2.1. Justifying public intervention

Several works have focused on this topic ([DUR 08], [CHA 13], among others), and our remarks will be in line with theirs in many respects. Traditionally, the goal of public intervention is to help localized companies in areas in decline or, on the contrary, in those on an economic upswing. The literature uses several arguments to justify this position: the existence of localized scale economies, externalities, the reduction of regional imbalances and credit constraints.

First, concentrating companies in a given area has some benefits. A denser market favors a more effective use of the existing infrastructures, while also driving companies to specialize more. We also note that the presence of companies that use highly specific components will encourage a provider to be situated near the clients, lowering the transport costs. When agents are in the same place, we should add to this the decrease in the transaction costs involved in exchanging goods or transferring knowledge, causing the appearance of technological externalities [DUR 08].

Second, public policies make it possible to internalize externalities by favoring the geographic concentration of the actors. Providing public aid to a company can be analyzed as an advance that will be paid back by this company, whose project is successful. This also means that public bodies fund the import of technological knowledge that would have tended to spread. Moreover, if public action is oriented towards the creation of human capital, externalities may be produced by this human capital whose mobility ensures the dissemination of knowledge.

Third, public interventions have a redistributive aspect. In several countries, they are implemented with the aim of slowing down the decline of certain geographic areas (we will come back to this point during the analysis of competitiveness poles) by promoting the creation of companies. The principle of this action derives from the idea that public aid should be targeted at areas rather than people.

Finally, there are faults in the credit market that make small companies, which are often discriminated against, suffer [FAI 12]. The funded projects must still be profitable and companies, which are constrained by access to credit, must be actually innovating. In this field, public authorities can get involved directly by means of ad hoc agencies. In countries with regional stock exchanges, for example the United States, small companies may obtain equity through venture capital. On the contrary, when regional bodies have jurisdiction and authority, a regional credit system may be involved in co-financing or the provision of loan guarantees. This method, which is implemented in Germany, prevents the private sector from running risks that are too high. Nonetheless, we face the issue of knowing whether public bodies are always competent enough to distinguish between companies and, at a push, whether it is a comparative advantage for public bodies to behave like venture capitalists¹.

In relation to these arguments, public policies have two aspects: they must take into consideration the specific nature of a geographic location (existing industrial base, presence of universities and training centers, etc.) and the identity of the beneficiaries. Different levels of intervention may be involved. By incorporating a city in a cluster, neighborhood policies (for example, the implementation of specific infrastructures) amplify the dynamics of the city and, consequently, make it possible to direct resources which will favor the creation of new companies towards a specific area. A public agency may ensure the creation of a network among the companies.

As for the choice of target, public action directed at clusters tends to favor new activities, based on the hypothesis that these activities produce significant positive externalities. The problem becomes trickier once, for different reasons, intervention wants to take into consideration companies as well. For example, the works carried out in the USA indicate that a more arbitrary policy (targeted at certain companies) may turn out to be counterproductive. For example, removing non-compete clauses favors start-ups to the detriment of established firms, increasing the instability of the cluster. However, let us point out that the public policies used in this country in favor of small companies are not the policies inherent to the existing firms because of their small size and the difficulty involved in determining the firms rather than the sectors that are turning a profit.

Overall, the issue is the following: how can we build a coherent group of actors in a given area in order to favor innovation, growth and employment? Agrawal et al. [AGR 12] wonder about what combination of actors is most favorable for innovation. In line with the anchor tenant approach, these authors think that innovation is favored when, in a given area, there is a large innovative company coupled with a sufficient number of small businesses that are themselves innovative. This structure makes it easier to capture the externalities in the area while also keeping a fertile ground for the creation of start-ups, which are often founded by employees coming from large established companies. Moreover, this structure becomes richer when certain areas possess institutional assets (universities, research centers, etc.) that attract existing start-ups. It is difficult to identify clearly, in the works mentioned, the role of local innovation in the growth of local employment. However, establishing connections between these actors is crucial for the creation of a link between regional wealth and innovation.

If the externality-based argument becomes the leading one, we still need to acknowledge that whether externalities are actually produced, the cluster policy becomes meaningful only if these spillovers are nonlinear or if the companies that produce externalities of all kinds fully benefit from this effect. This emphasizes the issue of the degree of interaction between the parties involved and the optimal size of a cluster.

In this context, the externality dynamics lead us to group companies as much as to limit their clusterization. We can easily imagine a scenario where each operator sends and receives different flows from its different neighbors when fifty of them interact but we can also see that these flows decrease significantly beyond this threshold. It seems wise then to avoid strengthening all the operators in a single location because of the existence of markedly decreasing returns on the externalities. As we have seen, limiting the scale may be a consequence of diversity. The creation of another ecosystem would be justified only if it were coupled with a specialization directed by public authorities towards non-redundant technologies and productive processes.

We can point out a few things before the end of this section. We must first admit the existence of natural clusters, namely groupings of companies due to strong interactions that make us think there may be significant spillovers among them. Krugman, while revisiting Marshall, describes the process as follows². First, a small number of producers support local providers of specialized products (inputs), favoring the creation of external scale economies. Later, this grouping gives rise to specialized competitiveness poles that benefit workers and especially the flexible workforce of localized companies. Finally, knowledge spillovers take place due to the movement of workers from company to company. In terms of innovation, this results in opportunities to reduce transaction costs due to the possibilities of exchanging more tacit and highly specialized knowledge within this localized industry.

The argument often supported by the so-called natural clusters remains paradoxical: it simultaneously makes it possible to justify a cluster policy and to question, at least in part, the effectiveness of public action. Thus, we still need to justify public intervention through its coordinating and driving function in terms of cluster development. Constrained by the existing base built on private strategies, public policies must take into consideration the current stage of development of natural clusters and consolidate it. This may be done by establishing relationships between university research and the creation of small companies or by avoiding excessively narrow industrial monocultures. Coherent technological diversity (called related variety) is associated with more significant R&D spending.

The fact remains that public intervention is justified when the activity of the cluster increases. In this case, “private marginal returns (profit for the companies) are less than social returns (profit for society as a whole) and, from this point of view, we may think that the geographic concentration generated without public intervention will be suboptimal” [DUR 08]. However, as the authors of this article aptly point out, the fact that various mechanisms (economies of scale, externalities, financial shortcomings) are at work in different markets (inputs, idea, credit) to increase local returns implies that it will be difficult to identify which specific factors bring about the appearance of economic profit in clusters. Two elements should be taken into account. On the one hand, public authorities must have the expertise and skills required to analyze the specific mechanisms at work in each cluster. In a hi-tech cluster, knowledge spillovers are predominant and public action differs in this case from those involving the production of the qualifications required by companies. On the other hand, we should take into consideration the autonomous mechanisms at work within a cluster, especially the role played by large companies, at once as a spin-off site and a place of selection of the providers of products or knowledge. Here, we once again come across the systemic aspect of the formation and development of clusters. At any rate, thanks to the often complementary action of public and private actors, the costs of public action may be limited. From this perspective, public authorities have the possibility of redeploying existing resources instead of allotting new financial resources, especially in relation to the stage of development of the ecosystem.

Public policies have another role to play through immigration policies. Recent studies carried out in the USA indicate that immigrant scientists and engineers have been responsible for more than half of the increase in this qualified population since 1995 [KER 10]. “Immigrant” entrepreneurs are also commonly involved in the development of start-ups for the marketing of new technologies. In this context, immigration policies may create “supply shocks” that quantitatively and qualitatively influence innovative workforces in different areas.

Finally, let us point out that the structuring of activities into clusters may represent forces that go beyond the existence of spillovers. The presence of an appealing and dynamic large city near or within a cluster (San Francisco Bay, New York City) or of world-renowned institutions (universities, research centers) may cause groupings that go beyond the scope of strictly economic effects.

2.2. Innovation and production ecosystems and open innovation

Innovation and production ecosystems, such as we have analyzed them, rely on the adoption of open innovation forms that need to be spelled out. On a theoretical level, the externalities involved in making innovation more open mobilize arguments of three kinds [ROP 13]. The most widespread (micro-econometric) approach emphasizes the idea that the firms operating in R&D-intensive sectors benefit from spillovers of informal knowledge that increase their productivity. In this context, pure knowledge spillover effects are linked to the average R&D spending of the sector. We have acknowledged that these effects are strengthened when they occur within regional clusters.

A conceptually different approach highlights how actors (companies, laboratories, universities, etc.) establish cooperative relationships in the shape of partnerships or joint ventures. The argument used here is that the knowledge created by a firm may benefit other organizations, provided that the latter are involved in the processes of collective innovation. In this case, the organization, or even the scale, of the R&D process is altered and the relationships built this way increase the innovation outputs of each actor as patents or new products. An extreme form of this process takes place when the inventors’ mobility favors knowledge spillovers, even when they work in another region and organization, due to the persistence of social relationships.

Besides these private benefits, a third mechanism suggested by the authors emphasizes the positive externalities that will concern organizations not involved in any cooperation relationship. In other words, for an unchanged level of R&D spending, “an increase in the average degree of openness in a sector may result in beneficial externalities which may raise the innovation productivity of the sector” [ROP 13]. We have pointed out that the knowledge created by universities and public laboratories could look like a public good and play a significant role in the development of technologies with multiple potential marketable applications driven by private organizations. The channels used to disseminate knowledge may take on the shape of social interactions, interpersonal networks or publications. We will emphasize that this effect may be strengthened in relation to regionally concentrated networks insofar as they enable face-to-face relationships between the members of the network. Consequently, the availability of a public or semi-public good (we once again come across the idea of common goods or common resources) may be especially advantageous to the creation of positive openness-related externalities by increasing the productivity of the firms’ innovation beyond the point that would be reached by privately investing in the production of knowledge.

This analysis has implications in terms of public policy, which may aim to spread open innovation practices. In particular, when the information produced by laboratories and universities has the properties of a public good, firms are encouraged to invest in spite of almost non-existent private returns for basic research developed inhouse.

Nonetheless, we can make two remarks:

• – Let us recall first that Arrow [ARR 84] was the first author to recognize the existence of shortcomings in the market in terms of new ideas and technological information. The highly unstable results of intellectual activities (basic research, the transition phase between invention and innovation) produce “an inextricable tangle of objective uncertainties and decisions of the entrepreneurs (that) is certainly uninsurable” [ARR 84]. Whatever the rewards imagined by entrepreneurs may be, this will result in a weakened incentive to produce knowledge. Besides these observations, we should also mention the idea that each innovation, by its very nature, requires specialized technical knowledge. Coupled with a market that is still in the making, each innovation (especially in the hi-tech field) can only be appreciated by an individual or people who have experience in this business segment. This means that the nature of technological information is not that of a public good that may be easily made available to users, subject to suitable contractual procedures. Technological information has an “amorphous quality” [ZEC 96], and this feature may be extrapolated to knowledge: it is hard to appreciate, it becomes valuable only as input integrated in a product and the producer’s knowledge is greater than that of future users. This justifies the development of inter-organizational collaborations involving producers and users working in tandem to assess, dissect the characteristics of and determine the scope of application of a piece of technological knowledge not “ready for use” (see Chapter 1 for innovative practices). In this context, replacing the theory of incentives with relational analysis as the foundation of the ecosystems needs more than wishful thinking in favor of open science, requiring us in particular to redefine the way we think about the nature of the public good of technological knowledge and information.
• – Secondly, the objectives assigned to public policy in this field may go against the process of institutionalization and legitimization of entrepreneurial science which, for example, has taken place in the USA in the last three decades. Several authors have emphasized that “the expansion of the commercial application of science may well lead to the de-institutionalization of open science” [COL 06, p. 346], compromising thus the interaction between the results of research and the development of proprietary technologies. Other research carried out on a significant sample of American companies indicates that the largest of them are relying less and less on their internal research. At the same time, they are getting more actively involved in the commercial application and the protection of existing knowledge. We cannot dismiss altogether the idea of a rejection of open science in favor of patents or secrecy.

However, the two most significant explanatory elements seem to be the following [ARO 15]:

• – There is a strong and negative relationship between the diversification of activities and the publication behavior. Companies that are refocusing their attention on their central skills reduce their investments in science and rely on start-ups regarded as more effective for the development of their inventions through capitalist and contractual relationships.
• – The globalization of markets, approximated in this study by the proportion of Chinese imports in overall American imports, provokes a reduction in long-term investments in the tangible and intangible capital of the companies studied. The globalization pressure, which derives from low-cost countries, directs firms towards more applied research and short-term incremental innovations.

In this context, it is interesting to think about ecosystems, insofar as we acknowledge that technological and market opportunities significantly depend on the dynamic interactions between public and private actors.

A recent report from MIT has noted that the main problem currently encountered concerns a lack of engagement from the private sector. This is seen in the absence of major R&D laboratories such as those developed within companies (Xerox Parc, Bell Labs) investing alongside the state in the 1960s [MIT 13].

2.3. Industrial structures

In terms of the type of industrial structures suited to these forms of organization, our aim is not to recommend public policies; these tend to promote “extreme structures”, oriented either toward the consolidation of big businesses – although these structures do have a decisive responsibility in terms of technological choices – or toward small, entrepreneurial companies, with skill sets falling into clearly defined technological niches. Through an analysis at the product–market level, Aghion et al. [AGH 05] showed that uniform industrial structures are not optimal for innovation. More recent projects have reached the same conclusions, highlighting the importance of another dimension: the organization of R&D work [AGR 12]. Results obtained for different regions in the USA (268 Metropolitan Statistical Areas or MSAs) indicate that the effects of the organization of R&D on innovation (measured in terms of the number of patented innovations, weighted using citations) increase as the diversity of the MSA in question increases – i.e. with the co-existence of both large and small laboratories within an ecosystem. The emergence of small structures is not independent of the technological trajectories taken by major operators. While big businesses have an undeniable advantage in terms of idea production (Schumpeterian model), these larger units do not only commercialize innovations in connection with their own established trajectories. The problem does not lie in a lack of competence at the level of big businesses but rather in the relationship between the organizational hierarchy and innovation. Managers are clearly, and understandably, unwilling to divert resources, in terms of both time and money, to support employees taking original and uncertain “outside the box” approaches. Pressure from the financial markets incites companies to cut costs (reducing the number of laboratories, limiting commitment to costly and high-risk initial phases) which do not immediately further the company’s existing commercial position.

There are two ways of limiting the effects of hierarchical relationships: inter-organizational collaboration (consortia, joint ventures, etc.) and spin-offs. The spin-off approach allows innovators to create small structures with the capacity to produce and commercialize innovations which would otherwise have been abandoned; this approach is thus one way in which the complementarity of structures can affect innovation. The spin-off approach is favored in cases where big businesses, following precise technological trajectories, produce potentially useful knowledge which lies outside the scope of their own research projects. However, it is not limited to the commercial sphere. In the USA, academic staff or scientists affiliated with universities or public laboratories may be encouraged to create their own companies to develop and commercialize innovative concepts. These structures are referred to as supported spin-offs [BLO 11]. Scientific entrepreneurship is funded by universities or laboratories, acting as venture capitalists, giving inventors access to the resources and talents they need in order to create a company.

It is important to note that the localization of innovation, at the company rather than regional level, is not an independent variable; it is strongly conditioned by the chosen technological options, the organizational structure of big businesses, and the types of funding which are considered to be most suitable. The greater the distance between an innovation and the technological “heart” of a company, the greater the risk that the company will lack the organizational capacity to develop the innovation. The organizational flexibility that permits a firm to engage with both mature and emerging markets can be dangerous, insofar as a new technology may create ruptures and reduce the potential of the existing skills acquired by the business [CHR 97]. A spin-off approach, supported by venture capital, provides a set of alternative solutions to a technological problem. One extreme example can be found in the case of Cisco. Due to the lack of visibility in technological evolutions and in the corresponding market, the company recruited and funded start-ups to explore technological solutions within the framework of a technology race; the “prize” was integration of the start-up into Cisco, which then took charge of the commercialization phase.

Note, however, the possibility of developing alternative forms of funding other than venture capital. Innovative investment projects, developed by groups of large and small companies with strong interconnections in terms of production and/or R&D programs may be subject to evaluation and certification by public agencies. This suggestion was made by Aglietta et al. [AGL 12] in terms of the implementation of clean technologies. These investments would be financed through indivisible funding bonds (reserved for the main operators in cases of individual action by the companies involved) or by injections of capital from specialized institutions.

2.4. Conclusion

Public action to encourage innovation and production ecosystems is based on a number of arguments, which will not be cited in full here. Note simply that, as in the case of geographic economics, benefits may be obtained by the grouping of companies stemming from the improved use of infrastructures, the reduction of transaction costs, and the existence of financial and technological externalities.

However, companies have long sought to profit from these gains and to internalize the externalities produced through the creation of natural clusters. In this context, public authorities must take a different approach if they wish to trigger the creation of a new cluster for example or to reinforce the current status of an existing cluster. The variety of these forms of organization, the need for precise identification of the specific mechanisms generating economic benefits and the need to correctly define appropriate actions, all pose problems in this context.

Public authorities can also influence the course of events by extending open innovation practices. In the cases where knowledge produced by universities and research centers is of a public or quasi-public nature, the availability of this information encourages companies to invest, in spite of the often-limited private returns offered by internal fundamental research. However, inter-organizational collaboration is necessary in order to derive full profit from the knowledge produced, with the aim of clearly defining the technological potential and the field of application of this knowledge. The relational aspect considered in Chapter 1 is particularly important here, re-centering reflection on the nature of knowledge as a collective asset.

Inter-organizational collaboration lies at the heart of questioning regarding industrial structures. Uniform industrial configurations are not optimal for the purposes of innovation. The co-existence of large and small units fosters the exploration of technological possibilities which may be uncertain and are often situated outside of the programs adopted by major firms. In this context, the creation of spin-offs from either large companies or universities is essential in maintaining the vitality of an ecosystem. This is illustrated, with certain nuances, by our analysis of American clusters.