Understanding policy learning in regional innovation policies: lessons from the Galician case

1 Understanding policy learning in regional innovation policies: lessons from the Galician case Manuel González-López ICEDE Research Group – Departamento de Economía Aplicada Universidade de Santiago de Compostela [email protected] Funding Bodies: This research has received the support of the project EURIPER (EU Regional and Innovation Policies and Peripheral Regions) funded by the Jean Monnet Projects Action of the Erasmus+ Programme (Project Reference: 587410-EPP-1-2017-1-ES-EPPJMO- PROJECT). 2 Abstract The objective of this paper is to analyze how learning happens at regional innovation policies in a multilevel governance framework. We explore the case of Galicia (Spain), as an example of a region owning wide competencies on innovation policies and affected by two other government levels, the Spanish and the European ones. Our methodology has consisted of in-depth interviews with the persons who bore the major responsibility for innovation policies during the last 25 years. Our findings suggest that the interactions with other government levels act as a key channel for policy learning. Sometimes learning is associated to simple imitation of the national level, as it has been the case of the Galician innovation policy since its emergence. This might lead to questionable results because neither the previous policy-making trajectory nor the region particularities are taken into account. Other times learning is needed because a new political rationality is imposed by other government levels. This has been the case of the Smart Specialization strategy in Galicia, which nevertheless was also key to enhance social and political participation, a channel for policy learning not commonly used before. Key Words: Innovation, policy, regions, learning, Galicia 3 1. Introduction As regions have become a major actor at the implementation of innovation policies, regional innovation policies are increasingly receiving more attention from the literature. Most studies focus either on assessing existing policies or on proposing new measures to enhance regional innovation, nevertheless most of them have disregarded the policymaking process itself (Uyarra, 2009). Only the literature about policy learning ((B.-Å. Lundvall & Borrás, 1998); (Nauwelaers & Wintjes, 2008); (Borrás, 2011); etc.) has taken into account the policy-making perspective in order to assess how learning and change happen in innovation policies. Nevertheless more research is needed in order to understand how policy learning happens at the regional level, particularly in a multilevel governance framework. This paper aims at covering this gap by means of analyzing how has learning process occurred in the case of Galicia (Spain), as an example of a European region owning wide competencies on innovation policies but also affected by two other government levels (Spanish and EU). The paper is divided into four parts. In the first one, we introduce the literature about innovation policy and policy learning and later we propose an analytical taxonomy to approach policy learning in the regional context. In the second part, we describe the evolution of the Galician innovation policy during the last two decades, focusing on legal and institutional changes as well as on the instruments variations. The following section approaches the learning mechanisms, which have been operating in the Galician innovation policy during the reference period. We end with a conclusion including our main findings about how learning happens in regional innovation policies in a multilevel governance context. 4 2. Innovation policies and policy learning at the innovation literature a. The evolutionary perspective about innovation policies The evolutionary approach provides us with a robust theoretical rationality to understand how economies change (Nelson & Winter (1982), Dosi & Nelson (1994), J Stanley Metcalfe (1998), etc). This approach initially focuses on the importance of innovation in order to explain the dynamics of firms, industries and economies. As innovation is understood as the key factor for economic progress, the evolutionary advocates highlighted also the importance of innovation policies and thus developed an evolutionary rationality for such policies. In contrast with Neoclassical economics, the evolutionary approach not only focused on market failures in order to rationalize S&T policies under a linear view of innovation. They rather focused on the need to address systemic failures under the understanding that learning processes needed for innovation are the result of multiple interactions in systems where different agents and institutions are involved (J Stan Metcalfe (1995), J Stan Metcalfe & Georghiou, (1997), (Woolthuis, Lankhuizen, & Gilsing, 2005), etc). Such interactions, like for instance between universities an industries, are not always fluid and sometimes they do not even exist. Although there is a solid contribution of the Evolutionary school on the rationality and type of policies needed for innovation, contributions about the policy making process are more limited. Several authors have pointed out that policy-making should take into account the evolutionary character of innovation and economic change. Thus Metcalfe & Georghiou (1997) remarked that policy makers must be adaptive and must practice continuous learning, as the outcomes of their actions are subject to uncertainty. “Corresponding to our 5 view of market processes as adaptive constructs there follows a view of the policy maker as an adaptive agent” (…) “Operating in the context of complex innovation systems the policy maker must accept a considerable degree of indeterminacy and unpredictability in the consequences of policy initiatives. Complex systems are of necessity governed by ambiguity so there is a strong case for policy experimentation and policy learning” (J Stan Metcalfe & Georghiou, 1997, p. 7). Similarly, Van den Bergh & Kallis (2009) indicate that evolutionary thinking offers a distinct and useful perspective on public policy design and change. According to their views, that means taking into account bounded rationality of agents involved in the policy process, diversity of policy ideas and selection in a multidimensional environment and path-dependence of policies. The same authors indicate that “understanding policy evolution as a path-dependent process shifts the emphasis in analysis to earlier, historical events, which by design or accident may have locked-in institutional development to a certain path” (Van den Bergh & Kallis, 2009, p. 16). The authors also highlight that, because of political path dependence, the capacity to change public policy is much more limited than conventionally thought. Despite the potentiality of the Evolutionary school in order to understand innovation policy design and change, Uyarra (2009) has underlined that most studies from this school disregard the innovation policy process itself. This author criticizes that Evolutionary scholars are usually biased towards normative analysis (what policy makers ought to do) disregarding the positive analysis (i.e. what policy makers actually do). In similar terms than Van den Bergh and Kallis, Uyarra put the emphasis on the path-dependent nature of the policy making process: “(…) policies are adopted not on a tabula rasa, but in a context of pre-existing policy mixes and institutional frameworks which have been shaped through 6 successive policy changes. Past policy decisions would constrain the range of options available for current decision makers (…) Thus, when formulating policies, regions need not only take the knowledge and institutional base of the region as starting point, but also consider existing policy mixes and past policy history, for they will enable or constrain new policy goals.” (Uyarra, 2009, p. 132) In a recent paper Flanagan & Uyarra (2016) insist on the idea that most research about evolutionary innovation policies take a normative and rather theoretical viewpoint and very few are based on empirical studies. b. The Policy Learning approach The Evolutionary perspective discussed above identify policy learning processes as a major source for policy change and evolution ((J Stan Metcalfe & Georghiou, 1997); (Witt, 2003); (Van den Bergh & Kallis, 2009)). Policy learning has actually become a scholar subject of its own based on the ideas of the “learning economy” approach, developed by Bengt-Åke Lundvall (1992) and other authors. This approach characterizes modern economies as 'learning economies', in which knowledge is the crucial resource and learning is the most important process (Lundvall & Johnson, 1994). This approach has been particularly successful and widely applied within the fields of Economic Geography and Regional Development. The concept of ‘learning regions’ (Morgan, 2007); (Hudson, 1999); etc, attained a very relevant status within contemporary regional development discourses at the end of the 90s (MacKinnon, Cumbers & Chapman, 2002). B.-Å. Lundvall & Borrás (1998) established the basis of the understanding of innovation policies from the learning perspective. The authors refer to a new policy paradigm for innovation systems where the policy-making process is also learning and adapting 7 constantly to the new demands and conditions of the economy. Like some of the evolutionary authors named in the previous section, Lundval and Borrás consider that new practices of innovation policy evolve following an evolutionary logic rather than being created by deliberate design. Borrás (2011, p. 727) defines police learning as the “specific process in which knowledge is used in the concrete development of policy formulation and implementation”. As such, policy learning refers to the ways in which knowledge is created and used at all points of the policy making process, affecting rationales, design, operation and impacts of policy and policy mixes (Innovation Policy Platform, 2018). Policy learning is partly a process of institutional learning, which is defined as a dynamics process through which economic and socio-political institutions evolve gradually. In the seminal work mentioned above, B.-Å. Lundvall & Borrás (1998) identify four mechanisms through which policy learning can be encouraged: a) Policy evaluation for reformulation of innovation strategies. It is one of most common mechanisms for policy reformulation and it refers to the assessment of specific projects, programmes or whole policies. b) Technology assessment and forecasting. Unlike evaluation, technology assessment and forecasting are usually conducted externally and it has been largely used to fix priorities and strategies in the medium and long term. c) Enhancing the social and political participation in the definition of technological alternatives. It usually takes the form of representative committees where different stakeholders independent from government structures are represented. Apart from the traditional policy discussion in parliaments, there seems to be an increasing trend towards a more real social participation. 8 d) Learning from others. It refers according to the need to look carefully at policies designs and implemented in other countries (or regions). Further contributions have added to the understanding of learning processes by distinguishing between internal and external sources. As pointed by Borrás (2011) policy learning will partially depend of the level of capacities that organizations in the systems enjoy, combining knowledge sources from within and from outside that system. This primary distinction between internal and external learning sources has been extended by Nauwelaers & Wintjes (2008), who identify three different modes of policy learning in innovation, each of them might provide tacit and codified pieces of knowledge. a) Intra-organizational learning. It refers mainly to learning-by-doing, based on own practices, and occurs when policy makers and/or organizations learn from experiences. The relevant knowledge is mostly tacit, embedded in individuals or teams. However, part of it can be codified in the form of internal monitoring or evaluation reports. b) Intra-system learning. It refers to learning processes involving interactions with other agents of the innovation system. It includes both exchange of tacit knowledge (e.g. through meetings between people from the various organizations or employee mobility). Formal knowledge takes the form of evaluations, scoreboards or studies of innovation systems. c) Inter-System Learning. It involves comparisons and benchmarking between own system and other systems. It also uses tacit knowledge (e.g. through the interaction with other policy makers in meetings, discussions, etc…) and codified knowledge (taking the form of comparative scoreboards and international benchmarks). 9 c. A proposal for assessing policy learning at the regional level The contributions by B.-Å. Lundvall & Borrás (1998) and Nauwelaers & Wintjes (2008) constitute from our point of view two major analytical tools for assessing innovation policy learning channels. Nevertheless, as we will try to argue next, they must be updated and enriched, particularly if we aim at applying it to the regional context. First, we must take into account that nowadays innovation policies are affected by multilevel governance (OECD, 2010). That means that these policies are often implemented by different government levels and so, if we aim at analyzing innovation policies at the regional context, we must take into account the multi-level governance structure and the interaction between government levels (e.g. regional, national, EU). In this situation, coordination across levels of government becomes a key issue for achieving policy coherence. In the same token, it seems reasonable that policy learning at the regional level is affected by other levels of governance which might become (or not) channels to incorporate new knowledge. Secondly, from our viewpoint an additional learning channel or mechanism must be explicitly taken into account. We refer to new ideas and particularly those referring to theoretical approaches on the specific policy domain. Thus Sanz-Menéndez & Borrás (2001), in a study about the evolution of EU technology policy, argue that ideas and policy frames raised by expert communities have had a powerful influence in S&T EU policy reformulation. In a similar line, Mytelka & Smith (2002) point out that policy learning cannot be separated from the development of the field of innovation research itself. They analyze the links between the development of innovation theory since the late 70s and the 16 can be observed in the table, this period was characterised by a certain instrument “inflation” probably due to the budgetary bonanza of the time. The 4th (and last) R&D Plan (2011-2015) coincided with the middle of the economic crisis initiated in 2008. As pointed above, this is a period of important changes in the Galician innovation policy. In general, we observed two tendencies regarding policy instruments: a) there was a reduction of the number of instruments, towards more selective ones; and b) more focus was put on financial instruments, particularly interest rate subsidies for R&D and innovation projects. The explanation of this switch can be found in both learning experiences and ideological reasons (in 2009 the conservative party came into power again). On the one hand, the new government prioritized big innovation projects (so the instruments should be more selective) under the premise that during the previous period there were too many instruments which originated high bureaucratic expenditures. On the other, more emphasis was paid to financial instruments, avoiding not-refundable grants to which the government was quite reluctant, believing that it hided the real commitment of firms to innovation. It is also plausible that the budgetary reduction due to the economic crisis explained the change towards more selective instruments. Finally, the smart specialization strategy (RIS3), that is covering the period 2014-2020, included 19 instruments some of which were new while others were already present at previous plans. In this sense, there was a certain continuity between the RIS3 and previous plans. The main novelties refer to the programme of public procurement for innovation as well as some public-private strategic projects. Both instruments counted with a high portion of the budget dedicated to innovation policies. Following previous trends, more attention was paid to the business side of the innovation process. In any case, the most remarkable 17 feature regarding the instruments of the smart specialization strategy refers to a quite poor level of implementation, as only 7 out of 19 instruments were actually running during the first three years of implementation. Table 2. New instruments of the Galician Innovation Policy by period 4. Policy learning in Galician innovation policies In order to identify the learning mechanisms used by Galician innovation policy-makers we have followed the taxonomy presented in section two about policy learning mechanisms or channels. Table 3 shows how the mentioned mechanisms have affected policy learning (and so policy change) in the Galician case. The main inputs and insights for building the taxonomy have been the interviews conducted in our in-field research. As expected, the accumulation of internal experiences and learning by doing processes have been identified as major inputs for change and learning in Galician regional innovation policies. Apart from the first period when this policy actually started from zero, different type of learning processes shape the trajectory towards the consolidation of this policy. As we have seen in the two previous sections, both minor as well as major changes affecting the policy organization, design procedures or instruments, can be related to learning from internal experiences. An example of major policy change refers to the creation of the Galician Innovation Agency, which was largely due to the need for avoiding excessive bureaucratic costs and being more flexible and effective. Other change at least 18 partially due to experience learning refers to the move towards more selective instruments (lower number and with higher budgets) which also aimed at reducing bureaucratic costs and increasing their impact. Finally, the implementation of some instruments like the recent programs of public procurement for innovation has also been the result of a “learning by doing” internal process. As indicated by one of the persons interviewed “the adoption of these program has needed continuous interaction with internal financial and audit units because the general administration was not prepared to deal with such initiatives; in a sense we have been quite innovative in this regard”. Moreover, formal learning channels like evaluation and/or monitoring and technological assessment and forecasting have generally been little used in the Galician innovation policy. There is a very poor policy evaluation tradition in Galician administration (in similarity to many other European administrations). Evaluation and monitoring mechanisms have been formally included since the third R&D Plan but they have not been implemented so far (including the SMaSpe strategy). Regarding technology assessment and forecasting, some reports on R&D and technological trends were hired to private consultants or universities particularly to support the plans design. In some cases they were pointed as very useful by the policy makers but, as pointed out by one of the persons interviewed: “unfortunately in some cases these reports were stuffed in a drawer and not really used for further planning”. Learning through social and political participation for policy design and implementation has also been quite poor with the only exception of the last period. Most plans were only submitted to a process of formal socialization and political discussion, but not to systematized participative process. The most remarkable change happened during the 19 design of the SmaSpe strategy where the procedure followed the EU Commissions requirements and so working groups with stakeholders, policy-makers and experts were created, particularly in order to promote the entrepreneurial discovery process. Besides, there are plans for continuing such process of social participation during the implementation of the strategy. Moreover, one of the most used channel for policy learning comes from the interaction with the other two governance levels, the Spanish and European ones. As we have mentioned in previous sections, from the beginning the Galician innovation policy was very influenced by the Spanish policy and governance structure. Adaptation of organizational models and instruments has been a constant trend during all the period analyzed. Sometimes it is even difficult to distinguish between policy changes due to learning processes through the interaction with the Spanish government or simple imitation due to political or normative reasons. As we have previously indicated, the move from a General Secretary´s office, with a cross-sectional character (directly dependent from the Ministry of Presidency) towards a rather sectorial General Directory (dependent on Ministry of Industry) is an example of change due to imitation. In any case, even if policy change is due to simple imitation from other government level it also implies some kind of learning, notwithstanding the results of such learning process. At the end of the day, as pointed by Mytelka & Smith (2002) learning is usually a response to a problem that generates search with uncertain outcomes. In the words of one of the respondents: “This was a bad decision and time proved me right (…) Although it had not important effects on the short run it was the prelude of deeper changes happened later like the split of the innovation policy between industries and 20 universities, something that actually challenges the existence of an integrated innovation policy”. Likewise, the EU level has been also relevant as a source of learning and imitation for the Galician innovation policy, in particular during recent times. Again, the SMaSPe strategy is a key milestone as it was designed following the European Commission guidelines. As pointed out by the person responsible of the innovation policy at that time “the implementation of the strategy meant a great effort by officers and policy makers regarding of learning about the rationality and procedures of the SmaSpe”. Furthermore, as indicated in previous sections, the adoption of the SmaSpe implied that all the efforts of the innovation agency were put on its design and implementation. That meant the abandoning of the R&D plans that were already a consolidated trend in the Galician innovation policy, something with an impact that it should be assessed in the future. Learning from other regions and countries has also been pointed out as a relevant learning channel by the Galician policy-makers. Thus, most of the persons interviewed identified two regional models, the Basque and the Catalan, as mirrors to look at in different aspects and particularly when dealing with instruments. The Basque policy for technology centers and the Catalan model policies for human resources are two significant examples. These two regions have been the ones with accentuated industrial tradition but also with more ambition for deeper political autonomy in Spain. This has probably lead also to more ambition and experimentation in regard their innovation policies. Other models like the Nordic or the American ones were pointed out by some of the interviewed as references for specific aspects of the innovation policy. 21 Finally, regarding the theoretical rationales, according to the persons interviewed the Galician innovation policy has been largely influenced by some of the most remarkable theoretical approaches on regional innovation emerged in the last two decades. The 1st R&D Plan, pointing the emergence of a real innovation policy in Galicia, was an explicit attempt to incorporate a systemic view of innovation and leave aside the linear model. According to the person in charge of the policy at that time: “Before the 1st plan the innovation policy was not even a R&D policy but just a scientific research policy (…) in fact the policy depended from the DG of Universities, in the ministry of Education”. Moreover, the Regional Innovation Systems approach is well present for the design of the Galician innovation policy particularly since the 2nd Plan (2002-2005). As indicated in the previous section, several instruments aimed at enhancing the interaction of the different agents of the system (firms, universities, research centers, etc.). Other theoretical approaches have been in one way or another incorporated in the policy like for instance the open innovation theories (included in the rationality of the 4th R&D plan). Finally the Smart Specialization theoretical approach has indirectly been very influential for policy learning during the last period, with the implementation of the Galician RIS3. Table 3. Policy learning mechanisms 5. Conclusions 22 This paper has tackled the question of how innovation policy learning happens at the regional level. We have seen how innovation studies have poorly approached the policy learning and change mechanisms, probably because they have usually disregarded the policy-making process. Only the literature about policy learning has paid attention to this issue and so provides us with a rationality and analytical framework to understand policy learning and policy change. However, the insights provided by these contributions are not enough to understand innovation policy learning in the regional context. In particular, as we have tried to show in the Galician case, the interaction with other government levels constitutes one of the most used channels for policy learning. Sometimes learning is associated to simple imitation because of different reasons. In the Galician case, imitation of the Spanish innovation policy was a constant trend all the time, particularly during the initial stages, but also when political changes happened at the Spanish level. Other times learning through multi-level governance channels is needed because it is legally or politically mandatory to adapt to a new political rationality. The implementation of the Smart Specialization strategy constitutes also a clear (and maybe extreme) example of this kind of policy learning since its adoption was mandatory in order to gain access to structural funds. Moreover the need of adopting the Smart Specialization strategy has not only implied an important learning process in itself for the Galician policy-makers but it has also been key to enhance policy learning through other mechanisms. It is the case of learning through “social and political participation” since for the first time a systematized process of participation was used for designing and implementing the Galician innovation policy. Nevertheless, too much imitation from other government levels might not always lead to positive results because neither the previous policy-making trajectory nor the region particularities are taken into account. This seems to be the case of the Galician policy in the 23 90s, when the institutional set-up of the innovation policy was changed just to copy the Spanish one. 24 6. References: Borrás, S. (2011). Policy learning and organizational capacities in innovation policies. Science and Public Policy, 38(9), 725–734. https://doi.org/10.3152/030234211X13070021633323 Borrás, S., & Jordana, J. (2016). When regional innovation policies meet policy rationales and evidence: a plea for policy analysis. European Planning Studies, 24(12), 2133– 2153. Dosi, G., & Nelson, R. R. (1994). An introduction to evolutionary theories in economics. Journal of Evolutionary Economics, 4(3), 153–172. Flanagan, K., & Uyarra, E. (2016). Four dangers in innovation policy studies – and how to avoid them. Industry and Innovation, 23(2), 177–188. https://doi.org/10.1080/13662716.2016.1146126 Hudson, R. (1999). ‘The Learning Economy, the Learning Firm and the Learning Region’ A Sympathetic Critique of the Limits to Learning. European Urban and Regional Studies, 6(1), 59-72. Innovation Policy Platform. (2018). Policy Learning. Retrieved from https://www.innovationpolicyplatform.org/content/policy-learning Lundvall, B.-Å. (1992). National Systems of Innovation: Towards a Theory of Innovation and Interactive Learning. National systems of innovation Towards a theory of innovation and interactive learning. https://doi.org/10.1080/08109029308629360 Lundvall, B. Ä., & Johnson, B. (1994). The learning economy. Journal of industry studies, 1(2), 23-42. 25 Lundvall, B.-Å., & Borrás, S. (1998). The Globalising Learning Economy: Implications for Innovation Policy. Brussels. MacKinnon, D., Cumbers, A., & Chapman, K. (2002). Learning, innovation and regional development: a critical appraisal of recent debates. Progress in human geography, 26(3), 293-311. Mellizo-Soto, M. F., Menéndez, L. S., & Castro, L. C. (2002). Diseño institucional y preferencias políticas: O cómo equilibrar los intereses académicos en la política deficiencia, tecnología e innovación gallega. Documento de Trabajo, 2, 9. Metcalfe, J. S. (1995). Technology systems and technology policy in an evolutionary framework. Cambridge Journal of Economics, 19(1), 25–46. Metcalfe, J. S. (1998). Evolutionary economics and creative destruction (Vol. 1). Psychology Press. Metcalfe, J. S., & Georghiou, L. (1997). Equilibrium and evolutionary foundations of technology policy. Centre for Research on Innovation and Competition, University of Manchester. Morgan, K. (2007). The learning region: institutions, innovation and regional renewal. Regional studies, 41(S1), S147-S159. Mytelka, L. K., & Smith, K. (2002). Policy learning and innovation theory: an interactive and co-evolving process. Research Policy, 31(8–9), 1467–1479. Nauwelaers, C., & Wintjes, R. (2008). Innovation policy, innovation in policy: policy learning within and across systems and clusters. Innovation Policy in Europe. Measurement and Strategy, Edward Elgar, Cheltenham, 225–268.