Smart specialization and place leadership: dreaming about shared visions, falling into policy traps?

ABSTRACT The concept of smart specialization has rapidly acquired a central position in European policy and academic circles. It raises interesting challenges for the regional studies agenda. First, smart specialization is about not only policy formulation, implementation and evaluation but also pooling scattered resources, competencies and powers to serve both shared and individual ambitions. Thus, policy formulation and implementation need to be seen in a new light. Second, when smart specialization is seen as one of the platforms for aligning several actors to boost regional economic development, the need to understand agency in its multiplicity emerges as central. This paper argues that to achieve truly transformative smart specialization strategies, there is a need to investigate in more depth the multi-actor strategy processes and new forms of leadership, as well as to invest time and money in advancing related capabilities across European regions.


INTRODUCTION
The concept of smart specialization has rapidly acquired a central position in European policy and academic circles. It is an institutionalized European strategy approach that serves as an 'ex-ante conditionality' for the European Union Structural Funds (2014-20) (Martínez-López & Palazuelos-Martínez, 2015). According to Morgan (2017, p. 569), 'smart specialisation is the most ambitious regional innovation programme ever to be launched in the European Union', and 'it affords a unique opportunity to explore the interplay between institutions, innovation and development'. Capello andKroll (2016, p. 1393) argue that smart specialization is a major change in the ways innovation policies are formulated and implemented. They include in their list of important changes the fact that innovation is no longer banded together only with research and development (R&D) expenditures and employment; thus, policy understanding of the very nature of innovation has broadened beyond high-tech and product portfolios. Importantly, the European Union has finally matured to abandon the conviction that it might be possible to find a one-size-fits-all policy serving the very different regions of Europe (Capello & Kroll, 2016, p. 1393.
Smart specialization raises interesting challenges for the regional studies agenda. First, it is about not only policy formulation, implementation and evaluation but also pooling scattered resources, competencies and powers to serve both shared and individual ambitions. Thus, policy formulation and implementation need to be seen in a new light. Second, when smart specialization is seen as one of the platforms for aligning several actors to boost regional economic development, the need to understand leadership in its multiplicity emerges as central. To achieve truly transformative smart specialization strategies, there is a need to investigate in a more in-depth manner the multi-actor strategy processes and new forms of leadership providing them with directions, as well as to invest time and money in advancing related capabilities across European regions.
While the smart specialization literature is well informed by the latest advancements in regional development studies and economic geography, it is less aware about studies on multiactor policy processes and related studies on agency. Consequently, this paper argues that smart specialization calls for a better understanding of the ways power and governance systems are acted on, and how various actors actually come together to work collectively for regional development and innovation. For these purposes, it is argued that the place-based leadership approach would add an agentic lens in studies focusing on how various sectoral interests and individual ambitions can be overcome, and hence to see, as called for by smart specialization advocates, the place and economic development holistically instead of focusing only on organizational or sector-based interests (see also Beer & Clower, 2014;Collinge, Gibney, & Mabey, 2011;Sotarauta, 2005).
In essence, smart specialisation is about identifying 'the unique characteristics and assets of each region' (Heimeriks & Balland, 2016, p. 562) that underline competitive advantages, and 'mobilizing regional stakeholders and resources around an excellence-driven vision of their future' (Heimeriks & Balland, 2016, p. 562; see also McCann & Ortega-Argilés, 2013). This paper focuses on the latter issue by asking (1) what are the main obstacles that need to be overcome to realize the smart specialization ideal; and, more specifically, (b) what is the place of a shared vision in mobilization of heterogeneous groups of actors. In other words, its main aim is to identify the dynamics between smart specialization and place-based leadership, and by doing so connect tentatively two disparate fields of study. The emphasis is on the obstacles preventing shared visions from being constructed and stakeholders from being mobilized, and consequently, it is argued that by paying more attention to place-based leadership also in the context of smart specialization, we might gain additional analytical support for improving the governance capacity and policy capabilities in different European regions to use the new policy approach.
The paper is structured as follows. The next section examines the basic tenets of smart specialization (smart specialization ideal), as advocated by the European Commission and several investigators. The paper then moves on to discuss the policy traps (main obstacles) that were identified by drawing upon the literature on smart specialization and also place leadership and governance. The third section introduces place leadership as an approach that would allow one to learn more about how stakeholders are mobilized and their activities coordinated and also about policy traps and how they might be removed. The fourth section highlights the importance and challenges of imaginaries guiding smart specialization, and reminds both smart specialization and place leadership scholars how multidimensional a vision actually is. regional economic development paths based on micro-level operations (Benneworth, Pinheiro, & Karlsen, 2017;Lagendijk, 2007): The fundamental point here is the Hayeckian argument whereby the knowledge about what to do is not obvious. It is knowledge 'of time and place'; this is local knowledge which is dispersed, decentralized and divided. It is hidden and needs to be discovered. (Foray, 2016(Foray, , p. 1433 Smart specialization is defined as 'a process of discovery and learning on the part of entrepreneurs, who are the best positioned agents to search for the right types of knowledge' (Estensoro & Larrea, 2016, p. 1321. Consequently, the smart specialization guidelines put the entrepreneurial discovery process (EDP) in its core (Rissola, Kune, & Martinez, 2017, 24). The EDP is defined to be an: inclusive and interactive bottom-up process in which participants from different environments (policy, business, academia, etc.) are discovering and producing information about potential new activities, identifying potential opportunities that emerge through this interaction, while policymakers assess outcomes and ways to facilitate the realisation of this potential.
Moreover, according to the guidelines: the EDP pursues the integration of entrepreneurial knowledge fragmented and distributed over many sites and organisations, companies, universities, clients and users, specialised suppliers (some of these entities being located outside of the region) through the building of connections and partnerships. (Rissola et al., 2017, p. 24) The EDP is dependent on entrepreneurial agents, who, according to Coffano and Foray (2014), are the sources of the 'entrepreneurial knowledge' that again is the foundation upon which regional innovation strategies are formulated. Importantly, the concept of entrepreneurial agent does not only refer to Schumpeterian entrepreneurs (Shane & Venkataraman, 2000), but may take many forms, e.g., higher education institutions, public research institutes, independent innovators and firms. The entrepreneurial agents are supposed to be insightful and knowledgeable actors who see opportunities and are willing to take risk. Thus, EDP aims to move away from the identified problems of vertical policy programmes that are targeted to serve a selected industry or technology area to tackle needs that are not easy to identify or fulfil. Therefore, smart specialization is about identifying the characteristics, unique assets and competitive advantages of regions by assembling key actors and their resources around an excellence-driven vision (McCann & Ortega-Argilés, 2013, pp. 1298-1299. As an inclusive process, smart specialization is expected to answer one of the main challenges of policy-making: how to involve a wide spectrum of agents in policy design, and how to choose priorities in a multi-actor and multipurpose context that reaches beyond sectoral interests and public sector predominance. As an integrated and place-based economic transformation agenda, smart specialization, ideally, flows from outward-looking analysis to the creation of a vision, and then to setting objectives and priorities, and finally to implementation. The strategy process is supposed to be embedded in an inclusive governance structure, a capacity-building toolbox and an evaluation system. Smart specialization advocates for the regional government to bear the primary responsibility for formulating strategy and identifying key policy objectives (Rodríguez-Pose, di Cataldo, & Rainoldi, 2014).
Of course, as Foray points out, new development paths often emerge spontaneously in a decentralized manner without any specific government intervention to boost specific processes. In these kinds of cases, transformational processes are brought about by innovative entrepreneurs and the integration of a variety of knowledge to translate discoveries to economic success (Foray, 2014, p. 492). But, as is also argued, many regions lack a strong enough entrepreneurial vision to trigger future transformational processes; and, therefore, as reminded by Landabaso (2014, pp. 388-389), in some cases, the public sector is called to show courage in finding new development paths and coming up with a vision to unwrap the policy processes so that 'the non-usual suspects' may also have a seat and voice in regional development efforts.
Moreover, the smart specialization approach is expected to help 'inefficient regional administrations become accustomed to external connections and be confronted with practices and experience coming from outside, challenging inertia and clientelism which prevail in locked-in systems' (Rodríguez-Pose et al., 2014, p. 10). Smart specialization is thus expected to be of use not only in constructing a shared vision and pooling scattered resources but also in fixing the problems of poorly functioning governance systems and lack of leadership. Navarro, Aranguren, and Magro (2011), among many others, believe that the broad engagement of the main stakeholders resulting in a clear and shared vision is a prerequisite to solving the many regional development related issues. A strategy formulated in this way is expected to allow all the stakeholders to target their own actions, resources and capabilities to support a collective strategy. Foray's (2014) fundamental point at the beginning of this section includes a decisive set of questions and the smart specialization literature provides tentative answers, leading to an additional set of tricky questions: . Who knows what to focus regional development efforts on? Not a single actor; knowledge is dispersed. . How can dispersed knowledge be pooled to better serve the common purpose of regional development? By organizing a collective strategy process resulting in a shared vision and collective strategies. . What kinds of agency are needed? Public actors are expected to launch the process and mobilize the key stakeholders. Entrepreneurial agents are needed to identify the opportunities and reveal the ways to achieve them. Public actors are responsible, based on the constructed understanding of what should be done, for leveraging resources and powers to change the local conditions to support transformational processes. . New questions: what issues need to be taken account when aiming to construct a shared vision and mobilize entrepreneurs to contribute to policy formulation; which factors may prevent mobilization and visioning; and how should one lead complex multi-actor modes of policy formulation and their implementation? In other words, what kind of place-based leadership is needed to meet all the expectations?

SMART SPECIALIZATION TRAPS
The smart specialization strategy has failed to explain concretely how the concept could provide a common political rationale for a socio-economically and territorially diverse set of regions and nations facing different place-based challenges and different innovation modes, hence, quite legitimately, different general policy agendas. (Capello & Kroll, 2016, p. 1396 Capello and Kroll (2016Kroll ( , p. 1395 argue that the blunt smart specialization agenda is not an adequate policy answer to the many difficulties faced in practice. They add that in some European Union member states, the smart specialization exercise may have been carried out without any significant impact. Following and endorsing their view, some observations from earlier studies concerning the potential bottlenecks of smart specialization strategies are elaborated next, and the policy reality is discussed using five highly interlinked smart specialization traps. The metaphor of a trap refers to a situation where actors are misled into acting contrary to their interests or intentions and hence to an unpleasant situation that is hard to escape (Merriam-Webster Collegiate Dictionary). Next, the following five traps are discussed: the institutional conflict trap, governance trap, mobilization trap, shared vision trap and capability trap. These are treated as 'process traps'; traps related to substantial issues are not dealt with in this paper. Perhaps the most obvious of the smart specialization trapsthe implementation trapis not included. All too often, the problems of policy-making and strategic planning are reduced to questions related to why the intentions were not realized as planned. The traps cannot be solved simply by designing better strategies, engaging more stakeholders or hiring better consultants; it is much more complex than that. The implementation trap, or the implementation gap, would by no means be a novel target of interest (Sabatier & Mazmanian, 1980), but it would provide a misleadingly linear and straightforward framework with which to assess the lack of results. Here, an implementation gap is not seen as a trap as such but rather as a natural outcome of actors being trapped by one or more of the traps discussed below.

Institutional conflict trap
Many earlier studies have highlighted the importance of institutions in regional development (Gertler, 2010;Rodríguez-Pose, 2013;Zukauskaite, 2018); often, the lack of supportive institutions is seen as one of the main issues in failed efforts to boost regional development. Coffano and Foray (2014) state that also smart specialization policies require supportive institutions and strong policy capabilities at the regional level. Of course, as Capello andKroll (2016, p. 1396) find, the importance of institutional arrangements is hardly specific to smart specialization but is a determining factor in any sort of policy (Rodríguez-Pose, 2013). A quality governance system, political goodwill and non-corrupt actors are among the key building blocks of economic development and related development efforts (e.g., Robinson & Acemoglu, 2012;Rodríguez-Pose, 2013;Rodríguez-Pose et al., 2014). In the context of smart specialization, the existence or non-existence of supporting institutions is of utmost importance, but it is equally important to have institutions that are not in conflict with each other. In practice, actors are rarely provided with clear answers on what is and is not possible or on what is rewarded and what is sanctioned (Beckert, 1999, p. 780). In these kinds of situations, actors might fall into institutional gaps (van den Broek & Smulders, 2014, p. 158).
In many regions, institutional conflicts can harm smart specialization efforts more than is acknowledged. According to Grillitsch (2016, pp. 29-30), dissipating institutional integration can lead to political challenges that might undermine the institutional support for a smart specialization strategy. Moreover, as Grillitsch maintains, conflicting institutions can hamper networks, trust among key actors, their ways of engaging in collective processes and the coordination of different interests. Grillitsch (2016, pp. 29-30) summarizes the importance of 'institutional harmony' by saying that 'if the degree of integration is high, conflicting interests can be better mediated and a consensus built supporting the development of a shared vision for the development of the region'. Consequently, it would be important not to fall into an institutional conflict trap that might dissipate collective efforts, and thus also make construction of a shared vision more than challenging. Consequently, also trust-based social relations might be hampered or prevent them from being built.

Governance trap
The institutional conflict trap is closely related to the governance trap. A smart specialization strategy, as a bottom-up and entrepreneur-led collective endeavour, is dependent on subnational-level actors' autonomy and power to make choices and decisions for setting collective objectives, finding a shared vision and achieving place-based objectives (Barca, 2009;Tomaney, 2010). As Bentley, Pugalis, and Shutt (2017, p. 6) conclude, 'the system of governanceon a continuum from centralism to localismis a determining factor of the scope for place-based leadership of sub-national bodies', and thus for smart specialization as well. The governance and place leadership literature has argued that devolved, decentralized and/or localist governance systems provide better prospects for place-sensitive strategy processes and related leadership than centralist systems of governance. A centralized system may weaken the capacity to act at subnational levels (Beer, 2014;Bentley et al., 2017). All this calls for an understanding of governance that would boost smart specialization efforts instead of slowing them down. This may prove crucial, since smart specialization is, according to its ideal, a long process pursued by a heterogeneous group of actors via a process of collaborative governance.
Supportive modes of governance are crucial for smart specialization strategies, as in most countries, related development resources, capacities and policy tools are distributed across various ministries, regional development agencies and other organizations, which leads to complex public-public coordination side by side with complex public-private interactions (Capello & Kroll, 2016). Due to the imperfect communication and coordination between many parties, it is not easy to assemble and pool scattered expertise in a concentrated working group (Capello & Kroll, 2016, p. 1398. Therefore, any smart specialization effort should be organized according to the governance system in question. More importantly, if possible, the system ought to be renewed to support better smart specialization-related work.

Capability trap
The smart specialization literature, side by side with the main bulk of the regional innovation system and policy literature, emphasizes the capabilities needed in pursuing discoveries and innovation. However, as Laasonen and Kolehmainen (2017) observe, there is a lack of studies focusing explicitly on the capabilities needed in knowledge-based regional development and related innovation policies. The lack of capabilities and underdeveloped understanding of them may also trap smart specialization strategies. Navarro et al. (2011) expect governments to assume a more central role than that of mere facilitators in situations where regional actors lack scientific and technological capabilities or the capability to generate a systemic vision for smart specialization strategies. They also state that in regions having actors with these capabilities, governments may remain in a purely facilitative role. Navarro et al. seem to assume that governments have a choice: that they would somehow be capable of selecting a role and show natural talent in pursuing the tasks that come with varying roles. In many regions, this is not the case. The capabilities associated with smart specialization need to be consciously developed.
First, as Kroll, Muller, Schnabl, and Zenker (2014) maintain, many regional development agencies simply do not have enough competent staff to take care of smart specialization tasks adequately. Second, according to them, some regions rely too much on external consultants who do not have specific knowledge of the region and its resources, actors and main challenges (Kroll et al., 2014). Third, many regional officers do not have sufficient training to manage and organize smart specialization strategies but have experience in law, public administration or spatial planning; therefore, they simply do not have the capabilities and knowledge to mobilize actors and coordinate complex, continuously evolving processes (Kroll et al., 2014). For these reasons, many regional development officers have had a hard time learning how smart specialization procedures differ from traditional priority-setting industrial policies (Kroll, 2015). If a smart specialization strategy fell into a capability trap, it would have a hard time fulfilling the policy dream.

Mobilization trap
In smart specialization thinking, it is imperative to mobilize actors with knowledge and experience successfully. The smart specialization approach believes strongly in the capacities of entrepreneurs to identify and frame the technologies and opportunities to be focused on and selected as spearheads in regional development and innovation. There is plenty of evidence that innovative entrepreneurs are capable of perceiving emerging opportunities and working on them (Shane & Venkataraman, 2000). However, there is less evidence that entrepreneurs would be willing to contribute to collective and public interest-driven discovery processes.
Martínez-López and Palazuelos-Martínez (2015, p. 6) recognize that picking out actors is not 'the end of the road towards a successful entrepreneurial process of discovery'. First, there is no reason to expect that everybody selected and invited to join a process would be willing to invest their time and money in such a collective strategy as smart specialization. Second, there is no reason to expect that all the stakeholders would grasp and respect the importance of a collective strategy process or the rationale behind searching for a shared vision. Third, it remains to be seen whether mobilized entrepreneurial actors will be willing to open up their thinking to other entrepreneurial actors who are potentially their competitors. Fourth, political logic drives rent seeking, which is inclined to dilute the many efforts to concentrate funding on policy priorities or to reach a consensus on them in the first place or construct a shared vision (McCann & Ortega-Argilés, 2013). It would be overly idealistic to assume that various stakeholders would be able to leave their own incentives, logics, drivers, ambitions, visions and paymasters behind and think only of the region's best interests. Fifth, a bottom-up process of strategy formulation may run the risk of misallocating public resources, and that might be an outcome of an unbalanced mobilization of actors (Camagni, Capello, & Lenzi, 2014;Capello, 2014;Estensoro & Larrea, 2016). Hence, it is important to identify and appreciate different reasons to join a collective process. Sixth, as observed by Benneworth et al. (2017), the emergent, informal and flexible mobilization of actors may also have a counter-effect of implicitly directing development work to irrelevant issues or serving only marginal interests. Moreover, some actors may have a tendency to voice such radical visions that other actors would not be willing to accommodate or be capable of accommodating them (Horlings & Padt, 2011), potentially leading to decreasing commitment and thus deepening the mobilization trap. Seventh, the non-selection of actors not perceived as 'fitting' may increase social tensions; hence, something important could be left completely aside (Benneworth et al., 2017;Horlings & Padt, 2011).
The many-sided involvement of stakeholders is often a primary pursuit in place-based development ideals (Barca, 2009;Tomaney, 2010), but at the micro-level it is a delicate art to mobilize the right type of assembly for the issues in question. In addition, there is a connection to other traps: the ways actors are mobilized depends on the governance system and capabilities, and institutional conflicts may make mobilization harder than it should be.

Shared vision trap
Smart specialization emphasizes the significance of a shared vision in guiding multi-actor development work; a shared vision is believed to be the basis for collective action. However, the view of a shared vision is one-dimensional. It is seen to emerge from collective contemplation or as a result of entrepreneurial insight, but in the smart specialization literature it is not discussed in depth and its many dimensions are not appreciated. The main question is whether a truly powerful shared vision can be constructed to serve as expected as an initial step in the various efforts to boost structural change. Another question is whether it could serve as a mechanism for producing information about potential innovations, spillovers and structural changes. Ideally, this kind of visionary information should be used by the government when the time comes to make choices and allocate funds (Foray, 2016(Foray, , p. 1434. There are well-placed arguments calling for more dynamic and democratic regional development tools and practices as well as place-based leadership and strategy approaches leading to a shared vision, but in spite of well-intentioned aspirations, there is always a danger that a collective endeavour will wind up being an unhealthy pursuit of self-/party interests or a purely market-driven agenda (Nicholds, Gibney, Mabey, & Hart, 2017). This might derive from overly centralized decision-making (governance trap), conflicting regulative and normative rule systems (institutional conflict trap) or a biased assembly of actors designing a strategy and communicating their vision (mobilization trap). Various distributed and open knowledge-oriented strategy and policy processes contain a promise of inclusive policy-making, but they may also disguise institutionalized power inequalities under a nicely wrapped strategy package. Indeed, there are often difficulties in combining abstract visionary thinking with operational matters that leaves space for self-or party interests to hijack a collective arena. In multi-actor development work, shaky trust, the fear of losing autonomy, a multitude of communication problems and the inevitable differences in individual visions and interests are always present in one form or another (Horlings & Padt, 2011).
Consequently, there is a need to be aware of the political and social dimensions of regional development and innovation, and not to approach it purely as a technical or economic procedure. This may change the ways in which shared visions are approached and used. MacNeill and Steiner (2010) claim that leading a process is more important than organization or strategic planning in achieving strategic success. Healey (1997) points out that the process should focus on specific actions being invented through the inclusive and interactive process by (1) drawing, shaping and organizing attention, and then deflecting it to the questions and issues that need to be faced (Heifetz, 1994); and (2) formulating and reformulating problems and opportunities, which is central to addressing significance and leading belief systems (Sotarauta, 2016). In this thinking, smart specialization strategy is to be seen as an arena for discussions, battles and quarrels, and one day a region might be mature enough to construct genuinely collective strategies and shared visions. A smart specialization strategy process can thus also be seen as a quest for awareness building, learning a shared language and vocabulary for addressing the main issues, and constructing collective beliefs (Sotarauta, 2016).
The assumption is that a genuinely shared vision and collective mutually supporting action may be achieved if and only if we respect the multitude of visions and values and do not believe in one grand shared vision only. Of course, there is a need consciously to build the capacity to act and think collectively, but, in many regions, collective thinking is not embedded in the institutional environment or systems of governance. Smart specialization processes might be used as platforms for capacity building.

IS PLACE LEADERSHIP A PRECONDITION FOR A SUCCESSFUL SMART SPECIALIZATION STRATEGY?
Smart specialization guidelines suggest that smart specialization strategies might and should provide regional development policy-makers and practitioners with renewed leadership in the many efforts to boost regional development. However, one can also argue that smart specialization has a better chance to make a difference in regions with a well-established and shared place-based leadership. Charles, Gross, and Bachtler (2012) highlight the importance of leadership and argue that it is central to smart specialization processes. According to them, collaborative leadership and commitment are crucial in moving forward on this front. They argue that smart specialization is 'a requirement which needs to be based on strong regional consensus and/or close integration with domestic strategic programmes' (p. vi). In achieving this, the place leadership literature implies the importance of taking responsibility for choices and related decisions that are subsequently enacted (Bentley et al., 2017), and thus it is argued that the place leadership approach would contribute to smart specialization studies and strategies by digging deep into the ways individual or grouped leaders' capabilities are attached to a smart specialization strategy in a given context and place.
Without knowing it, Bentley et al. (2017, p. 5) construct a conceptual link between smart specialization, vision and leadership by saying that 'place-based leadership is considered to improve the capacity to generate future-oriented spatial visions as well as increasing the likelihood of realizing visions'. They add that leadership could be '(re)conceptualized as the capacity of the coming together of actors to realize (collaborative governance) ambitions' (p. 5). Nicholds et al. (2017) conclude it is possible to cultivate multilevel and shared leadership attuned to political strategies and aimed at balancing power among competing vested interests, relying on communication, negotiation and the management of broad sets of relationships. On their part, Charles et al. (2012) observe there are major differences between regions in their capacity to launch and implement impactful innovation strategies. According to them, the main differences are due to lack of legitimate leadership and sound managerial processes. Similarly, Kempton, Goddard, Edwards, Hegyi, and Elena-Pérez (2014) pay attention to the lack of collective leadership that makes it difficult to articulate the needs of a wider region and mobilize regional stakeholders. Moreover, as they point out, if the powers to act are limited and leadership poorly developed, it is difficult to accomplish coordination between local and national initiatives.
Consequently, place leadership would complement more established smart specialization studies by capturing the micro-dynamics of large-scale social and economic co-development processes, which incorporate a wide range of power and resource-related, individual and communitybased agendas as well as negotiations across wide spectrum of people, organizations, disciplines and professions (Nicholds et al., 2017, p. 3). This might prove useful, as place leadership operates in a normative and interactive space that is both multilateral and multidisciplinary (Budd & Sancino, 2016).
Place leadership is about a relational and as such interactive and collective form of agency shaping, as well as being shaped by place-specific and broader contexts (Beer et al., 2018;Collinge & Gibney, 2010;Hu & Hassink, 2017;Sotarauta, Beer, & Gibney, 2017), and thus it might serve the efforts to unravel the secrets of place-specific and socially embedded specialization efforts, too. This might prove important not only in supporting mobilization or the construction of a shared vision for smart specialization but also in identifying the covert forms of power and influence in it as well as skewed mobilization patterns and subsequent development assemblies. The many contradictory institutional pressures and conflicts affect development efforts in many ways (Benneworth et al., 2017), and understanding how these pressures influence also smart specialization processes may turn out to be crucial, as personal embeddedness boosts development initiatives, builds leadership capacity and is used when mobilizing regional public bodies or business elites and representatives of media (Raagmaa & Keerberg, 2017).
All in all, by definition, place leaders are actors who have the capacity to organize and reorganize social action with an ambition to change the social networks and institutions framing the factors in which smart specialization is embedded (Sotarauta, Horlings, & Liddle, 2012). Importantly, as Hu and Hassink (2017) maintain, place leadership does not usually produce instantaneous impacts on local/regional economic dynamics. Rather, it pools scattered resources, capabilities and powers for shaping institutions for future development as intermediaries in economic practices and interactions (Bathelt & Glückler, 2013). Of course, both the internal and external institutional arrangements of a region shape and frame its evolution, and place leaders aim to work across the many institutional divides. Place leadership appears as crucial to stimulating, coordinating and mobilizing smart specialization agendas with the responsibility to generate something new.
In sum, place leadership is the process of reconciling conflicting and competing interests aimed at generating collaborative advantage and an understanding of the challenges associated with transforming places as well as organizations and capabilities (Trickett & Lee, 2010, p. 434). Place leaders are contextually embedded agents who are able to identify, communicate, translate and influence place-specific challenges and opportunities (Bailey, Bellandi, Caloffi, & de Propris, 2010). The concept of place leadership might add analytical leverage in the efforts to understand how policy traps affect smart specialization practices and what kinds of capabilities need to be learned for removing the traps or navigating across them.

DISCUSSION: DO IMAGINARIES GUIDE SMART SPECIALIZATION AND PLACE LEADERS IMAGINARIES?
Under global capitalism, it is economic imaginaries that prevail, although most imaginaries also address the broader embedding of the economic in social (political, ecological) formations. Imaginaries serve to define subjects and objects of regulation and to articulate visions underpinning particular strategies and projects. (Lagendijk, 2007(Lagendijk, , p. 1199 Visions, brands, images, narrativesall sorts of imaginariesplay important roles in socioeconomic and political developments. They are among the most powerful providers of future directions for many actors, often implicitly. According to Allen and Cochrane (2007), the politics and agency revolving around regional development are attempting to construct wider visions of change, and as discussed above, also smart specialization approaches stress the importance of shared visions. This might indeed prove important in the attempts to set wide-reaching rules and principles within which individual ambitions and decisions may be reflected, thus opening up the search for alternative visions (Allen & Cochrane, 2007). Pike, Rodríguez-Pose, andTomaney (2007, p. 1255) stress the importance of framing: what local and regional development is in the present, and what it can or could be in terms of future visions. And, normatively, what it should bein the sense of people in places making value-based judgements about priorities and what they consider to be appropriate 'development' for their localities and regions.
By having a more explicit focus on place leadership, we might be able to shed a light on which actors or groups (or which interests) in reality influence dominant imaginaries framing smart specialization initiatives, as well as how and why they do so. In the long run, seeing regional development and related smart specialization strategies as contested arenas for the search for future visions might support specialization efforts better than simple technical instructions on how to design a better strategy and come up with a shared vision. Accordingly, a strategy process that is embedded in a deep sociocultural understanding of leadership and related power and influence systems in a specific place, as well as their strengths and shortcomings, would root strategies in the sociocultural-political-economic fabric of a place instead of its administrative machinery and a few selected stakeholders only. There is also always a need to keep in mind that clientelism is not absent from collective strategies and construction of shared visions; the mobilizers and/or the ones who are mobilized can use their position in the process to favour those who share their vision. Actors with formal institutional power and hence the capacity to influence wide networks can use their social positions to impose their specific interests and visions into a smart specialization strategy, and thus they might end up amplifying path dependency instead of enhancing branching into new directions. Fortunately, there are also visionary, entrepreneurial and inspirational leaders influencing in an open and inclusive debate the course of events both in-and outside the regional development regime. Hu and Hassink (2017, p. 4) show how pioneering leaders can influence actors' cognitions 'by inviting them into a new visionary context for future change'.
A vision is not only about communicating desired futures but also something to fight and argue about, to support or attack. In the hands of skilful place leaders, vision is a powerful tool for exercising interpretive leadership when navigating across the many smart specialization traps and/or working to remove them step by step. The best visions create a tension between the past, present and possible as well as imaginable futures that touches the deepest emotions of the stakeholders. Such visions work best if they acknowledge the competing visions of individual actors or groups of actors. This is important because smart specialization and related leadership is, by definition, dealing with multiple visions and efforts to align people around something common, allowing transformative action in collaboration. In this thinking, identifying the entire spectrum of visions and learning from them may be the precondition for something that can be labelled as shared.
Ultimately, as Pike et al. (2007Pike et al. ( , p. 1266 state, even though the many desires and expectations expressed in visions for regional development involve all sorts of conceivable problems, not having a vision would make regional development work even harder.

CONCLUSIONS
This paper argues that smart specialization necessitates finding new developmental strategies and also understanding the importance of place leadership; the ways actors navigate and/or remove the policy traps towards collective action. The concept of place leadership might add value in the many efforts to make smart specialization strategies fulfil policy dreams.
This paper adds to the literature on regional economic development and innovation, specifically smart specialization, by reminding one that the EDP and shared vision are not purely technical but essentially social and political processes. As such, there is a need to scrutinize the basic assumptions of smart specialization and improve its potential to produce hoped-for results, as well as to find its true place in the regional development chart by identifying the ways it might be used constructively for a collective benefit or misused to benefit narrow interests. Most importantly, there is a need for more studies focusing on the basic assumptions behind smart specialization, analytically examining both the processes related to it and its canonical position as the regional development approach.
This paper also adds to the smart specialization and place leadership literature by exploring the important but understudied link between these two core concepts. In this way, it joins earlier studies that call for a more nuanced understanding of agency in the context of regional development and innovation (e.g., Beer & Clower, 2014;Collinge et al., 2011;Lagendijk, 2007;Sotarauta, 2005Sotarauta, , 2016Uyarra, Flanagan, Magro, Wilson, & Sotarauta, 2017). This paper is a step forward, among many others, in the efforts to link purposive but interactive agency to the literature, which has, for some time, been criticized for seeing policy-makers and practitioners simplistically (Uyarra, 2010;Witt, 2003). Of course, given the uncertainties and complexities surrounding regional development and, more broadly, social and economic change, it should not be assumed that achieved and visible transformations are fully intentional. There may be a series of intentional triggers as well as complementing and competing intentions, but as has been shown, transformations are often creeping by nature, and the interplay between intention and emergence needs to be respected (e.g., Djelic & Quack, 2007;Garud & Karnøe, 2001;Sotarauta, 2016).
Place leadership could provide future studies on smart specialization with a conceptual toolkit to dig deeper into the processes, but not a magic bullet for how to manage them. All in all, following Trickett and Lee (2010, p. 434), one might argue that place leadership and smart specialization require a well-established spatial literacy of place and process.

ACKNOWLEDGEMENT
The author is grateful to Robert Hassink for his profound comments.