94703 Innovation, Technology & Entrepreneurship Policy Note October 2014  •  Number 6 Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place ABSTRACT Innovation policy is increasingly conceived as the shared responsibility of national and subnational governments, yet most countries are struggling with multilevel governance. This paper provides an overview of how different countries are dealing with regional innovation policy and of the institutions and tools used for national-regional coordination. The focus is on discussing the policy implications for developing countries, which have decentralized their innovation policies more recently than developed countries, but face more acute challenges with respect to critical mass and institutional capacity. These challenges call for a sequential and cautious approach to decentralizing innovation policy, seeking economies of scale and avoiding wasteful duplication across regions, while introducing methods to ensure that devolution of power is accompanied by capacity building in regional governments. Governments of developing countries are also advised to be cautious when using a regional innovation policy approach as a tool to reduce cross-regional income gaps. This is because policy efforts to counterbalance the natural trend toward agglomeration of innovative activities in core regions might lead to inefficient allocation of public funds. 1. INTRODUCTION intensive capitalism.” Building on the “national During the last two decades, regions and cities around innovation systems” framework (Freeman 1987; the world have become more active in developing Lundvall 1992), the literature on “regional innovation their own innovation policy agendas. In the words of systems” underscores the importance of subnational Florida (1995: 528): “Despite continued predictions governments as catalysts and coordinators of regional of ‘the end of geography’, regions are becoming actors’ innovative activities and interactions (Cooke more important nodes of economic and technological 2001; Howells 1999). The regional innovation systems organization in this new age of global, knowledge- approach is inspired by neo-Marshalian theories of industrial districts and clusters, in which regional policy is deemed to be critical to foster “agglomeration Jose Guimon effects” and “interactive learning,” while at the same Associate Professor time to bring scientific knowledge closer to local Department of Development Economics Autonomous University of Madrid, Spain industrial needs (Koschatzky and Kroll 2009; Laranja et E-mail: jose.guimon@uam.es al. 2008; Rip 2002). Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place – October 2014 1 While a regional approach to innovation policy can institutional framework and defining national strategy, be positive in many respects, aggregate empirical all the while seeking coherence and economies of studies have failed to find a significant relationship scale and avoiding wasteful duplication. In addition between decentralization and a nation’s innovative to top-down devolution, powers in innovation policy performance (Taylor 2007). In addition, despite the are also shifting upwards toward supranational growing trend toward decentralization, it would be organizations, particularly in the European Union. The misleading to conclude that regional level policy is need for interregional cooperation adds another layer replacing policy at the national level. The two tiers of complexity when the appropriate demarcation should be conceived, rather, as interdependent of innovation policies does not coincide with the and complementary. The notion of “multilevel existing administrative boundaries. For example, governance” refers to sharing responsibility Shanghai’s innovation system encompasses parts over policy design and implementation between of the surrounding provinces to create a larger different administrative and territorial levels. Central functional urban area. In sum, a major challenge governments maintain their key role in building for policy-makers is to streamline the interaction of different levels of government in order to improve the efficiency of innovation policy. Studies have sought to understand how public TABLE OF CONTENTS policies in developing countries have contributed to the emergence of subnational “knowledge hubs” 1. Introduction . . . . . . . . . . . . . . . . . . . . 1 (e.g., Chaminade and Vang 2008); however, the Regional Innovation Policy and 2.  literature on multilevel governance of innovation Multilevel Governance: Trends and policy has mostly focused on developed country Challenges . . . . . . . . . . . . . . . . . . . . . 3 experiences. This is partly because developing 2.1  The Division of Responsibilities Between countries have only very recently decentralized their Levels of Government . . . . . . . . . . . . . . . . 4 innovation policies. This paper aims to address this gap by exploring the challenges associated with 2.2  Institutions and Tools for National-Regional regional innovation policy and multilevel governance Coordination . . . . . . . . . . . . . . . . . . . . . . . 7 in developing countries. Many of the opportunities 3.  Implications for Developing and challenges of regional innovation policy do not Countries . . . . . . . . . . . . . . . . . . . . . 10 depend on a country’s level of development, but 3.1  Coupling Regional Innovation developing countries’ regional innovation systems Strategies with the Broader National often face higher levels of diversity, income inequality, Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . 10 and institutional instability than do similar systems in developed countries (Cassiolato and Martins 3.2  Regional Capacity Building . . . . . . . . . . . . 11 2000). The following section provides an overview of 3.3  Using Regional Innovation Policy how regional innovation policy unfolds in different to Close Income Gaps and countries and of the institutions and tools used for Address Societal Challenges . . . . . . . . . . . 12 national-regional coordination. Section 3 discusses 3.4  Monitoring, evaluation and the implications of regional innovation policy for experimentation . . . . . . . . . . . . . . . . . . . . 13 developing countries and Section 4 presents some concluding remarks. 4. Concluding remarks . . . . . . . . . . . . . 13 References . . . . . . . . . . . . . . . . . . . . . . 14 2 Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place – October 2014 REGIONAL INNOVATION POLICY 2.  Also recently, in Latin America, larger countries like AND MULTILEVEL GOVERNANCE: Brazil, Mexico, Colombia, and Chile have begun to move TRENDS AND CHALLENGES toward a regional approach to innovation policy. For Traditionally, regions have only been involved in example, in Brazil, the decentralization of innovation innovation policy in some federal countries, such as Germany, Switzerland and the United States. But regions in more centralized countries, such as THE DECENTRALIZATION OF INNOVATION BOX 1.  France, Sweden, Poland, Japan and South Korea, POLICY IN CHINA among others, have progressively gained new policy China recognized the need to decentralize innovation policy during their competences in this area. However, it should be sweeping economic reforms in the late 1970s and have substantially noted that the scale and scope of decentralization accelerated innovation policy efforts during the last decade. Currently, subnational governments contribute more than half of the country’s differ markedly from country to country (and even total public expenditure in R&D and innovation, but the division of from region to region within a country) depending policy responsibilities between the subnational and national levels is on their particular technological profiles, institutional not clearly defined by law. In practice, the central government controls basic research and areas of national interest (e.g., national defense, frameworks, and historical trajectories. For example, in health, security). The subnational governments oversee technological 2009, the subnational share of total R&D expenditure development and diffusion of existing technologies. The central govern- ment allocates 70 percent of its science and technology budget to R&D, by the public sector ranged from around 5 percent in while subnational governments allocate just around 30 percent and use Austria and Denmark to around 20 percent in Spain the bulk of their budget to other types of innovation support. and South Korea, 50 percent in Germany and China, Subnational governments primarily influence innovation policy by devel- and almost 80 percent in Belgium (OECD 2011). Note, oping special economic zones, scientific parks and industrial clusters. China’s establishment of special economic zones in the early 1980s however, that budget allocation (fiscal decentralization) provided incentives for export-oriented industrialization and tested the does not necessarily coincide with decision-making market economy and new institutions, while facilitating the absorption of modern technologies and management practices. More recently, in an power (political decentralization), because budgets effort to further develop technology-based clusters, Chinese regions and can be decentralized, while decisions are taken at the cities have launched science parks, incubators, and extension services, central level. In other words, the role of regions can which has encouraged local universities to engage in research and to es- tablish industrial linkages, induced venture capitalists to invest in SMEs, be a passive one, limited to the implementation of and attracted R&D from multinational firms. The national government nationally defined policy initiatives, or an active one, has sometimes intervened in cooperation with regional governments by selecting the most promising special economic zones and science parks with regions acting as participatory partners in national around the country and providing them with additional funding and policy processes and co-financing policy initiatives, or support. For example, since 1988, the national government has selected even serving as independent policymakers that develop and certified 84 national science parks through the Touch Programme. Regional governments continue to offer a range of complementary their own strategies and devote their own resources to policies for science parks, including free land allocation, infrastructure projects that they select. and facility support, as well as enhanced public services to companies located in the parks. In order to be admitted into the science parks, firms must be certified by regional regulators as “high-tech enterprises.” Many middle-income nations, especially the larger The process of decentralizing innovation policy in China is facing signifi- emerging countries, have also decentralized their cant difficulties. For example, in addition to the science parks designated innovation policies in recent years. For example, at national level, the proliferation of provincial and local initiatives has given rise to a “race-to-the-bottom” characterized by increasing in China, the central government has focused on competition based on tax incentives and subsidies. Li (2003) argues that fundamental research, while the provinces and an unclear division of labor among the different levels of government has sometimes resulted in inconsistencies between regulations adopted at municipalities have often concentrated on diffusion of different levels as well as in a less coherent implementation. According technology and supporting industrial clusters (Box 1). to a report by the World Bank and DRCSC (2013: 168): “The challenge for Similarly, although regions in Russia have very limited China is to arrive at a national innovation strategy that is cost efficient, optimally decentralized, rationally sequenced, and urban-centric.” policy competences and resources available for R&D, their relevance in innovation related support has Sources: Li 2013; Zeng 2011; OECD 2009a; World Bank and the Development increased substantially, with a focus on technology Research Center of the State Council 2013. parks, incubators and venture funds (Spiesberger 2013). Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place – October 2014 3 policy began in the mid-1980s with the establishment research organizations and universities, competitive by the National Research Council of twelve science parks research funding, tax exemptions, investment in in cooperation with state and municipal governments venture capital, etc.), and “soft instruments” (e.g., and universities. Later on, with the development of voluntary standardization, codes of conduct, public- networks of business incubators, innovation policy private partnerships, etc.). Sometimes the division of shifted further from a centralized approach to more power between national and regional governments is bottom-up, regional and local initiatives in cooperation clear; in other instances, those powers are shared. For with universities (Etzkowitz et al. 2005). example, in most countries, the national government wields exclusive power over the intellectual property In lower-income countries, however, the decentralization rights regime, while both the national and regional of innovation policy remains at its infancy, not least levels share responsibility over developing scientific because science and innovation policy is underdeveloped infrastructure. A specific type of policy instrument, such due to very limited available funds. Central governments as matching grants, can be used simultaneously at both across Africa dominate science and innovation policies levels, but must have different targets. For example, even in the most advanced countries, like South the national level could target large-scale projects that Africa. However, South African provincial governments involve the production of new knowledge, and the have developed the authority to implement regional regional government could target smaller programs development initiatives that have begun to target that focus on technology diffusion. Building on a innovation, like the Blue IQ Initiative (Gauteng Province) review of the literature, Table 1 provides a tentative and the Cape IT Initiative (Western Cape). In Morocco, framework to assess a country’s multilevel division despite having a predominately centralized science of innovation policy competences. Note, however, and innovation policy, the national government is that in light of the large differences across countries, increasingly adopting a decentralized perspective on these classifications do not intend to be prescriptive or innovation policy, including new initiatives to develop indicative of the most efficient approach. technology clusters and science parks in selected regions. For example, the Innovation Cities program, Drawing clear lines of separation between national launched in 2011 to boost innovation in four cities and regional innovation policy competencies, however, (Casablanca, Fez, Marrakech and Rabat), emphasizes the would be counterproductive since certain policy building of new research centers and business incubators instruments are best executed if they are shared in the vicinity of universities in order to foster university- among different levels of government. This marks industry cooperation (Basrhir 2011). These examples a more cooperative approach that can benefit from suggest that although decision-making still remains complementary strengths around a single public highly centralized, cities and regions are also gaining policy issue. A typical example of a national policy ground as the loci of innovation policy in Africa. that encompasses the regional level is the centrally coordinated cluster initiatives that provide support to selected regions. These initiatives result from a The Division of Responsibilities 2.1   bottom-up program design driven by a competitive Between Levels of Government call for proposals open to all regions (Eickelpasch and The levels of differentiation within innovation policy Fritsch 2005). In France, for example, the Poles of instruments can vary largely from country to country. Competitiveness program (“Pôles de Compétitivité”) Borras and Edquist (2013) classify the different was launched in 2005 to support research-intensive innovation policy instruments as “regulations” (e.g., clusters throughout the country. Within this program, intellectual property rights, statutes of universities, clusters deemed to be internationally significant receive public research organizations, bioethical regulations, additional funding from the national level, while those etc.), “economic transfers” (e.g., en block support to that are not considered a national priority are left in the 4 Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place – October 2014 TABLE 1  National vs. Regional Responsibilities in Innovation Policy MORE OFTEN NATIONAL MORE OFTEN REGIONAL Mode of innovation Knowledge generation Knowledge diffusion and exploitation Target groups Public research labs, universities, large firms Small firms, startups, spinoffs Infrastructure Universities, public R&D labs Incubators, science parks, special economic zones, technology transfer offices Regulations Intellectual property rights Building permits, infrastructure development Economic transfers Tax deductions, large grants to new R&D projects Smaller grants to fund business innovation Soft instruments Standardization, codes of conduct Networking and brokerage services Human capital Higher level education, postgraduate scholarships Technical training, lifelong learning, internships Linkages International linkages, scientific collaboration Public-private partnerships, cluster development Sources: Chaminade and Vang 2008; Karo 2012; Koschatzky and Kroll 2009; Laranja et al. 2008; OECD 2011; Perry and May 2007; Ritzen and Soete 2011; Todtling and Trippl 2005; World Bank 2010. hands of regional governments. In the 1990s, Germany, continuous information sharing. For instance, national- through the BioRegio program, sought to stimulate regional coordination should seek to streamline and its biotechnology sector by promoting clusters in harmonize administrative templates and procedures selected regions. The central government shepherded for requesting R&D incentives, in order to reduce the program’s creation, finance, administration, and transaction costs for target firms and researchers. selection of target regions. However, the policy was With the information they compile about regional essentially regionalized as each region designed and policy instruments, the central government can create implemented its specific plan. information platforms, or “one-stop shops,” for grant applicants. These might include, for example, websites In practice, a survey of OECD countries revealed the listing national and regional R&D grants, regional cluster lack of a clear division of powers between national and development initiatives, science parks, etc. regional innovation policies and a large overlap in the kinds of instruments used (OECD 2011). Overlapping In general, the appropriate level of intervention for a national–regional policy instruments creates synergies given innovation policy instrument depends on several only if both levels complement one another. This different factors, including the type of targeted actors, complementariness can arise in the way the instruments the efficiency of the innovation delivery mechanisms, are structured in terms of their target actors or across and the strategic scale of the instrument, among others. the range of service needs. Overlapping instruments Moreover, different regions have different administrative leads to negative results, however, only if they create capacities to implement innovation policies. Even if redundancies by failing to distinguish among target regions have sufficient budgets and competencies, groups or topics. For example, in a study of Korea’s some can still lack public employees with the necessary regional innovation policy landscape, Hassink (2001) skills and experience. Conversely, central governments noted that a lack of coordination between national tend to operate closer to the knowledge frontier due to and regional governments resulted in fragmentation their size, international networks, and capacity to attract and overlapping functions. The proliferation of public qualified workers. However, returns to scale could also support programs at different levels can lead to higher decrease, and, especially in large countries, the center transaction costs, more bureaucracy, and confusion could get overwhelmed. The key, then, is to match among target firms. To address these risks, national and allocation of authority with administrative capacities. regional governments can develop mutually acceptable eligibility and reporting criteria, clear mechanisms for Central governments should ensure the coherence of co-financing policy instruments, and clear systems of regional innovation strategies by seeking economies Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place – October 2014 5 of scale and reducing fragmentation. In the absence BOX 2 DECENTRALIZATION OF INNOVATION POLICY of an appropriate central oversight, decentralization IN THE RUSSIAN FEDERATION can result in inefficiencies and governance problems. Ritzen and Soete (2011) argue that if all regions Russia is a large and highly heterogeneous country with a wide variety of income levels, industrial specializations, and innovation capacities strive to achieve research excellence, the majority across its 83 regions (or “federal subjects”). Russia is, therefore, an will miss the mark, often leading to the creation appropriate prototype for differentiated regional innovation strate- gies. Traditionally, regions were not involved in R&D and innovation and subsequent decline of overambitious research policy, but in recent years some have developed their own innovation infrastructure in locations with weak endowments, strategies focused on promoting business innovation through regional also known as the “cathedrals in the desert” venture funds, technology parks and incubators. At the strategic level, federal government efforts to support this trend and further syndrome. In addition, the central government should empower regional governments appear explicitly in the 2011 “Strategy be attentive to possible gaps in combined regional for Innovation Development” and in the innovation policy chapter of its 2020 strategy (Gokhberg and Roud 2012). At the operational strategies that could result in neglected policy areas level, a relevant example is the 2012 cluster initiative, in which the at any level of government. In particular, investments federal government provides matching funds to the most promising with benefits that expand beyond the regional and technologically advanced regional clusters in order to resolve infrastructure bottlenecks. In parallel, the R&D funding allocation constituencies will tend to be underfunded without system continues to shift from block funding based on soviet-era cen- central government intervention. For example, if no tral planning decisions to a competitive, project-based R&D funding system (Spiesberger 2013). region sets out to specialize in medical technology, These developments have begun to lead to a more competitive al- the central government should consider necessary location of science and innovation funds across regions, which has the creation of at least one public medical technology resulted in a tendency toward the most advanced regions absorbing research center and launch a national competition to a larger share of federal funding. While the government is aware of the advantages of concentrating resources in “regional engines” of fund it.1 national innovation performance for the sake of overall efficiency, it is also concerned with mounting regional inequalities (OECD 2012). The federal government has also reaffirmed its primary objective of National governments often struggle to create promoting balanced regional development as stated in its current innovation policy that promotes balanced technological “Strategy of Diminishing Disparities Between Regions Until 2015.” development across economic levels and geographic In the same vein, the Council for Research for Productive Forces has developed four new programs targeting innovation in less developed regions. The positive link between technological regions. innovation and economic growth implies that in order to close income gaps across regions, policies Sources: Gokhberg and Roud, 2012; OECD, 2012; Spiesberger, 2013. should seek to close innovation gaps. The problem is that most national innovative efforts tend to be heavily concentrated in a few core regions. And, this concentration is nearly insurmountable due to in Box 2. Also in China, the central government is the cumulative nature of technological capabilities, attempting to revert mounting income inequalities economies of scale, agglomeration effects, and across regions by providing more support to indivisibilities. This self-reinforcing process often innovation in underdeveloped regions. For example, aggravates regional divergence, which comes with the Revitalization Plan for Higher Education Institutes important economic and political implications. In in Mid- and Western China (2012–2020) funds R&D sum, central governments charged with allocating projects in order to strengthen universities in less innovation funding must balance promoting developed regions. scientific excellence through regional competition and concentrating resources in the core regions, on the one hand, and encouraging convergence and equality among their regions by nurturing lagging 1 The so-called ‘subsidiarity principle’ refers to the intervening role of the next higher level of government in the case that the regions, on the other (Crespy et al. 2007). The Russian lower level in the federal system does not intervene to address Federation struggles with this dilemma as discussed a societal need (Bermann, 1994). 6 Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place – October 2014 Institutions and Tools for National- 2.2   Regions and localities are also increasingly creating their Regional Coordination own autonomous innovation agencies to implement The discussion in the previous section emphasized regional innovation strategies that act as counterparts the importance of national-regional coordination. to the national agencies and ministries responsible for Such coordination should be fostered at all stages of innovation policy. Some regions have established new, the policy cycle, from strategic formulation to policy autonomous agencies, while others rely on the evolution design, implementation, monitoring and evaluation. of preexisting structures, like regional development In this section we further discuss some of the main agencies. An example of the former is Ruta N, a local coordination tools and provide country examples to agency established in 2009 in Colombia’s second largest illustrate the policy options available. city, Medellin, in order to execute the city’s new Strategic Plan for Science, Technology and Innovation 2011–2021, which seeks to boost entrepreneurship, business Consultation Bodies, Territorial Representatives, and innovation, and cluster development. An example of Agencies the latter is the Agency for Economic Development of The decentralization of innovation policy is often the Greater ABC Region in Brazil, which was created in accompanied by the creation of new national-regional 1998 with a broader regional development mandate, institutions for dialogue and agenda-setting that are but has adopted a stronger focus on promoting adapted to specific circumstances within each country innovation in recent years (Zhang 2010). In particular, and that build on existing institutional frameworks. the agency has established two business incubators These may take the form of policy councils made up of and several programs to promote innovation and to high-level representatives from the national and regional upgrade the regions’ automobile parts, computer governments or territorial representatives from the hardware, and plastics clusters. The agency’s board of relevant ministries and national agencies. For example, directors is comprised of representatives of the regional in the Czech Republic, the national innovation agency governments, the seven regional municipalities, business CzechInvest has established regional offices that support associations and firms, higher education institutions, innovation policy implementation. These agencies have and trade unions. This helps improve not only vertical played an important role over the past few years in coordination, but also public-private partnerships supporting the development of science parks and cluster and horizontal coordination among the different initiatives across the country’s regions. In Mexico, formal stakeholders of the regional innovation system. national-regional consultation occurs via the National Conference of Science and Technology (S&T), which also A more recent trend is the creation of multilevel promotes regular dialogue between national and state agencies for the joint design and implementation of level S&T councils. Moreover, the National Council for innovation strategies. A case in point is Innovation S&T (CONACYT) has established regional delegations Norway, a new national agency launched in 2010 that liaise with subnational governments and oversee and jointly owned by the regional (41 percent) and the implementation of federal programs (OECD 2009). the central (51 percent) governments. The new agency has increased regional involvement in the In addition to these kinds of formal institutional design and funding of innovation policies, while processes, most countries also use national-regional providing an opportunity to pool funds and exploit working groups and ad hoc meetings as instruments complementarities and synergies across regions. of dialogue and coordination. Moreover, a frequent practice is to invite regional representatives to the working groups and national innovation strategy Project Co-Financing and National-Regional Funds development consultation processes, which provides a In some instances, co-funding mechanisms are associated forum for building consensus and strategic alignment. with individual, large-scale projects, such as scientific Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place – October 2014 7 infrastructure; whereas, in other cases, they have a selection process was that it fostered information more open and longer-term scope. There are many sharing among regions to avoid overlaps. possible design options, but the general idea is to share responsibility over the selection and funding of science In other instances, governments have promoted and innovation programs in order to create synergies the creation of joint national-regional trusts to fund and strengthen policy coherence. A possible mechanism innovation projects for regional actors on a competitive for this is to establish a matching grants system in which basis. For example, Mexico’s “mixed funds” (Fondos the central government agrees to match every dollar, up Mixtos, FOMIX) were jointly established in 2002 by the to a certain limit, that regional governments dedicate to central and regional governments as a trust fund for innovation projects on the condition that private firms science and innovation projects. As of 2013, more than also match the contribution (World Bank 2010). 35 mixed funds had been established across the Mexican States and also in some municipalities (Ciudad Juarez, La Project co-financing is a widely used instrument Paz, Puebla). This system was meant to foster innovation to develop strategic investment projects and capacity at regional level and to better articulate federal infrastructure whose benefits span beyond the regional and regional support for innovation. Each fund has its constituencies. For example, a public research center in own technical committee and evaluation commission to a region can collaborate with firms from other regions issue calls for proposals and select projects for funding in the country. Similarly, a university from one region within their respective regions. The total budget for can accept students from another region on a reciprocal the 2001–2012 period was around $US 580 million, of basis and their graduates can end up working in which 45 percent came from the regional governments different regions. As referenced earlier, such initiatives and 55 percent from the central government. However, tend to be underfunded without central government according to a review of this program by the OECD, in investment. Under these circumstances, project co- many cases, the system suffered from a lack of well- financing addresses fiscal imbalances and encourages articulated demand from the states, especially from the convergence of objectives. The decision of the central less developed ones (OECD 2009b). government to co-finance projects is often the result of a national competitive solicitation open to all regions. Other governments have created new funds to finance the implementation of regional innovation strategies, In China, for example, the National Science and which often requires regions to develop their own Technology Infrastructure Program provides a roadmap innovation strategies. The regional government, in for investing in research infrastructure by fostering cooperation with a central government evaluation inter-regional integration and sharing infrastructure. commission, selects and finances regional projects that During the 11th five-year-plan period (2006–2010), the comply with national goals. The Colombia’s Fund for government invested around $US 275 million out of Science, Technology and Development, for example, which 20 percent came from subnational governments has financed regional innovation initiatives since 2012 (Li 2013). Likewise, in Spain, a national program has through mineral resource royalties (Box 3). selected a limited set of so-called “singular technical- scientific installations” in regions that are co-financed with national funds. These are comprised of large National-regional contracts investments to build unique public research labs Many countries rely on contracts or bilateral and equip them to engage in cutting-edge scientific agreements between central and regional research. In 2007, the Spanish regional and national governments with regard to their mutual governments agreed to create 24 new “singular commitments, the assignment of decision-making technical-scientific installations” thru 2015, in addition powers, the distribution of financial contributions, to the 37 existing ones. A major advantage of the and mechanisms to monitor and enforce contracts 8 Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place – October 2014 BOX 3 COLOMBIA’S NEW ROYALTIES FUND FOR STI Although some of Colombia’s 32 regions (or “Departamentos”) were already engaged in the design and implementation of science, technology and innovation (STI) policy, the scale and scope of regional involvement in this area has increased substantially since 2012, following the creation of a new fund for STI. New 2011 legislation governing state royalty administration from national mineral resource exploitation (including oil) contains a provision to allocate 10 percent of the royalties to the new fund for STI. Currently, this represents around $US 500 million annually, which is about the same amount that the country invested in R&D prior to the establishment of the new system. The royalties fund for STI is distributed to regions based on a formula that considers popula- tion, unemployment and poverty indicators, so that poorer regions get more funding per head than richer ones. The Fund is administered by Colciencias, the R&D funding agency of the central government, but the regional governments select the projects they would like to fund in their regions. A committee established by Colciencias approves selected projects after ensuring that they conform with the rules, expected quality, and feasibility. Ultimately, regional governments need to develop their own innovation strategies to better select the most likely projects to be awarded monies from the royalties fund. This new system has strengthened regional autonomy in innovation policy and has promoted a more systematic design of regional innovation strate- gies. However, early experience also points to a number of challenges. The first problem is that even though a country’s less advanced regions receive a larger part of the funds, they lack the absorptive capacity required to design and develop the right kind of projects to promote innovation. In order to address this issue, Colciencias has established an Office of Regionalization that has supported so far the drafting of 27 regional development plans for STI, which serve as a reference for project formulation. A more profound challenge of this new fund is its charge to balance two conflicting strategies: the first is oriented around scientific excellence and performance, and the second aims to reduce regional disparities. Another problem is the lack of the- matic coordination across regions and between regions and the state in which each region pursues its own strategies in an isolated manner. For these reasons, Colciencias is currently trying to stimulate collaboration among regions as a way to better align regional strategies and to help underdeveloped regions to catch-up by learning from more advanced regions. However, up until now, the results have been disappointing because the central govern- ment cannot practically impose interregional collaboration, and regions have a tendency to allocate all of the available funds only within their own boundaries (i.e., to local firms and universities without involving partners from other regions). The risk is that this may lead to fragmentation, wasteful duplication, and other inefficiencies. Sources: Cuervo and López 2013; OECD 2014. (OECD 2010). Contracts are used for co-financing improve national-regional coordination in support of and joint programming innovation policy instruments. R&D and innovation in areas of common interest. In this Furthermore, if properly designed, they can hold levels way, the Spanish government provides preferential loans of government accountable to act beyond political at low interest rates for regional investments in science mandates, which contributes to the continuity and and innovation until 2015, with a 10-year pay back stability of innovation policy strategies regardless of period. Within the contract, each region commits to its the political cycle. National-regional financing contracts own quantitative objectives for meeting the national take many forms. Sometimes they finance specific, plan’s 2015 targets. A Joint Monitoring Commission large-scale scientific infrastructure whose benefits with representatives at both levels monitors progress. In spill over regional borders. In other instances, they France, relationships between the regional authorities coordinate, manage and finance broader innovation and the central government are organized by seven-year policy strategies. The OECD has highlighted the contracts called State-Region Project Contracts (Contrat “relational” nature of contracts, which means that de Projet État-région). These broad contracts cover all they can play an important trust-building, information- policy areas and set out the financial transfers provided sharing, and capacity-building role that facilitates by the central government to meet regional policy multilevel governance (OECD 2010). Unlike a general objectives, with a specific chapter dedicated to research call for proposals, relational contracting is more and innovation. These contracts also regulate the interactive and serves as a vehicle for managing a development of the Poles of Competitiveness program cooperative relationship over time. mentioned earlier. Through these kinds of programs, the French government is shifting toward a network- For example, since 2009, the Spanish central oriented and bottom-up approach to innovation policy government has established contracts (“convenios”) to (Barmeyer and Kruth 2012). Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place – October 2014 9 IMPLICATIONS FOR DEVELOPING 3.  BOX 4. SMART SPECIALIZATION AND NATIONAL- COUNTRIES REGIONAL COORDINATION IN POLAND As argued above, many developing countries Poland is a unitary state divided into 16 regions (“voivodships”), each are struggling with multilevel innovation policy with substantial regional autonomy. In order to receive funds for innova- governance, even more so than developed countries, tion during the current 2014–2020 European Union (EU) financial per- since they are often confronted with higher levels of spective, Poland had to design new Research and Innovation Strategies for Smart Specialization at both the regional and national levels in com- regional diversity, income inequality, and institutional pliance with the new European Commission framework. A World Bank shortcomings. These challenges call for a sequential team assisted the Polish government in this process, and their report provided a useful illustration of the challenges associated with national- and cautious approach to decentralizing innovation regional coordination of innovation policy (Piatkowski et al. 2014). In policy, introducing methods to ensure that devolution general, there was considerable overlap and lack of coordination across of competences is accompanied by capacity building in regions, and between the regional and the national strategies. This is at least partially because Poland developed many of the regional innova- regional governments. It also requires the development tion strategies before finalizing the national innovation strategy. The of new institutions and national-regional coordination World Bank report argues that “system consistency requires some gen- eral structure and sequencing, allowing lower documents enough time tools, such as those discussed in Section 2.2. The rest and space to adjust to hierarchically senior documents” (Piatkowski of this paper further discusses some implications for et al. 2014: p. 35). In addition, the report found that in order to avoid regional innovation policy in developing countries. duplication and to maximize synergies, national level leadership should be strengthened by adopting a more active coordinating role, providing enhanced procedural support to regional governments, and mapping out potential industrial specializations throughout the country. Coupling Regional Innovation Strategies 3.1   In many cases, regions were unable to design strategies that differenti- with the Broader National Strategy ated themselves from other regions. Their strategies tended to be very generic and lacked a clear link to the region’s specific characteristics The first step for regions that embrace an active role in and endogenous potential. According to Klincewicz (2014: p. 22), “the innovation policy is to develop a multi-annual strategy regions selected similar specializations, related to the most popular, that determines the main problems and defines a set broadly defined technology areas, without major intra-regional differen- tiation.” In some instances, this was due to an attitude of “mechanical of priorities in the short-, medium-, and long-term. The fulfillment” of smart specialization strategies. Policy-makers felt that strategy should reflect the broader national context they had adequately complied with EU funding requirements and fell short of constructing tangible regional socio-economic transforma- and avoid unnecessary duplication and fragmentation tion. Moreover, innovation strategies have been aligned with capacities of efforts, as well as seek opportunities for joint in implementing agencies rather than with firm needs, leading to an programming with the central government and excessive focus on public research institution agendas rather than on market demands (Kapil et al. 2014). neighboring regions. Regional and national innovation To improve national-regional coordination, Poland relied on territorial strategies should be coherent in terms of their contracts signed between the central government and the individual objectives, target indicators, and policy instruments, regions in 2014. The aim was to define the region’s tasks and objectives as well as the mechanisms for disbursement of EU funds allocated but this is often hard to achieve (see Poland’s recent to the region for 2014–2020. Importantly, these contracts strive to experience in Box 4). Forming coherent policy means delineate a clear line of demarcation between the responsibilities of that regional governments link their strategies with the central and regional governments, and thus set the framework for regional intervention. those at the national level. The central government should also design its national strategy after jointly Sources: Kapil et al. 2014; Klincewicz 2014; Piatkowski et al. 2014. considering the strategies of the regions. A key element of regional innovation strategies is to 2 In 2011, the European Commission created the Smart decide in which sectors or technologies to specialize. Specialization Platform “to provide information, methodolo- For this purpose, the notion of “smart specialization” gies, expertise and advice to national and regional policy makers, as well as to promote mutual learning, transnational is becoming increasingly influential in Europe and cooperation and contribute to academic debates around the other regions around the world.2 In essence, this concept of smart specialization.” This was followed by the approach advocates for concentrating resources in 2012 publication of the “Guide to Research and Innovation Strategies for Smart Specialization (RIS3).” Both are also useful a limited set of priority areas that are clearly aligned tools for countries outside the EU and are available at: http:// with regional strengths and competitive advantages. s3platform.jrc.ec.europa.eu/ 10 Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place – October 2014 Beyond industrial targeting, the priority areas can interactions. However, in practice, policy instruments be “activities” (e.g., advanced manufacturing, clean are often selected based on the continuation of energy, e-health, research labs, design, or headquarter previous schemes or pressure from specific lobbies and functions) or “key enabling technologies” (e.g., ICT, interest groups, rather than on visionary considerations nanotechnologies, biotechnology, etc.), which cut of policy mix interactions and critical assessments of across several industries. As regions try to determine the most acute problems (Borras and Edquist 2013). pathways for smart specialization, it is critical that When selecting the innovation policy mix at regional they connect with the national vision and that they level, it is first necessary to assess: (1) what is being consider the specialization choices of other regions done at the national level, (2) what else is needed, in the country, so as to build synergies and avoid (3) how is it possible to tailor national programs to duplications. However, a frequent scenario in many the regional context, and (4) what opportunities lagging regions is the difficulty in finding clear would complement national programs with regional strengths or market trends in which to determine their initiatives? Indeed, regional innovation policy does not most appropriate specializations. Therefore, they often occur in a vacuum, but rather within the context of a choose a more flexible strategy that encompasses broader national innovation system. enabling policies for entrepreneurship and market selection, rather than ex ante targeting (Correa and Ultimately, regional innovation strategies should be Guceri 2014). operable and easily understandable to the public. They should introduce a clear plan of action for Regional innovation strategies should rely on an inclusive implementation, while allowing for flexibility. Drafting a consultation process, including representatives of local, strategy and agreeing on objectives is the easiest part, regional, and national governments as well as all relevant while the biggest challenge lies in implementation. stakeholders (academia, firms, entrepreneurs, NGOs, Policy-makers should be careful to avoid any etc.). An expert group should assess regional strengths, incongruity between strategy and implementation. weaknesses, opportunities, and threats (SWOT), For example, in the case that regions design strategies using methodologies, such as production and trade for the sole purpose of securing more funds or that indicators, bibliometric indicators, technology foresight, they comply with higher level requirements, but then benchmarking, and independent expert assessments. In continue with “business as usual.” addition, regional innovation strategies should provide a stable budgetary framework and optimize all available financial resources, including decentralized funding (i.e., 3.2  Regional Capacity Building decentralization of allocations in centrally developed Many regions in developing countries lack the critical programs), autonomous funding (i.e., regional mass and capacity to tackle the costly, risky, and discretionary budget allocation), and regional power of path-dependent nature of innovation. Recently, taxation (Cooke et al. 1997). much of the push for decentralization in developing countries has primarily been on the expenditure side, Policy-makers should ensure that regional innovation while local revenue capacity has remained weak strategies translate into measurable goals and (Gadenne and Singhal 2013). This is a source of plan how specific policy instruments will achieve tension between the national and regional levels of those goals. Choosing the appropriate policy mix government and requires new systems of funding, is a complex process that involves selecting a set of as discussed in Section 2.2.2. Beyond funding, different and complementary policy instruments to lagging regions in developing countries are often address the constraints identified in the innovation characterized by “organizational thinness,” that is, by system. The objective is to unveil and maximize a lack of innovative firms, efficient public agencies, synergies, or positive interactions between policy and adequate knowledge infrastructure to support instruments, while avoiding duplication and negative collective learning (Asheim and Isaksen 1997). Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place – October 2014 11 Against this background, it is important that the link between innovation and economic convergence/ national government support the devolution of divergence. However, as examples from Mexico, competences with measures to build regional Russia, China and Colombia show, this is a challenging capacity. First, central governments can offer regional task. One challenge is an equity/efficiency dilemma, governments information, technical assistance, and in which concentrating public resources in the most training and advice. Second, the national government advanced regions helps to build critical mass and larger can establish mechanisms that identify and share best outputs (in terms of aggregate innovation performance regional practices. Third, they can foster collaboration and economic growth) than would concentrating among regions, especially where the regions face public support in underdeveloped regions. A second problems of critical mass or in the case of counter- challenge is what Oughton et al. (2002: 27) call the productive inter-regional competition. In particular, “regional innovation paradox,” which refers to the the central government can encourage collaboration contradiction between the comparatively greater need by providing incentives to joint regional initiatives. to spend resources on innovation in lagging regions For example, in Mexico, the National Science and and their relatively lower capacity, as compared to Technology Council launched the program FORDECYT more advanced regions, to absorb public funds. The to provide funds to groups of states that face common Mexican mixed funds and the Colombian royalties thematic challenges or a functional geographic need fund cases demonstrate this effect. Third, innovative (OECD 2009b). During Poland’s accession to the actors from the periphery to the core are often blocked European Union, the government formed a macro- from beneficial agglomeration effects found in more region called Eastern Poland, bringing together the advanced regions (McCann and Ortega-Argiles 2014). five poorest regions of the country in order to secure For example, lagging regions’ investments in advanced additional European funding. However, despite the training and support for entrepreneurs could potentially common challenges facing these regions, they could move toward core regions. not agree on a common innovation strategy, so the government abandoned the program in 2014 Addressing these challenges requires policies that (Piatkowski et al. 2014). increase regional capacity to absorb innovation-related funds and closer interactions between government, industry and universities (Oughton et al. 2002). In order Using Regional Innovation Policy to 3.3   to avoid the risk of regional fund migration to core Close Income Gaps and Address regions, investments should focus on regional strengths Societal Challenges and local industrial needs, and they should be directed, Innovation policy in developing countries should align wherever possible, toward “local fixed assets” (e.g., a more closely with the broader development agenda for local center of competence, a local university research fulfilling the basic needs of citizens, given that a large program to address the needs of a regional cluster, share of the population faces acute shortages of food, etc.). Another line of action is to encourage core water or electricity, as well as inadequate education regions to orient their research agendas toward the and healthcare. If properly managed, decentralization needs of peripheral regions. of innovation policy can contribute to this agenda, fostering bottom-up, participatory approaches that Collaboration between regions can build strong promote social innovation (Mercado 2012). absorptive capacity in lagging regions and reorient leading region agendas toward larger development As discussed in Section 2.1, national governments often challenges (e.g., health, hunger, education, rural attempt to use regional innovation policy to reduce development, etc.). However, policy-makers should income disparities across regions because of the direct carefully address barriers to cross-regional collaboration, 12 Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place – October 2014 such as political pressure facing regional policy makers context of policy. Moreover, because the returns on R&D to properly capture regional investment benefits, as in and innovation often appear only after sustained, long- the case of Colombia’s new royalties fund (Box 3). term investments, they require a long-term policy vision. Despite the need for stability, the policy mix also needs to adapt to new challenges and opportunities. Thus, Monitoring, evaluation and 3.4   program monitoring and evaluation should strive for a experimentation balance between stability and experimentation, allowing Monitoring and evaluation systems should formulate sufficient flexibility for the redesign of programs during precise objectives, identify relevant indicators, set their life cycles, according to changing conditions, realistic targets, and devise appropriate incentive interim assessments, and user feedback. mechanisms. Indicator systems should help inform both short-term decisions and long-term strategies. However, it is very difficult, in practice, to identify cause 4. CONCLUDING REMARKS and effect relationships between regional policy and Despite the trend toward decentralization, an excessive innovation outputs. Beyond the general difficulties enthusiasm toward the devolution of competences to associated with the intangible and long-term nature developing country regional governments would be of the returns of R&D and innovation, the effects of risky. National governments should take a strong role in regional programs are often intertwined with those of innovation policy by building institutional frameworks, national programs, such that it is difficult to attribute defining national strategies, seeking coherence and outcomes to a particular level of intervention. economies of scale, and avoiding wasteful duplication across regions. Yet, greater regional involvement in In any case, the central government should monitor innovation policy requires central governments to regional performance within nationally funded shift from omnipresent control to facilitation, capacity programs to ensure that the agreed upon objectives building and support. Policymakers should continue are met. Moreover, the central government’s to reflect on how to better approach the challenges monitoring and evaluation practices allow for regional of multilevel governance. This paper has offered an comparisons, standardize evaluation methods, and overview of different policy options and presented a serve as a platform for sharing experiences across review of recent international experiences, with the hope regions. Indeed, evaluations should aim to identify of fostering better-informed policy experimentation in successful regional programs and replicate them in developing countries. However, we still lack sufficient other regions. However, the monitoring and evaluation data on the correlation of policies and outcomes both system also needs to include the bottom-up dimension. because the examples we have reported here are still For example, the central government should establish very young, and because of the inherent difficulty in mechanisms to incorporate feedback from the regions measuring policy outcomes and institutional changes. to improve national policies. Central governments can also act as focal points for collecting, evaluating, and Finally, we would like to stress that a “one-size-fits-all” disseminating international experience. approach to regional innovation policy is inappropriate (Holbrook and Salazar 2004; Todtling and Trippl 2005). Policy evaluation should also determine whether regional Instead, a “flexible, decentralized policy process that innovation strategies are addressing potential strategy- takes the diversity of circumstances into account implementation incongruities (see section 3.1). Regional is needed” (World Bank 2010: 253). Regions are innovation strategies should provide a certain degree of significantly different from each other, each with its continuity as a sign of commitment to firms and other own technological trajectories, institutional settings, organizations, which make innovative decisions in the governance abilities, and industrial strengths and Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place – October 2014 13 weaknesses. And, this warrants a differentiated Botelho, A. J. 2013. “ERAWATCH Country Reports approach. Innovation policies and knowledge hubs 2012, Brazil.” Seville, Spain: Institute for Prospective in high-technology regions are necessarily different Technological Studies, European Commission. from those that would be appropriate for industrial production zones or yet from those that could Cassiolato, J. E. and Martins, H. M. 2000. “Local facilitate a resurgence of science and technology Systems of Innovation in Mercosur Countries.” in underdeveloped regions. From the perspective Industry and Innovation 7: 33–53. of the central government, this calls for a flexible and asymmetric approach to decentralization across Chaminade, C. and Vang, J. 2008. “Globalization of regions that takes into account each region’s unique Knowledge Production and Regional Innovation technological endowments, administrative capacities, Policy: Supporting Specialized Hubs in Developing and political wills. Countries.” Research Policy 37: 1684–1696. Cooke, P. 2001. “Regional Innovation Systems, ACKNOWLEDGEMENTS Clusters, and the Knowledge Economy.” Industrial The author is indebted to Paulo Correa, Natasha and Corporate Change 10: 945–974. Kapil, Esperanza Lasagabaster, Marcin Piatkowski and Jean-Louis Racine for their valuable guidance and Cooke, P., Uranga, M. G. and Etxebaria, G. 1997. suggestions. “Regional Innovation Systems: Institutional and Organizational Dimensions.” Research Policy 26: 475–491. REFERENCES Asheim, B. T. and Isaksen, A. 2002. “Regional Correa, P. and Guceri, I. 2014. “Research and Innovation Systems: The Integration of Local ‘Sticky’ Innovation for Smart Specialization Strategy: and Global ‘Ubiquitous’ Knowledge.” Journal of Concept, Implementation Challenges and Technology Transfer (27): 77–86. Implications.” Washington DC: The World Bank. Barmeyer, C. and Kruth, K. 2012. “The Pôles of Crespy, C., Heraud, J-A. and Perry, B. 2007. “Multi- Compétitivité: Regional Innovation Clusters with a Level Governance, Regions and Science in France: French Touch.” In: Heidenreich, M. (ed.). Innovation Between Competition and Equality.” Regional and Institutional Embeddedness of Multinational Studies 41: 1069–1084. Companies. Cheltenham: Edward Elgar. Cuervo, C.J., and López, L. 2013. “Regalías Para Basrhir, F. Z. 2011. “Morocco Boosts Innovation,” la Ciencia, la Tecnología y la Innovación: Caso Nature Middle East, Published online 15 May 2011, Colombiano”, in: OCyT (ed.) Observando el Sistema doi:10.1038/nmiddleeast.2011.57. Colombiano de Ciencia, Tecnología e Innovación, Bogota: Observatorio Colombiano de Ciencia y Bermann, G. A. 1994. “Taking Subsidiarity Seriously: Tecnología: pp. 369–407. Federalism in the European Community and the United States.” Columbia Law Review 94: 331–456. Eickelpasch, A., and Fritsch, M. 2005. “Contests for Cooperation — A New Approach in German Borras, S., and Edquist, C. 2013. “The Choice of Innovation Policy.” Research Policy 34: 1269–1282. Innovation Policy Instruments.” Technological Forecasting and Social Change 80: 1513–1522. 14 Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place – October 2014 Etzkowitz, H., Carvalho de Mello, J.M. and Almeida, Koschatzky, K., and Kroll, H. 2009. “Multi-level M. 2005. “Toward ‘Meta-Innovation’ in Brazil: The Governance in Regional Innovation Systems.” Evolution of the Incubator and the Emergence of a Ekonomiaz 70: 44–59. Triple Helix.” Research Policy 34: 411–424. Laranja, M., Uyarra, E. and Flanagan, K. 2008. “Policies Florida, R. 1995. “Toward the Learning Region.” for Science, Technology and Innovation: Translating Futures 27: 527–536 Rationales into Regional Policies in a Multi-level Setting.” Research Policy 37: 823–835. Freeman, C. 1987. “Technology Policy and Economic Performance: Lessons from Japan.” London: Pinter. Li, Y. 2013. “ERAWATCH Country Reports 2012, China.” Seville, Spain: Institute for Prospective Gadenne, L. and Singhal, M. 2013. “Decentralization Technological Studies, European Commission. in Developing Economies,” NBER Working Paper No. 19402. Lundvall, B.A. 1992. “National Systems of Innovation: Toward a Theory of Innovation and Interactive Gokhberg, L. and Roud, V. (2012) The Russian Learning.” Pinter, London. Federation: A New Innovation Policy for Sustainable Growth, in: Dutta, S. (Ed.) Global Innovation Index McCann, P. and Ortega-Argiles, R. 2014. “Smart 2012: Stronger Innovation Linkages for Global Specialization, Regional Growth and Applications to Growth, pp. 121–130. Fontainebleau, France: INSEAD European Union Cohesion Policy,” Regional Studies, and WIPO. forthcoming. Published online: 14 Jun 2013. Hassink, R. 2001. “Toward Regionally Embedded Mercado, A. 2012. “Social Inclusion or Social Innovation Support Systems in South Korea?: Case Illusion: The Challenges of Social Inclusion, Social Studies from Kyongbuk-Taegu and Kyonggi.” Urban Participation and Social Cohesion in Venezuelan S&T Studies 38: 1373–1395. Policy.” Science and Public Policy 39: 592–601. Holbrook, A. and Salazar, M. 2004. “Regional OECD 2009a. “Regional Innovation Systems in China: Innovation Systems within a Federation: Do Insights from Shanghai, Sichuan and Liaoning.” In: National Policies Affect All Regions Equally?” OECD Reviews of Innovation Policy: China 2008. Innovation: Management, Policy & Practice 6: Paris: Organization for Economic Cooperation and 50–64. Development. Howells, J. 1999. “Regional Systems of Innovation?” OECD 2009b. “OECD Reviews of Regional Innovation: In: D. Archibugi, J. Howells, and J. Michie (eds.), 15 Mexican States”, Paris: Organization for Economic Innovation Policy in a Global Economy, Cambridge: Cooperation and Development. Cambridge University Press. OECD 2010. “Contracts for Supporting Multi-level Kapil, N., Piatkowski, M., Navarrete, C. 2014. “Poland Governance of Innovation Policy.” Paris: Organization Smart Growth Operational Program Review.” for Economic Cooperation and Development. Washington DC: World Bank. OECD 2011. “Regions and Innovation Policy.” Paris: Klincewicz, K. 2014. “ERAWATCH Country Reports Organization for Economic Cooperation and 2012, Poland.” Seville, Spain: Institute for Prospective Development. Technological Studies, European Commission. Regional Innovation Policy and Multilevel Governance in Developing Countries: Between a Rock and a Hard Place – October 2014 15 OECD 2012. “OECD Reviews of Innovation Policy: Spiesberger, M. 2013. “ERAWATCH Country Reports Russia 2011,” Paris: Organization for Economic 2012, Russia.” Seville, Spain: Institute for Prospective Cooperation and Development. Technological Studies, European Commission. OECD 2014. “OECD Reviews of Innovation Policy: Taylor, M. Z. 2007. “Political Decentralization and Colombia.” Paris: Organization for Economic Technological Innovation: Testing the Innovative Cooperation and Development. Advantages of Decentralized States.” Review of Policy Research 3: 231–257. Oughton, C., Landabaso, M. and Morgan, K. 2002. “The Regional Innovation Paradox: Innovation Policy Todtling, F., and Trippl, M. 2005. “One Size Fits All? and Industrial Policy.” The Journal of Technology Toward a Differentiated Regional Innovation Policy Transfer 27:97–110. Approach.” Research Policy 34: 1203–1219. Perry, B., and May, T. 2007. “Governance, Science World Bank 2010. “Innovation Policy: A Guide for Policy and Regions: An Introduction.” Regional Developing Countries.” Washington, DC: World Bank. Studies 41: 1039–1050. World Bank and the Development Research Center Piatkowski, M., Szuba, T. and Wolszczak, G. (2014). of the State Council 2013. “China 2030: Building “Review of National and Regional Research and a Modern, Harmonious, and Creative Society.” Innovation Strategies for Smart Specialization (RIS3) Washington DC: World Bank. in Poland”. Washington, DC: World Bank Group. Zeng, D.Z. 2011. “How Do Special Economic Rip, A. 2002. “Regional Innovation Systems and the Zones and Industrial Clusters Drive China’s Rapid Advent of Strategic Science.” Journal of Technology Development?” World Bank Policy Research Working Transfer 27: 123–131. Paper 5583. Ritzen, J., and Soete, L. 2011. “Research, Higher Zhang, M. 2010. “Competitiveness and Growth Education and Innovation: Redesigning Multi-level in Brazilian Cities: Local Policies and Actions for Governance within Europe in a Period of Crisis.” Innovation.” Directions in Development, Private Sector UNU-MERIT Working Papers 2011–56. Development. Washington DC: The World Bank. Shin, J. C. 2009. “Building a World-Class Research University: The Brain Korea 21 Project.” Higher Education 58: 669–688. 16 www.worldbank.org 16