Innovation in Clusters. Estelle Vallier

of his cluster concept. He transfers to territories an industrial logic of building assets in a competitive relationship within a market (Porter 1993). Specialization enables a comparative advantage to be put forward. In a capital mobility context, we are witnessing an explosion of competition between territories to attract the most qualified workers and businesses. Advantages are not given, but deliberately constructed (Carré and Levratto 2011, p. 360). From the point of view of economic development policies, this implies a shift from a territorial policy focused on local redistribution issues to a policy focused on externalities and attractiveness issues (Béhar et al. 2013, p. 20).

As early as 1989, David Harvey, in a critical view, drew attention to this shift from “managerialism to entrepreneurialism” in local policies. The “managerialism” of the 1960s was focused on services to residents, whereas “entrepreneurialism” is based on policies oriented towards an exogenous economic development strategy (Harvey 1989, p. 3). Harvey, too, places this orientation in the context of the recomposition of capitalism since the 1970s (the transition from a Fordist-Keynesian regime of capital accumulation to a regime of flexible accumulation), which has led to the development of interurban competition for resources, jobs and capital. This competition has been largely accompanied, even encouraged, by the state, notably through the logic of the call for projects, which acts as a means of territorial differentiation (Brenner 2004). By entering into this competition, local public actors have become veritable promotion agencies for their territory in order to attract investors, especially from overseas. The standard-bearers of this promotion are therefore the specific advantages that the territory possesses compared with its rival territories (Carré and Levratto 2011, p. 336). Beyond this mission of promotion, the public actor must take charge of the economic and social regulation that was previously assumed by the large Fordist enterprise that du Tertre and Laurent define under the term territorial regulation as:

      The capacity of a localized and institutional social system to guarantee accumulation processes in its territory, to maintain economic activities there, to develop and protect its natural resources and technological achievements and to guarantee social peace. Territorial regulation thus results from the necessity to regulate the conditions through which sectors are deployed and resources are allocated (Laurent and Du Tertre 2008, p. 15).

      In this context, the public actor must encourage, mobilize, coordinate and structure interactions and resources (Béhar et al. 2013, p. 17). Thus, using the logic of Porter’s comparative advantage, Anna Lee Saxenian questions the difference in development between the American West Coast innovation hub, Silicon Valley, and the East Coast hub, Route 128 (Saxenian 1994). By looking at the structure of the social networks of the two clusters, Saxenian opens the way to an abundance of literature on the role of geographical proximity, on the typology of commercial and scientific exchanges and on the place of non-market interindividual relations in the coordination of actors (Lanciano-Morandat et al. 2009, p. 178).

1.3. The valleys of knowledge: interindividual relations as a source of innovation

      The end of the 20th century saw the emergence of a craze for valleys conducive to industrial and technological development, thanks in particular to the informal links forged by geographical proximity.

      1.3.1. Informal links in the heart of Silicon Valley

Since Saxenian’s work on Silicon Valley, the cluster concept is based on the idea that it can be crossed by interindividual relations that are conducive to informal exchanges, which are themselves conducive to the development of scientific and industrial collaborations (Saxenian 1996). The author explicitly refers to Porter’s comparative advantage. Indeed, she attempts to uncover the advantage that Silicon Valley has over its Bostonian counterpart, Route 128. For her, the development of the Californian cluster is strongly linked to the structure of the social networks that cross it. Based on regional data, Anna Lee Saxenian notes that Route 128 created three times fewer jobs in the early 1990s than its Californian neighbor (Saxenian 2000). It was less efficient, particularly in terms of its volume of exports of electronic products and its ability to concentrate the most successful national companies. According to the author, economists neglected the role of social networks in the success of such industrial concentrations, and, for their part, sociologists implicitly accepted this neglect, reinforced by the idea that market processes are not suitable objects for sociological study, since social relations play only a minor role in them (Granovetter 1985). Thus, faced with two distinct industrial systems with similar characteristics (similar technology, set up after the war, etc.), Saxenian proposes an approach based on social networks in order to distinguish them and explain their different developments. For her, Silicon Valley is based on dense learning networks between actors from related technologies that encourage entrepreneurship and experimentation. Competition is intense, but it is supported by informal communication systems that create mutual adjustments and learning.

      In the eastern United States, Route 128 is more composed of significant autarkical structures and the links between individuals are therefore characterized by a principle of corporate loyalty (Saxenian 2000). In the relationships between businesses, customers and suppliers, individuals refer first and foremost to the vertical hierarchy of the organization to which they belong, before circulating information horizontally between firms. For Saxenian, the performance of Silicon Valley, or its advantage, to use Porter’s terminology, depends on local social and institutional determinants, such as professional or commercial organizations, far more than economic and fiscal factors (wage differentials, real estate costs, local taxes). For example, The Semiconductor Equipment and Materials International or The Software Entrepreneur’s Forum organize meetings, dinners, seminars and training courses and participate in the creation of these social networks (Grossetti 2004, p. 170). From this point on, we can observe how interactions embedded in a given framework (in this case the businesses in the Valley) can be decoupled in another context (professional association, for example).

In another study of Silicon Valley, Michel Ferrary particularly observes the social networks of the French community living and working in the Valley. These French expatriates meet in “non-economic social institutions” (the French school in Palo Alto, the Maison Française, the film club, French restaurants), ethnic celebrations (the July 14 national holiday, the Beaujolais Nouveau day), newspapers and associations with a more economic purpose (Doing Business in French, Eurotrash, Silicon French). They develop strong ties of friendship, support and appreciation through cultural and social activities during which they meet (Dibiaggio and Ferrary 2003). According to Ferrary, they mobilize this community in the economic activities in which they are engaged. We can therefore see that literature highlights the virtuous development of these territorial concentrations by emphasizing the effects of geographical proximity in the constitution of social networks.

      1.3.2. The relational logic essential to geographical proximity

Since the beginning of the 1990s, the economics of proximity (Bellet et al. 1993, 1998; Rallet and Torre 1995; Gilly and Torre 2000; Dupuy and Burmeister 2003; Pecqueur and Zimmermann 2004; Torre and Filippi 2005) has sought to take account of the conditions necessary for the coordination of agents (Talbot and Kirat 2005, p. 9). This question has been the subject of research concerning the links, or otherwise, between geographical proximity and social interactions. The abundant work has led to “the emergence of a kind of specific scientific community, with its own conferences and regular publications” (Bouba-Olga and Grossetti 2010, p. 2). In the course of their work, this school has managed to define a certain number of types of proximity, but a concern for readability has led its authors to restrict the canonical forms of proximity to two (ibid., p. 4). Thus, they agree to say that alongside geographical proximity is organized proximity. The former is doubly relative: on the one hand, it is relative to the means of transport and, on the other hand, it proceeds from the judgment of the individuals on the distance that separates them from other individuals or groups. The latter is not geographical, but relational in nature and is based on an organization’s ability to have its members interact (Talbot and Kirat 2005). Beyond this vertical distinction, the authors have made a horizontal division within organized proximity between the logic of belonging