203Third-Generation Innovation Policy: System Transformation or Reinforcing. . .

Innovation policy has its origins in research policy. The linear model of innovation originally developed and diffused following Vannevar Bush’s work (1945) for President Franklin D. Roosevelt in Science: the endless frontier, laid the foundations for modern policies related to innovation. Bush argued that public investments in R&D would spill over to industry and in turn result in the development of new technologies that would subsequently benefit consumers and result in economic growth. This linear view of innovation remained dominant for more than half a century and is still very influential for governments aiming to support innovation and economic growth. Today, however, most research would refer to such efforts as R&D policies rather than innovation policies.

This first generation of innovation policy gained widespread acceptance and was used for many decades. It was eventually questioned in the 1980s by Kline and Rosenberg (1986), who proposed a different framework which they referred to as “the chain linked model.” In this model, the innovation process was instead conceived of as non-linear, iterative, interactive, and hence more unpredictable than the linear model. One important implication of this model is the idea that the impulse for innovation may come from other places than the organizations in society that pursue basic science. There is in this sense less unilateral emphasis on universities, research institutes, and corporate R&D departments than the perceived change agents. Similar ideas were advanced by Nelson and Winter (1982), Freeman (1987), and Lundvall (1992) and paved the way for a stream of academic literature using the terms innovation systems or systems of innovation. Policies inspired by the innovation system approach are often thought of as the second generation of innovation policy. In contrast to the first generation, these policies were often designed to support the linkages between knowledge creation and commercialization, and more devoted to bringing actors together in novel network constellations. These policies also put increased emphasis on intermediaries and their role as facilitators for change and innovation by providing good conditions and support to networks involving both academia and industry. The triple helix approach can be considered a framework rooted in the same tradition (Etzkowitz & Leydesdorff, 2000).

The third generation of innovation policy is more aimed toward certain grand challenges and is in this sense more outcome oriented, whereas the preceding generations were more aimed at (1) generating an increased supply of R&D and (2) creating conditions for commercialization. Several scholars have paved the way for the emergence of this approach. Schot and Steinmueller (2018) used the term “transformative change,” Borrás and Edler (2014) wrote about “socio-technical systems,” and Geels (2004) introduced the notion of “system innovation.” One important characteristic of these approaches, as well as the policies drawing on them, which distinguishes them from the first and second generations of innovation policy, is that they pay more careful attention to demand and adoption of innovation in society. Consequently, the networks, or “systems,” that these scholars and policymakers define when analyzing and supporting innovation in society, become more complex by also including civil society and the consumer market. Thus, the outer boundaries of the system become less straightforward to define, and thereby so does the scope of actors shaping the target population of any policy initiative. One J.-E. Bergkvist et al.

204way to handle this increased complexity is to focus less on single actors, networks, or aggregates of actors, and more on the universal norms and regulations that the literature refers to as institutions (Scott, 1995). The underlying assumption is that the institutional framework of a society defines the behavior of actors. These institutions are defined at different layers, and it is crucial for policymakers to understand the dynamics between these layers.

Although the third generation of innovation policy, as described above, draws on an eclectic set of related ideas, there are some central ideas upon which this policy rests. Below follows a review of these central ideas and their implications for transformation.

3

System Transformation

The idea that institutions of various type interplay at different layers in society, shaping and challenging collective action, has received widespread attention in the literature underpinning the third generation of innovation policy. Frank Geels (2004) coined this idea and his proposal of how to handle it in empirical research: the multilevel framework. The framework may be applied to specific industries or the economy as a whole. According to Geels, innovations are usually nurtured in what the framework refers to as niches, i.e. parts of the economy that are sheltered from direct opposition or competition. These may be R&D departments in a priceinsensitive application such as within the military, within the public sector, or among universities. They may also be entrepreneurs subject to incubator programs or other constructions temporarily sheltering them from competition.

When technologies have been further developed within a niche, they subsequently enter various socio-technical regimes. A regime is an established and ordered part of the economy such as a specific industry. The regime is populated by other complementary and/or competing technologies; there are firms, customers, suppliers, and institutions maintaining power balance and order in the regime. Each actor in the regime posits capabilities and incentives making them more or less willing and able to accept a technology that comes from an alternative niche and tries