank-size rule, etc.) must not be considered as predictive and normative models, but as indications for the interpretation of the complex phenomena of regional development. It is evident that, beyond a certain threshold, cumulative growth may lead to various social and environmental problems, underlining the need for more polycentric strategies. 4 See, among the others, Storper (1993), Amin and Thrift (1994), Malmberg (1996), Porter (1998), Steiner (1998).

1.2 Innovation and industrial clusters An industrial cluster may be defined as a spatial concentration of enterprises operating side by side with local organizations (such as universities and research centres) sharing the same social and cultural local environment. The success of industrial clusters has often been explained with reference to the particular advantages they offer in the field of market transactions (Williamson, 1975; Scott, 1988). However, more recent theoretical contributions tend to focus on the specific mechanisms that favour innovative processes (Malmberg and Maskell, 2002). In this framework it is useful to distinguish the relations between local economic actors in two analytical dimensions: the horizontal one, referring to enterprises carrying out the same functions in the value chain, and the vertical one, relative to different but complementary activities. The horizontal dimension of the cluster regards enterprises that, side by side, perform the same activities and are, therefore, potentially competitors. As already sensed by Alfred Marshall, the enterprises holding specific knowledge tend to develop different solutions for the execution of the same function. Co-localized enterprises can directly observe their own competitors, imitating the most successful solutions (Mamberg and Maskell, 1997; Loasby, 1999). In fact, inside the cluster, each enterprise knows the points of strength and weakness of their adversaries, as well as the quality and the cost of production factors. Through the observation, the discussion, and the comparison of the different practices, enterprises along the horizontal dimension learn and improve constantly, favoured by the sharing of the same socio-economic context and the use of the same languages and codes. The vertical dimension refers to enterprises carrying out different (complementary) activities in the value chain, i.e. enterprises connected by input-output relations. Suppliers and customers are attracted by the opportunities offered by the localization inside the cluster, like the possibility to establish privileged relationships (trusting, rapid, informal) with other subjects (Gertler, 1996; Danson and Whittam, 1999). The vertical dimension may expand thanks to the start-up of supply enterprises as a consequence of the decentralization of some phases of the production cycle, or thanks to the specialization of some enterprises focusing on particular tasks. The specialization in more and more specific functions favours the increase of the local stock of knowledge, as enterprises can concentrate on problems and inefficiencies otherwise neglected. Moreover, the presence of a high disintegration of the production process determines the need to build stable networks in order to coordinate the activities of the various enterprises, with a subsequent facilitation in the sharing of knowledge (Rosenberg, 1994). From a perspective of polycentrism as decentralized concentration, industrial clusters represent the nodes of a reticular and polycentric territorial structure. The aim of an innovative polycentrism corresponds on the one hand with the promotion of several innovative clusters, and on the other hand with the promotion of functional interactions between the clusters as a medium for the diffusion of knowledge and enhancing competitiveness through the previously mentioned potential synergic effects. In the following chapters, this theoretical framework will be translated into analytical parameters (Paragraphs 2 and 3) in order to analyse the polycentric and innovative phenomena in the Mediterranean Latin Arc (Paragraph 4). Moreover, this study will furnish some elements of discussion regarding the concept of polycentrism, which will then be discussed in the final pages (Paragraph 5).

2. Methodology 2.1 Data and territorial statistical units The research utilises, beside traditional input and output indicators concerning innovative processes, data concerning the presence of organizations and institutions that influence or may favour innovation in the territory. Such analysis starts with the identification of some areas characterized by a high endowment of technological functions in terms of enterprises, universities, polytechnics, and research centres operating in high-technology fields5. These are the main actors involved in the process of creating, diffusing, and locally reproducing (i.e. training and educational activities) knowledge (Saxenian, 1998). The identification of innovative territorial systems occurred with the intention to transpose into the empirical analysis the main characteristics of the innovative cluster evidenced from the literature:
-

proximity between socio-economic actors and, consequently, concentration of actors and resources in a limited space; specialization, based on the division of labor, in specific industrial sectors; variety of subjects and organizations operating in the territory; relativity of the size of the cluster: one small concentration of economic subjects may not be very meaningful in a strongly developed regional context, but it may constitute a regional development engine in one peripheral area.

Considering the proximity between the subjects implies the choice of an adequate territorial statistical unit of reference, pertinent to the study of local operating mechanisms of the cluster (interaction between local enterprises, solidarity and competition, imitation). In this framework, regional scales (Nuts 2 and Nuts 3) appear excessively wide, while considering the communal level would hide the relationships between the subjects localized in the wider metropolitan areas. This is the classic theoretical-methodological problem, often unsolved, of the concrete definition of self-contained territorial units, that is to say, the identification of the boundaries of a local territorial system in functional terms, and not on the basis of administrative partitions. The main instrument used for this purpose is the identification of a daily urban system, or the area in which the greater part of the daily life of the population is carried out; the local market within which the exchanges of goods, information, and services between human individuals and groups occurs. To this view, the data presented were initially collected on the communal scale and then aggregated to the scale of local systems: in the Italian case, local labour systems (sistemi locali del lavoro; Istat, 1997) were utilised, in France the zones demploi (Insee, 1998), while in Spain, the urban areas, as defined by the Direccin General de Vivienda (2000). Then, the share of enterprises located in every local system in ratio to the total regional (Nuts 2) high-tech industrial endowment was calculated for each of the six high-tech sectors considered, using the formula: Csl = Isl / Isr
where: C share (concentration) of enterprises I number of enterprises s high-tech sector
5

(1)

l r

local system (local labour system zone demploy or urban area) region (Nuts 2)

Six hi-technology sectors has been detected by means of Acs (1996) classification. They basically concern: i) biotechnology and biomedical; ii) information technologies and services; iii) high technology machinery and instruments, iv) defence and aerospace; v) energy and chemicals; vi) high technology research.

For the purposes of this analysis, those local systems with the following characteristics were considered (potentially) innovative clusters: 1. they concentrate at least one third of the total regional industrial endowment with reference to at least one of the six sectors considered; 2. they are characterized by the presence of at least one faculty or polytechnic, and at least one research centre, all of which are operating in a high-tech sector. The first criterion answers the need to identify the systems in relative terms, evidencing local situations of concentration and industrial specialization6. The established quota (one third within any of the six high-tech sectors), identified as a result of numerous empiric tests, potentially allows the identification of more than one industrial cluster within every region. The second criterion, referring to the combined presence of the variables considered, stresses the variety of actors involved in innovative processes. In fact, it must be remembered that the industrial cluster is not simply an area characterized by an elevated concentration of enterprises, but a place in which a number of interacting actors constitutes the basis of socio-economic evolutionary dynamics. This methodology has allowed the identification of 15 industrial clusters:
-

in Italy: Turin, Milan, Genoa, Padua, Bologna, Florence, Rome, Naples; in France: Lyon, Toulouse, Montpellier, Marseille-Aubagne; in Spain: Barcelona, Valencia, Seville.

The analysis was then extended in relation to these 15 local systems, by introducing additional data relative to the socio-economic framework conditions of the territorial systems, concerning the presence of institutions and organizations involved in the processes of technological innovation in the territory. Besides the use of traditional input and output variables7, data concerned8:
-

the presence of organizations involved in knowledge production and diffusion: besides the previously mentioned number of enterprises, academic institutions and research centres, science and technology parks were also considered, constituting important tools for innovative territorial policies; international aperture of the local system, measured through the following proxies: the number of international flights (inbound and outbound) during the year, the number of hotels, the number of international research projects involving local actors9, the presence of headquarters or offices of international organizations (as OECD or UN) and European Info-points. The international aperture of the cities represent a fundamental component of the urban innovative milieu, as pointed out by Amin and Thrift (1992);

The formula does not represent a real local specialization index: it has been created in order to identify local situations of specialization in relation to the general set up of the region. 7 These are expenditures (share of GDP) and employment (share of active population) in R&D activities; patent applications every 1.000 inhabitants. 8 All data refer to year 2000, with the exception of those concerning unemployment, Gdp pro capita PPP, Evca, Info-points and science parks (2001). Statistical sources: i) enterprises: Amadeus database by Bureau Van Dijk. Here are enterprises satisfying at least one of the following criteria: a) at least 1 million operating revenue; b) at least 2 million total assets; c) at least 15 employees; ii) banks: The Banker, 2001, n. 907; iii) international organizations: The Europa World Yearbook 2000, Europa Publications, London, 2000; iv) students, universities, polytechnics, research centres, scientific associations, museums and libraries: The World of Learning 2000, Europa Publications, London, 2000; v) science parks: membership directory of International Association of Science Parks, www.iaspworld.org; vi) patents, motorway density, flights, GDP pro capita PPP and unemployment: Regio 2002 database, Eurostat, Bruxelles. Data from Regio 2002 has not been detected at the local system scale, but at Nuts 2 scale or, where possible, Nuts 3 scale; vii) members of EVCA: www.evca.com; viii) hotels: The Hotel Guide cd-rom, The Hotel Guide, Meggen, 2000; ix) Info-Points Europe: www.europa.eu.int; x) fairs and expositions Guida Mondiale delle Fiere, Pianeta, Torino, 2000. 9 Data refers to the number of participations in international research projects and research networks financed by the European Commission during the 5th Framework Programme. Source: http://www.cordis.lu

the presence of facilities for innovative enterprises such as business and financial services, fairs and exhibitions, members of the European Venture Capital Association (EVCA), which is an organization supporting investments, particularly towards small and medium size enterprises operating in high technology sectors. Business and financial services are considered to be a supporting element for the start-up of firms, particularly in dynamic and innovative sectors (Mork and Yeung, 2001). Fairs and exhibitions represent a fundamental tool for the diffusion of information concerning new technological solutions proposed by enterprises (Rubalcaba-Bermejo and Cuadrado-Roula, 1995; Istat, 1998); cultural milieu, particularly important for the definition of social assets favouring an innovative climate (Aydalot, 1986; Saxenian, 1998). The presence of a cultural milieu has been measured in terms of proxies referring to the presence of museums and libraries of an international caliber, and universities and research centres operating in humanistic and basically non-technological fields10 socio-economic and welfare conditions, measured in terms of GDP pro capita and unemployment rates. The presence of a low ability to generate value together with high unemployment may be the consequence of a weak capacity to innovate, but at the same time situations of uncertainty, instability, and dissatisfaction may hinder innovative processes (Sweeney, 2001).

2.2 Statistical elaboration The elaboration has been carried out using some tools of multivariate statistics and, in particular, the calculation of the principal components. This statistical instrument allows to reduce the original number of variables (more than 20 in our case), replacing them with a smaller group of components (new variables) nearly as informative as the original group of variables. In our case, the calculation of the principal components, carried out on the basis of the standards of SPSS statistical software, allowed the identification of 5 components, comprising 86.93% of the original variance. The first component expresses both the level of internationalization and the scientific potentialities of the local systems. In fact, the component presents elevated correlations with variables concerning the international aperture, the presence of research centres and the participation in international scientific networks. The component evidences therefore a strong statistical tie often evidenced in literature: the ability to produce and to use knowledge is closely connected, in fact, with the insertion of the territorial system in the supra-local networks. The second component describes the technological and productive milieu, introducing high correlations with the presence of subjects operating within (or in close contact with) the industrial sector, of business services, of codified knowledge directly usable in production cycles (i.e. patents). The technological and productive milieu contemporarily constitutes both a cause and a consequence of the achievements of innovative processes: on the one hand, a dynamic milieu is able to utilize and to transform innovations resulting from the research activities into competitive advantage; on the other hand, the innovative milieu can stimulate the demand for innovations and energize research activities, for example, through financing. The third component describes the potentialities of the territorial systems in the local reproduction of knowledge, introducing high correlations with the existence of academies and science parks. The ability to reproduce locally the knowledge coming from outside the cluster plays a fundamental role in technological competition, favoring the use in the industrial sectors of innovations coming from research centres and academies: this allows the use and the imitation, within the cluster, of external
10

The category has been detected in a residual manner with respect to high technology sectors.

knowledge, while the formation activities carried out by the universities increase the availability of skilled and specialized human capital. The fourth component introduces a high correlation with GDP per capita and one inverse correlation with unemployment rates. The component describes therefore the conditions of socio-economic welfare, an aspect of remarkable importance not only for the well-being of the population of the local system, but also for innovative processes: innovation is not a phenomenon taking form in laboratories and research centres isolated from the local territorial context, but it is a social process involving a number of different actors and organizations. A climate of diffuse well-being can therefore facilitate social relations, favoring the spread of trust, cooperative attitudes, and personal satisfaction. From a closely economic point of view, moreover, the presence of positive economic conditions favors an increase in the local demand of goods and services that promotes the industrial development (Morck and Yeung, 2001). Finally, the fifth component is explained from traditional input variables relative to the innovative process, that is the expense and employment rates in research and development activities (in both public and private sectors). It is now possible to classify the local systems on the basis of the factorial scores obtained. The proposed classification concerns not only the technological performances of the local systems, but especially the endowment of factors and context conditions that may enhance and favour innovative processes: in this sense, it represents the potentialities of the local system in the field of technological innovation. For this reason, some rather different cases compare in the same groups: a first level refers to strongly innovative systems. Here all the five factors present high scores and, at the same time, there are no meaningful points of weakness. Barcelona and Milan represent by far the most innovative nodes, while Lyon and Rome represent second level systems; the second group includes moderately innovative systems, a wide category including local systems with high technological potential, but with evident problems concerning one of the five components. Here we find Naples, Marseille-Aubagne, and Montpellier, presenting socioeconomic problems (above all in terms of unemployment), together with a certain industrial weakness in the cases of Naples and Montepellier. Turin is included here as a consequence of low levels of international aperture, while the case of Florence is peculiar: even if the system does not reveal specific weaknesses, the technological potential is not really high from every point of view. For this reason Florence is included among the moderately innovative systems, and not in the previous typology; the third group includes the traditional systems, i.e. urban systems characterized by good economic and industrial performances, but scarce levels of innovation. In the cases of Bologna, Padua, and Valencia there is therefore the risk of traditional specializations and the necessity to increase an innovative capacity. Finally, a last group refers to declining systems. Here a low innovation capacity is linked to evident phenomena of unemployment and industrial decline, particularly in Seville. In addition to a better situation in the field of innovation activities, Toulouse is included in this group because of phenomena of industrial recession and deindustrialization. A similar industrial situation is present also in Genoa, but it must be pointed out that the city is actually experiencing an extensive transformation of the old economic base (with an increasing importance of the maritime port and important investments in the cultural sector). This process does not influence the old data used (as reported in note 7, data basically refers to 2000).

3. Hierarchy and polycentrism in the Arc Beyond the innovative potential, it is important to estimate the role carried out by the various systems in regard to their regional systems: in fact, one local system can act as an innovative engine for its wider (regional) territorial contest, while another one may simply represent one hierarchically dominant area located in a weaker regional context. Considering this potentially different cluster role, the regional organization of the technological phenomenon has been analyzed within the theoretical framework of polycentrism, distinguishing the territorial systems that have activated processes of diffusion and development from those that, on the contrary, seem to be isolated from regional technological dynamics. For this purpose, the theoretical concept of polycentrism has been transposed in the analysis emphasizing two different analytic dimensions: the presence of several centres, or a widespread territorial structure, and the existence of relations between the centres, or the degree of interdependence within the regional structure. Therefore, the concentration of high-tech enterprises (Formula 1) have been considered once again. The index supplies two important elements of analysis: first of all, the presence of numerous industrial concentrations describes basically widespread regional productive structures, while the presence of a few industrial areas is characteristic of polarized situations. In the second place, it is presumed that the presence of common industrial and technological specializations between several local systems within the same region allows a fast spread of knowledge and innovations, on the basis of the mechanisms described with reference to the horizontal and vertical dimensions of the cluster. The presence of these shared specializations can therefore reveal the degree of interaction (also in this case, obviously, potential) present between the various local systems. Therefore, one cartographic representation has been used so as to highlight the territorial systems characterized by meaningful industrial concentrations and the contiguities in the technological specializations between geographically close local systems: in Figure 2 the innovative systems previously analyzed are represented by dark circles, while the bright ones represent those local systems that concentrate at least 20% of the total regional endowment of enterprises in at least one of the high-tech sectors under consideration. Finally, those local systems close to each other (in the same region, adjacent or separated by another local system) sharing one or more specializations are connected with lines of different thickness, in relation to the number of shared industrial specializations11.

11

More in detail, the presence of one common specialization has been classified, with reference to Figure 2, as modest industrial contiguities, of two as significant, of three as intense.

Fig. 2 Concentrations and contiguities in industrial specializations

Industrial contiguities as potential interactions

It is now possible to classify the regions on the basis of the two analytic dimensions of polycentrism previously evidenced. The first refers to the degree of territorial spread, identified by the number of meaningful industrial concentrations. We will distinguish between regions characterized by an elevated degree of spread (five or more industrial concentrations), an intermediate degree (three or four), and a low degree of industrial spread (one or two), i.e. situations of strong polarization. The second dimension regards the number and the intensity of contiguous industrial specializations (the lines in Figure 2). In such a way it is possible to distinguish the situations of fragmentation from those of interdependence12, which is presumed will favor the diffusion of knowledge through imitation processes along the horizontal dimension of the cluster. It is now possible to reduce those results to an unique analytical dimension, concerning the degree of polycentrism of the regions, classifying them in the three categories as evidenced in Figure 313.

12

The classification is the following: low interdependence (0-1 contiguities), medium (2-3) and high (4 or more). The connections of double intensity (2 common specializations) have been considered as weight 2, while those of triple intensity as weight 3. 13 Lombardy represents a peculiar case: Milan is characterized by one concentration of enterprises so high to obscure, de facto, the important endowment of enterprises in the rest of the region and, for this reason, observing Figure 2, one can argue that the region would have to be classified as strongly polarized. To avoid this mistake, the statistical analysis for the region has been run again, excluding data concerning Milan, and this second analysis confirmed the strong polycentrism of the region (for the details see Vanolo, 2003). This is a feature of the Lombardy region already evidenced in the literature, for example by Pompili (2002). For these reasons, the region has been classified, in Figure 3 and hereinafter, as polycentric.

10

Fig. 3 Hierarchy and polycentrism in regional structures


Provence-Alpes-Cte dAzur Languedoc-Roussillon Lombardy Andalusia Regi n de Murcia Rhne-Alpes Tuscany Campania Lazio

Polycentrism

high

Diffusion

medium

Midi-Pyrnes Catalonia Veneto Liguria

Polarization

Intermediate situations
Emilia Romagna Piedmont Comunidad Valenciana

low low

medium

high

Interdependence

4. Towards some conclusive hypotheses The results of the analyses, both those concerning innovative potentialities (with reference to the classification proposed in Paragraph 2.2) and those concerning hierarchical or polycentric territorial regional structures within the Mediterranean Latin Arc (Figure 3), have been further summarized using Figure 4. Fig. 4 Technological potential and polycentrism

Technological potential

high

Barcelona

Milan Lyon, Rome

medium Padua low Genoa, Toulouse low

Turin Bologna, Valencia

Florence, Naples, Marseille, Montpellier

Seville medium high

Degree of polycentism

It is not possible to determine here the existence (and the logical direction) of a causal relation between polycentrism and technological potential. However, the hypothesis concerning the supposed facilitation of knowledge diffusion in polycentrical contexts seems to be conceivable: Figure 4, in fact, suggests a certain direct correlation between the two variables, with the sole exclusion of the two cases of Barcelona and Seville. Moreover, Figure 4 allows the identification of some typologies of local systems:

11

Milan, Lyon and Rome confirmed their role as important innovative engines. In fact, they combine a high technological potential with a marked polycentric regional structure. It can be supposed, therefore, that they carry out an important propulsive role for the development of their own regional contexts. In particular, Milan represents the main Mediterranean innovative engine; other local systems present strong widespread and innovative potential, even if less intensely than the previous ones. This is the case of Florence, Naples, Marseilles, Montpellier, and Turin. These centres can be considered as second level innovative engines; a particular case is represented by Barcelona, characterized by optimal technological performances (a little weaker than those of Milan) accompanied by one strongly hierarchical regional territorial structure. It seems therefore that the system acts more as a centripetal pole than as a innovative engine; another specific case, diametrically opposite to the one of Barcelona, is relative to the Andalusian polycentric system. The region, even if characterized by modest innovative performances, introduces high levels of polycentrism; Bologna and Valencia are characterized by a limited innovative potential and a moderately polycentric regional structure. The two centres are therefore classifiable as traditional systems; Genoa, Toulouse, and Padua (even if the latter with lightly better innovative performances) represent weak poles, characterized by insufficient performances in relation to both dimensions of the analysis.

Figure 5 presents these results visually, evidencing some elements of the territorial structure of the Mediterranean Latin Arc:
-

with the exclusion of the Barcelona pole, the Iberian area reveals itself to be extremely weak under the innovative profile. Nevertheless, it is important to notice the elevated polycentrism of the Southern area, with particular reference to Andalusia and Murcia, while the Catalan technological potential appears nearly exclusively concentrated in Barcelona; in Italy, the Western Po axis (Piedmont and Lombardy) and the Centre regions introduce elevating technological potentials and polycentric territorial structures. The point of contact between these two areas is the hinge of Bologna, characterized by elevated levels of socio-economic welfare, but from limited technological capabilities. The most problematic territorial systems are Padua and Genoa, characterized by technological and industrial weakness; the French-Italian axis constitutes one of the main elements of strength of the Latin Arc, thanks to the two innovative engines of Milan and Lyon. It must be noticed how Genoa and Toulouse present problematic socio-economic and technological situations, let alone by scarce polycentric structures.

12

Fig. 5 Innovation in the Mediterranean Latin Arc


Industrial contiguities as potential interactions

5. Conclusions: some theoretical remarks about areal and reticular polycentrism The empirical analysis can supply some cues, of a theoretical order, for a deepening of the concept of polycentrism. In particular, polycentrism has been identified, on the regional level, by the intersection of two analytical dimensions: 1. the first refers to the more common and intuitive meaning of polycentrism, that is the presence in the territory of several polarities within a basically widespread industrial fabric; 2. the second dimension considers the eventual presence of meaningful relations and interdependences between the centres. Splitting the concept of polycentrism into these two analytic dimensions allows us to introduce one simple, but nevertheless useful, terminological distinction: areal polycentrism will indicate those territorial configurations in which it is satisfied only the first analytic dimension, relative to those areas characterized by the presence of several polarities, but without any significant relations and interdependences between them; the reticular polycentrism will refer, instead, not only to the presence of several poles of development (area polycentrism), but also of networks and interactions between them.

The two terms introduce huge differences in terms of policies and strategies, as summarized in Table 1.

13

Table 1 Areal and reticular polycentrism


Areal polycentrism
Several isolated poles

Reticular polycentrism
Several interdependent nodes

Innovation Competition Innovation policies Objectives

Punctual innovative processes Specialization Diffusion-oriented Diffusion

Innovative systems Complementary networks Network-oriented Local development

The distinction between areal and reticular meanings assumes particular importance with reference to the mechanics of learning and production/reproduction of knowledge. While in areal polycentrism the innovative phenomena are essentially punctual, regarding the pole inner technological capabilities, in the case of reticular polycentrism the ability to innovate becomes a regional feature, deriving from the interaction between several poles; the development of complementary networks is therefore worthwhile. This perspective assumes particular importance in relation to the processes of territorial competition: while in the case of areal polycentrism the strategies of territorial competition regard, above all, the single poles, in the case of reticular polycentrism it is possible to formulate, more than the strategies on an urban scale, also cooperative strategies for the development and the promotion of the regional system as a whole. Moreover, it is interesting to consider how the two meanings of polycentrism are characteristic of two distinguished approaches to innovation policies. The areal perspective is distinctive of the traditional diffusion-oriented policies, aimed at the diffusion of knowledge towards the peripheral local systems. Reticular polycentrism is coherent with network-oriented policies: this is the more recent orientation of the Communitarian policies, aimed to stimulate not only the diffusion of knowledge, but also the promotion of networks, so as to pursue the participation of the peripheral and marginal local systems in the processes of creating value and knowledge. The mere availability of knowledge, in fact, does not constitute an advantage for a centre if there is no innovation demand and, further more, the capability to transform knowledge into competitive advantage for the local enterprises. The formation of a regional polycentric system, instead, can allow the functional specialization and the insertion of the centres into the economic processes on a regional scale. After all, areal polycentrism seems to recall the theoretical contributions of postwar regional sciences (in particular Myrdal, 1957), where much of the debate regarded the promotion of the spread (transfer of capital and productive factors) and backwash effects (of development) from the dominant pole towards the rest of the territory. Instead, reticular polycentrism answers wider objectives of local development for the regional system. In fact, reticular organization is not only the basis for development-diffusion effects, but also defines the modalities of collective learning of the territorial systems. Reticular polycentrism must not therefore be considered as a sum-zero distribution process, but as a development path based on the activation of those endogenous processes operating only on a supra-local scale. This interpretation of reticular polycentrism, deriving from the superimposition of the concepts of diffusion and territorial interaction, may be useful in the theoretical framework. First, this is applicable to different scales, such as the European one, where the objective of forming one European Research Area, (European Council of Lisbon, March 2000), can be interpreted as an objective of polycentric technological development. Secondly, it becomes explicit how hierarchy and polycentrism are not in opposition and do not exclude each other. The presence of dominant centres of differing importance and visibility is not in contrast either with the theoretical definition, or with the

14

political objectives of reticular polycentric: in a reticular systems, the major urban centres may act as gateways for participation in the hierarchical lower urban systems in the supra-local networks. Finally, reticular polycentrism is characterized by wide superimpositions with the concept of industrial cluster. In fact, as evidenced in the empirical analysis, polycentrism can be thought as a sort of network of networks: the nodes of the polycentric regional structure are constituted by the local systems rooted in the territory, i.e. clusters and networks of local actors. The promotion of networks therefore constitutes a multi-scalar objective, indispensable for the promotion of innovation both at local level (horizontal and vertical dimensions of the cluster), and on wider scales (regional or Communitarian polycentrism, insertion into international research networks). In this framework, the polycentric and cluster oriented policies, if opportunely contextualised (or, to use a different expression, territorialized), may promote strong synergies.

References
Acs Z. (1996), American high-technology clusters, in J. de la Mothe and G. Paquet (eds.), Evolutionary Economics and the New International Political Economy, Pinter, London, pp. 183-219. Amin A. and Thrift N. (1994), Globalization, Institutions and Regional Development in Europe, Oxford University Press, Oxford. Aydalot P. (ed.) (1986), Milieux innovateurs en Europe, Gremi, Paris. Bailey N. and Turok I. (2001), Central Scotland as a polycentric urban region: useful planning concept or chimera? Urban Studies, v. 38, n. 4, pp. 697-715. Boulifard C. (1994), Images et scnarios de dveloppment de l' Arc Mditerranen: un tat des Lieux, Quaternaire, Paris. Brunet R. (1989), Les villes europennes, Datar-Reclus, La Documentation Franaise, Paris. Conti S. and Salone C. (2001), Il sistema urbano europeo fra gerarchia e policentrismo, in Presidenza del Consiglio dei Ministri Dipartimento per i Servizi Tecnici Nazionali (ed.), LItalia nello spazio europeo. Economia, sistema urbano, spazio rurale, beni culturali, Gangemi, Roma, pp. 53-68. CRPM - Confrence des Rgions Priphriques Maritimes dEurope (2002), Study on the Construction of a Polycentric and Balanced Development Model for the European Territory, Crpm, Rennes. Danson M. and Whittam G. (1999), Clustering, innovations and trust: the essentials of a clustering strategy for Scotland, in M. M. Fisher, L. Suarez-Villa and M. Steiner (eds.), Innovation, Networks and Localities, Springer, Berlin, pp. 66-84. Datar (2000), Amnager la France de 2020. Mettre les territoires en mouvement, La Documentation franaise, Paris. Dematteis G. (1995), Progetto implicito: il contributo della geografia umana alle scienze del territorio, Franco Angeli, Milano. Direccin General de la Vivienda, la Arquitectura y el Urbanismo (2000), Atlas estadstico de las reas urbanas en Espaa, Ministerio de Fomento, Centro de Publicaciones, Madrid. European Commission (1994), Europe 2000+. Co-operation for the Spatial Development of Europe, Office for Official Publications of the European Communities, Luxembourg. European Commission (1999a), Sixth Periodic Report on the Social and Economic Situation and Development of the Regions of the European Union, CEC, DGXVI, Luxembourg. European Commission, Committee on Spatial Development (1999b), ESDP. European Spatial Development Perspective. Towards Balanced and Sustainable Development of the Territory of the European Union, Office for Official Publications of the European Communities, Luxembourg. Gertler M. S. (1996), Worlds apart. The changing market geography of the German machinery industry, Small Business Economics, v. 8, n. 1, pp. 87-106. Hall P. (1993), Forces Shaping Urban Europe, Urban Studies, v. 30, n. 6, pp. 883-898.

15

Hall P. (2001), Christaller for a global age: redrawing the urban hierarchy, Globalization and World Cities Study Group and Network Research Bulletin, n. 59; http://www.lboro.ac.uk/gawc Insee (1998), Atlas des zones demploi, Insee, Paris. Istat (1997), I sistemi locali del lavoro 1991, Istat, Roma. Istat (1998), Statistiche sulla ricerca scientifica e l' innovazione tecnologica, Istat, Roma. Kloosterman R. C. and Musterd S. (2001), The polycentric urban region: towards a research agenda, Urban Studies, v. 38, n. 4, pp. 623-633. Krugman P. (1991), Geography and Trade, MIT Press, Cambridge (Ma). Loasby B. J. (1999), Industrial districts as knowledge communities, in M. Bellet and C. LHarmet (eds.), Industry Space and Competition. The Contribution of Economists of the Past, Edward Elgar, Cheltenham, pp. 70-85. Malmberg A. (1996), Industrial geography. Agglomerations and local milieu, Progress in Human Geography, v. 21, n. 4, pp. 573-582. Malmberg A. and Maskell P. (1997), Towards and explanation of industry agglomeration and regional specialization, European Planning Studies, v. 5, n. 1, pp. 25-41. Malmberg A. and Maskell P. (2002), The elusive concept of localization economies: towards a knowledgebased theory of spatial clustering, Environment and Planning A, v. 34, n. 3, pp. 429-449. Meijer M. (1993), Growth and decline of European cities: changing position of cities in Europe", Urban Studies, v. 30, n. 6, pp. 981-990. Morck R. and Yeung B. (2001), The Economic Determinants of Innovation, Industry Canada Research Publication Program, Ottawa. Musterd S. and van Zelm I. (2001), Polycentricity, households and the identity of places, Urban Studies, v.38, n.4, pp.679-696. Myrdal G. (1957), Economic Theory and Underdevelopment of Regions, Duckworth, London. Norton R. D. (1979), City Life-Cycles and American Urban Policy, Academic Press, New York. Pere A. and Toms S. (1998), El Arco Mediterraneo Espaol: sus perspectivas como espacio de futuro, Revista valenciana destudis autonmics, n. 22, pp. 23-41. Pompili T. (2002), Leconomia del sistema policentrico lombardo, Scienze Regionali - Italian Journal of Regional Science, v. 1, n. 3, pp. 73-109. Porter M. E. (1998), Clusters and the new economics of competition, Harvard Business Review, v. 76, n. 6, pp. 77-90. Rosenberg N. (1994), Exploring the Black Box: Technology, Economics and History, Cambridge University Press, Cambridge. Rubalcaba-Bermejo L. and Cuadrado-Roura J. R. (1995), Urban hierarchies and territorial competition in Europe: exploring the role of fairs and exhibitions, Urban Studies, v. 32, n. 2, pp. 379-400. Saxenian A. (1998), Regional systems of innovation and the blurred firm, in J. de la Mothe and G. Paquet (eds.), Local and Regional Systems of Innovation, Kluwer, Norwell, pp. 29-43. Scott A. J. (1988), New Industrial Spaces, Pion, London. Steiner M. (ed.) (1998), Clusters and Regional Specialisation, Pion, London. Storper M. (1993), Regional worlds of production: learning and innovation in the technology districts of France, Italy and the Usa, Regional Studies, v. 27, n. 5, pp. 433-455. Suarez-Villa L. and Walrod W. (1997), Operational strategy, R&D and intra-metropolitan clustering in a polycentric structure: the advanced electronics industries of the Los Angeles basin, Urban Studies, v. 34, n. 9, pp. 1343-1380. Sweeney G. (ed.) (2001), Innovation, Economic Progress and the Quality of Life, Edward Elgar, Cheltenham. Vanolo A. (2003), Per uno sviluppo policentrico dello spazio europeo. Sistemi innovativi territoriali nellEuropa sud-occidentale, Franco Angeli, Milano. Williamson O. (1975), Markets and Hierarchies. Analysis and Antitrust Implications. A study in the Economics of Internal Organization, Free Press, New York. Zipf G. K. (1949), Human Behaviour and the Principle of Last Effort, Addison Wesley, Cambridge (Ma).

16

Você também pode gostar