Contextualising the context in policy entrepreneurship and institutional change

While there is a considerable literature concerning policy entrepreneurship and institutional isomorphism, significantly less literature has emerged addressing the impact of context on policy and institutional entrepreneurship and of the interactions between various contexts and agency. In this article, we demonstrate that the actions of entrepreneurs in the public sector are most likely to generate policy and institutional changes when they are reinforced by complementarities arising from context-dependent, dynamic interactions among interdependent structures, institutions and agency-level enabling conditions.


Introduction
Policy entrepreneurs as individuals and the process of policy entrepreneurship (Ackrill & Kay, 2011) have long been central topics in public policy research as a means of exploring policy change (Barzelay & Gallego, 2006;Béland & Howlett, 2016;Capano, 2009;Gunn, 2017). This is because policy entrepreneurs are recognised as key individual actors in the policy-making process (Béland & Howlett, 2016;Cairney & Jones, 2016;Jones et al., 2016;Mintrom & Norman, 2009;Saetren, 2016;Saurugger & Terpan, 2016). In particular, policy entrepreneurs sell policy ideas to elite decision-makers and thus play a key role in governmental agenda setting (Kingdon, 1995). In this context, it is argued that '[n]o entrepreneur alone will ever be enough to cause policy reform; we always require an account of the context' (Ackrill, Kay, & Zahariadis, 2013, p. 879) and '[ag]ency and context … should not be viewed in isolation but as linked through strategy' (Zahariadis & Exadaktylos, 2016, p. 62). Further, it is also recognised that policy entrepreneurship is a 'context-specific activity': 'The skilled advocacy of a policy idea, or skilled brokering' , will only produce reform in some political and temporal contexts: this is the causal structure behind the maxim 'ideas have their time' (Ackrill & Kay 2011, p. 76). However, despite these widely held views, past literatures have under-examined the interactions between macro, meso and micro-level contexts, and the agency of policy entrepreneurs in the policy entrepreneurship process.
Thus, in this article, we call for an eclectic integrative approach to examine the process of entrepreneurship and institutional change. We argue that the effect of interactions between contextual influences and agency-level enabling conditions on the actions of individual agents is fundamental to entrepreneurship in the public sector and thus to institutional change. Our aim here is to understand who these individual actors are, and how interactions within macro, meso and micro level contexts inform what they do and deliver specific public policy-making processes (see also Bakir, 2009aBakir, , 2009bBakir, , 2013. We argue that policy entrepreneurship is not just include 'selling ideas to policy-makers and selecting ideas by policy-makers' (Ackrill & Kay 2011, p. 78). It also includes institutional entrepreneurship in the policy-making when policy entrepreneurs steer the translation of these ideas to policy decisions and implementations to deliver institutional outcomes as promoters of policy ideas and elite-decision-makers. Thus, policy entrepreneurs function as institutional entrepreneurs by translating policies into formal rules and/or established practices when they are enabled by structural, institutional and agency-level complementarities (i.e. incentives reinforcing one another that inform the behaviour of agents) that prevail over contradictory incentives whilst operating in all stages of public policy-making process (Bakir, 2009a(Bakir, , 2013. Our key interest is to highlight linkages between policy entrepreneurship and institutional entrepreneurship activities. We present our overview in three sections. First, we offer a review of key agents in policy and institutional change, what motivates them, and when and how policy change happens. Second, we discuss macro-level context (i.e. the structural and institutional faces of context), and micro-level context (i.e. agency-level enabling conditions). Then we highlight how complementarities arising from these various contexts influence policy and institutional entrepreneurship that bring about policy and institutional change, with special reference to a structure, institution and actor-based (SIA) integrative framework (Bakir, 2013(Bakir, , 2017. Section four summarises the main argument (Bakir & Jarvis, in press).

Who are the key individual agents in policy change?
In the last three decades, there has been growing interest among political scientists and public policy scholars in understanding the role of human agency in policy change. Unsurprisingly, there are variants of definitions that relate to public sector entrepreneurship. Frohlich and Oppenheimer (1978), for example, offered one of the most common definitions of 'political entrepreneur' in political science: 'an individual who invests his own time or other resources [e.g. vision, skills, expertise, and leadership] to coordinate and combine factors of production to supply collective goods ' (1978, p. 68). In a similar vein, Kingdon (1995, p. 122) noted that the defining characteristic of 'policy entrepreneurs' is 'their willingness to invest their resources -time, energy, reputation, and sometimes money -in the hope of a future return' . They 'could be in or out of government, in elected or appointed positions, in interest groups or research organizations ' (1995, p. 122). They display 'social acuity [e.g. operating in networks], defining problems, building teams [e.g. promoting and maintaining advocacy coalitions], and leading by example' (Mintrom & Norman, 2009, p. 651). They sell ideas to powerful actors in politics and ensure their implementation; '[w]orking with those who have the formal power and resource control (typically referred to as policy champions)' , a policy entrepreneur 'seeks acceptance of the innovative idea in law or executive fiat, and the eventual implementation of the innovative idea into practice' (Roberts & King, 1991, p. 152; see also Schneider, Teske, & Mintrom, 1995, p. 3). In this widely held view, policy entrepreneurs are not decision-makers in public sector and their role is limited to influencing governmental agenda setting. Roberts and King (1991, p. 152) identify four types of 'public entrepreneurs' in terms of their behavioural patterns: political entrepreneurs, who hold elected leadership positions in government; executive entrepreneurs, who hold appointed leadership positions in government, bureaucratic entrepreneurs, who hold formal positions in government, although not leadership positions; and policy entrepreneurs, who work from outside the formal governmental system to introduce, translate, and implement innovative ideas into public sector practice. (1991, p. 152) Chwieroth adds another type: norm entrepreneurs 'who strategically promote greater acceptance of a particular practice or behaviour' for organisational change (2008, p. 483). For Chwieroth, their success depends on their discursive influence prevailing over alternative views, which is conditional upon their authority and/or formal 'position within an organization ' (2008, p. 493). Bakir offers a detailed analysis of institutional entrepreneurship and identifies institutional entrepreneur as another type of an individual actor in the public policy-making (Bakir, 2009a). Here an institutional entrepreneur 'refers to an individual who mobilises ideas, resolves conflicts, and steers their implementation for policy and institutional changes' during the various stages of public policy-making (Bakir, 2009a, p. 572). This perspective highlights multiple identities of an individual agency operating in different ideational realms to deliver policy and institutional outcomes. Here a policy entrepreneur can become an institutional entrepreneur when, enabled by various contexts. Institutional entrepreneur purposely initiates institutional change in all stages of public policy-making processes by building consensus and coalitions, and exercising power (Bakir, 2009a).

What motivates actors to engage in entrepreneurial activity?
To explain entrepreneurship in politics and policy, previous studies have focused more on the motivations of entrepreneurs to offer individual level explanations for their actions. They are considered to be rational individuals, motivated by a 'desire for power, prestige and popularity, the desire to influence policy, and other factors in addition to any money income derived from their political activities' (Bakir, 2003;Schneider et al., 1995, p. 11). In addition to instrumental rationality, however, these agents are also guided by cognitive rationality, namely their beliefs and ideas (Bakir, 2009a).

When and how does policy change happen?
The Multiple Streams Approach (MSA) originates in Kingdon's ([1984] 1995) ground breaking work on governmental agenda setting that established a framework of interpretation involving three streams of problems, policies and politics within the policy process. It is only when these disparate streams are brought together that policy change may occur (Kingdon, 1995, pp. 16-18, 172-179). This coupling results from policy entrepreneurs taking advantage of a brief opportunity (a 'policy window') 'to push for [a] solution or to focus attention on a certain problem' (Kingdon, 1995, pp. 179-184). Such 'windows of opportunity' may occur due to a crisis, shift in public mood or election of a new government. The role of policy entrepreneurs is different than norm (or ideational) entrepreneurs. Policy entrepreneurs are not only responsible for 'selling' a policy to key decision-makers, but must also ensure the convergence of disparate policy streams. Recently, there have been efforts to extend MSA by aligning it with policy cycles and advocacy coalition framework to other stages of public policy, including policy formation and implementation (Howlett, McConnell, & Perl, 2015, 2016, as well as showing its theoretical and empirical contributions to public policy theory (Cairney & Heikkila, 2014;Cairney & Jones, 2016).
Past literature discusses the agency of policy entrepreneurs with special reference to their strategies (e.g. 'engaging with others to clearly demonstrate the workability of a policy proposal' , 'demonstrat[ing] their trustworthiness and their commitment to their ideas for policy change') and their attributes (e.g. 'having deep knowledge of relevant procedures and the local norms that serve to define acceptable behaviour'), both of which increase the likelihood of a policy change (Mintrom & Norman, 2009, pp. 653, 656). There is, however, little research that examines how contextual influences constrain or enable policy entrepreneurship.

What is context in policy entrepreneurship scholarship?
It has been widely recognised that context matters in informing policy entrepreneurship.
Ironically, the early emphasis on the individual as change agent appears to have served as an inhibitor to theorization. In any given instance of policy change, it is usually possible to locate an individual or a small team that appears to have been a driving force for action. But in all such cases, the individuals, their motives, and their ways of acting will appear idiosyncratic. And idiosyncrasy does not offer propitious grounds for theorization. To break this theoretical impasse, policy entrepreneurship needed to be studied in a manner that paid attention simultaneously to contextual factors, to individual actions within those contexts, and to how context shaped such actions. (Mintrom & Norman, 2009, p. 651) When referring to context, scholars are mainly focused on the context of policy in policy formation and implementation rather than the macro contexts within which a policy entrepreneur is embedded. For example, '[k]ey sources of context include the "national mood, " interpreted by policymakers, and the policy conditions in each case, such as levels of congestion, fuel availability, and pollution when policymakers consider transport policy' (Cairney & Heikkila, 2014, p. 377; see also Kingdon, 1995;Mintrom & Vergari, 1998).
Moreover, agency-level micro contexts are conflated with structural and institutional contexts. For example, contexts that strongly influence the decisions and their translation into implementation by entrepreneurs are defined by 'the positions and relative power of decision makers in their relationships to one another' (Roberts & King, 1991, p. 152).
Here, this context is limited to 'a historically contingent context, including both formal and informal institutions ' (1991, p. 152). In a similar vein, others describe individual roles and resources as providing context to policy entrepreneurs: '[w]hile entrepreneurs have the capacity to choose behaviour, explanations of policy outcomes cannot be devoid of the context (institutions, roles, and resources) that regulates social interaction' (Zahariadis & Exadaktylos, 2016, pp. 62, 76, our emphases).

Discussion: how complementarities arising from macro and micro levels of context inform entrepreneurship
There are several weaknesses in the policy entrepreneurship literature summarised above (Bakir & Jarvis, in press). They can be grouped under two main categories. The first relates to the three major weaknesses of MSA. First, MSA is 'ahistorical' as it views policy streams independent of each other (Bakir, 2009a;Capano, 2009;Weir, 1992, pp. 191, 192). 'It does not focus on how previous policies affect current debates and institutional innovation, or institutional contexts that shape how, why, and where policy reforms take place' (Bakir, 2009a, p. 573; see also Schmidt, 2008b, p. 20). Second, the actions of policy entrepreneurs' are analysed at the national, regional  or systemic levels (Alimi, 2015). They are also considered key actors in global public policy and transnational administration (Stone & Ladi, 2015). However, they can also simultaneously operate at systemic, national, and micro-organisational levels, working towards macro-and micro-institutional reforms in internationalised policy domains (Bakir, 2009a). Third, their roles and actions are not always limited to articulating policy innovations, placing them onto governments' agendas or even policy formulation itself (2013, p. 574). They may operate not just in policy formulation, design, and/or agenda setting process but in various stages of public policy-making. In contrast to widely held view that policy entrepreneurs promote ideas and operate outside decision-making processes (Mintrom, 2000), policy entrepreneurs are endogenous to policy-making processes (see Ackrill & Kay, 2011;Bakir, 2009a).
The second category of weaknesses relates to context. If individual entrepreneurs in the public sector play a key role in policy and institutional change, we need to understand what factors enable entrepreneurship in macro-and micro-level contexts. We need theoretical lenses to further our understanding of how and why agents (e.g. entrepreneurs) embedded in their context are motivated and enabled to take actions that deliver policy and institutional change.
This article argues that explaining and exploring the interactions among various contexts and agents should be at the heart of research on policy and institutional entrepreneurship bringing about policy (institutional) change. Although there has been recent statements that '[t]he [future] research agenda endures: how to understand policy entrepreneurs' situations in the broader context' (Ackrill, Kay, & Zahariadis, 2014, p. 879), the above review demonstrates that the previous literature generally displayed weak theoretical and empirical engagement with the role of micro-and macro-level contexts in informing policy entrepreneurship.
There are three main problems in contextualising policy and institutional entrepreneurship. The first issue relates to macro-level context. Specifically, the concepts of structure and institution are combined or conflated in previous literatures. Thus, it fails to appreciate the analytical value in understanding their interactions with one another as well as with actors. Although context has taken on multiple meanings, at the macro-level we find it most helpful to consider it as composed of structures and institutions. Structures refer to the broader material and cultural contexts within which actors and institutions are embedded (Bakir, 2017, p. 226). Change in political regime, macroeconomic structure, and government are some examples of material structures, whilst policy and political traditions, socio-economic norms, and public philosophies are examples of cultural context. These structures both constrain and enable agency. We thus highlight the relevance of Gidden's concept of 'duality' to illustrate the interdependence among structure and agency. In structuration theory, '[s] tructure … is not to be conceptualized as a barrier to action, but as essentially involved in its production' (Giddens, 1979, p. 70; see also Berger & Luckmann, 1967). However in contrast to this view, which elides the distinction between structure and agency (Giddens, 1976(Giddens, , 1979, Archer (1995Archer ( , 2003 is right to call for paying 'attention to the analytical separation of structure and agency in order to appreciate and analyse their interactive relationships' (Bakir, 2013, p. 13). We also adopt this perspective.
Institutions refer to formal and/or informal rules that guide the behaviour of agents through the logic of appropriateness and/or the logic of instrumentality (Campbell, 2004;Campbell & Pedersen, 2001). Some examples of formal institutions include constitutions, laws, and regulations, whilst informal institutions include normative and/or cognitive ideas. However, institutions are embedded in material and/or cultural structures; they are defined as 'a relatively stable collection of rules and practices, embedded in structures of resources that make action possible (…) and structures of meaning that explain and justify behaviour -roles, identities and belongings, common purposes, and causal and normative beliefs' (March & Olsen, 2008, p. 691, emphasis in the original).
The second main weakness relates to micro-level context (i.e. agency-level enabling conditions) that informs the decisions and actions of entrepreneurs. The social position of the entrepreneur needs to be further elaborated. Individuals may occupy multiple social positions with multiple identities in various stages of domestic policy-making processes. These multiple identities, for example, include multiple hats, such as decision-maker, academic, framer, and broker, which enable policy entrepreneurs to operate in different ideational realms, including programmes, paradigms, and discourse, building coalitions, generating consensus, and resolving conflicts in domestic political economic struggles. At individual level, it is these multiple identities that empower policy entrepreneurs in public policy-making (Bakir, 2009a). When such individual level complementarities are reinforced by structural (e.g. a socio-economic or political crisis) and institutional (e.g. institutional failures) incentives that prevailed over contradictory ones, policy entrepreneurs become institutional entrepreneurs steering various domestic and transnational policy networks and advocacy coalitions towards institutional change (see, for example, Bakir, 2003Bakir, , 2009aBakir, , 2013. As Bakir (2013) illustrated earlier, organisational institutionalists may offer a helping hand to public policy scholars in their quests for identifying the role of context in institutional entrepreneurship: 'institutional entrepreneurship represents the activities of actors who have an interest in particular institutional arrangements and who leverage resources to create new institutions or to transform existing ones' (Maguire, Hardy, & Lawrence, 2004, p. 657;see also DiMaggio, 1988:14;Rao, Morrill, & Zald, 2000, p. 240). 'How can actors change institutions if their actions, intentions, and rationality are all conditioned by the very institution they wish to change?' (Holm, 1995, p. 398). Seo & Creed (2002, p. 226) refer to this as the 'paradox of embedded agency' . The theory of institutional work in organisational institutionalism aims to address this paradox (Battilana & D' Aunno, 2009;Battilana, Leca, & Boxenbaum, 2009). Organisational institutionalists refer to 'enabling conditions' at the field, organisational, and individual levels to identify those conditions which allow for institutional entrepreneurship and resolution of the paradox of embedded agency (Battilana, Leca & Boxenbaum 2009, p. 67). Specifically, 'there are three main enabling conditions for the agency of actors that have been identified in organisational institutionalism: field-level (i.e. industry/sectoral-level) enabling conditions (i.e. a 'jolt or crisis that precipitates action and a degree of heterogeneity and incomplete institutionalisation of practices, values, and norms'); organisational-level enabling conditions (i.e. a 'position in the organisational field or, more broadly, in the institutional environment'); and individual-level enabling conditions (i.e. the actors' social position and existence of institutional entrepreneurs 'who adopt leadership roles in episodes of institution building') (Bakir, 2013, p. 14. see also Battilana & D' Aunno, 2009, pp. 37-42). Like organisations, human actors occupying multiple social positions play a key role in institutional entrepreneurship 'because it might affect both actors' perception of a field … and their access to the resources needed to engage in institutional entrepreneurship …' (Battilana, Leca & Boxenbaum, 2009, p. 76; see also Bourdieu, 1977 and Lawrence, 1999 cited in Battlina, Leca and Boxenbaum).
The third main weakness is that there is little research that examines how interactions within and across structural, institutional, organisational and individual levels create complementarities that enable or constrain institutional entrepreneurship (Bakir, 2013; see also Bakir, 2017). In a similar vein, organisational institutionalists recognise, 'organizations' and individuals' social positions influence actors' likelihood of engaging in institutional entrepreneurship not only independently, but also jointly, through interaction' and 'there is a need to further explore' this interaction (Battilana, Leca & Boxenbaum, 2009, p. 78). If there are various contextual influences that inform actor behaviour, then we need to understand how their interactions affect an agent's actions. Institutional complementarity 'stems from the interactions between the influences that different institutions have on agents' behaviour' (Cited in Bakir, 2013, p. 14;Amable, Ernst, & Palombarini, 2005, p. 313;Campbell, 2011;Hall & Soskice, 2001). Complementarity based on reinforcement 'occurs when various national institutions mutually reinforce similar incentives that affect agents' behaviour' (Bakir, 2013, p. 14;Crouch, 2010, pp. 121, 122). This complementarity concept has been adapted to structural complementarity, highlighting a similar influence of structures on agential actions (Bakir, 2013(Bakir, , 2017. The various sets of structural and institutional influences, or complementarities, inform agential actions by reinforcing similar incentives.
In addition to under-appreciation of the complementarities arising from macro-level contexts that reinforce entrepreneurship, there is also a need for a fuller understanding of the similar effects of micro-level complementarities. Structural and institutional influences cannot account for the actions of agents completely independent of agents' resources and capabilities. However, such resources and capabilities are not limited to institutional entrepreneurs' social positions where political economic struggles take place. Indeed, these agents translate policy ideas into institutional outcomes, due in part to their own characteristics and skills. This view also recognises that institutional entrepreneurs have multiple social positions with corresponding identities.
They may play multiple roles in the various stages of domestic policymaking process, such as facilitating legislative adoption and executive/bureaucratic implementation of policy ideas. A policy entrepreneur, who mobilizes ideas and utilizes discourse toward policy and institutional changes, may simultaneously perform various roles, such as decision maker, academic, framer, and mediator that enable him to operate in different ideational realms such as programs, paradigms, and discourse. (Bakir, 2009a, p. 574) Indeed, as Campbell (2004, p. 100) noted, 'ideas do not emerge spontaneously or become important without actors, and so it is important to situate these actors and theorize their roles vis-à-vis ideas in any account of institutional change' . This view rightly calls to 'differentiate among types of ideas' [i.e. programs, paradigms, frames and public sentiments] that affect institutional change and 'different types of institutional entrepreneurs [i.e. decision makers, theorists, framers, and constituents] operating in different ideational realms ' (2004, p. 101, our emphasis).
Further, 'ideas provide linkages between structures, institutions, agents and political processes' (Bakir, 2013, p. 11; see also Béland & Cox, 2011;Blyth, 2012;Schmidt, 2008aSchmidt, , 2008bSchmidt, , 2011 through institutional work. Institutional entrepreneurs have several individual-level resources that enable them to operate in different ideational realms, which contribute to their power and leadership in purposeful entrepreneurship activity for institutional change. This mainly includes their multiple identities. There are four main 'enabling identities' (i.e. decision-maker, mediator, academic and framer) that reinforce entrepreneurship activity at the individual level in policy and institutional change processes (Bakir, 2009a, p. 574). These agents are decision makers as politicians, which empowers them to make policy decisions; they are mediators with multiple memberships, experience and a central position in domestic and international policy networks in an issue area that enables them to mediate ideas effectively within and among these networks to resolve conflicts and engineer a compromised solution in political economic struggles; they are academics with strong theoretical and intellectual backgrounds that give them knowledge authority; they are framers with strong coordinative and discursive skills to build and maintain broad coalitions within and outside the politics supporting the change. Institutional entrepreneurship activity is most likely when entrepreneurs have most of these identities that enable them to operate in different ideational realms. We argue that this view recognises entrepreneurs not only 'within a set of social relationships and institutions' (Aldrich, 1999, pp. 81-85;Campbell, 2004, p. 74) but also incorporates their individual qualities. It is these identities that influence the success of institutional entrepreneurship at the micro level. These multiple identities as agency-level enabling conditions influence entrepreneurship activity in distinctive but interrelated and complementary ways in a dynamic institutionalisation process. This view contrasts with the identification of various agents such as entrepreneurs, brokers and framers with distinct roles in the policy change process (Campbell, 2004, p. 101;Christopoulos & Ingold, 2015;Peters, 1997;Svensson & Öberg, 2005).
If institutional entrepreneurship is a context-specific activity in public policy-making, then we need a process-oriented, context contingent, dynamic and integrative framework joining macro, meso and micro level influences that enhance or constrain entrepreneurial action. The Structure, Institution and Agent-based (SIA) integrative framework (Bakir, 2013(Bakir, , 2017 may offer a step forward to identify and understand policy and institutional entrepreneurship as a single venture that necessitates consideration of multiple layers and integrated levels of context relevant analysis. The eclectic SIA framework, based on transparent and rigorous inductive qualitative research, suggests that complementarities arising from various contexts that reinforce similar incentives and prevail over contradictory ones for agency behaviour help us to think more rigorously about multiple levels of analysis and causal mechanisms behind the policy and institutional outcomes (Bakir, 2013(Bakir, , 2017. It conceptualises causal mechanisms linking cause and effect as complex and dynamic processes of interactions within and between the structural, institutional, and actor-levels that reinforce similar incentives for agents' decisions and actions. Thus, understanding various complementarities that consistently reinforce and encourage certain actions is central to causal mechanisms for explaining why and how institutional entrepreneurs act as they do to deliver collective outcomes. This framework does not rank the relative importance of various complementarities. They manifest themselves at least to some degree in influencing actions that generate outcomes. Some complementarities will be stronger than others as their nature and impact are time and context specific.

Selected empirical examples
Extensive macroeconomic and microeconomic policy and institutional reforms in Turkey following the 2001 domestic economic crisis illustrate the importance of complementarities arising from interactions among structures, institutions and agents reinforcing policy and institutional entrepreneurship (Bakir, 2009a). Structural complementarities reinforcing entrepreneurship came with the 2001 financial and economic crises in Turkey that coincided with a crisis of legitimacy in Turkish politics spurred by strong public distrust and anger and the ineffectiveness of a weak coalition government, which together opened a window of opportunity for entrepreneurship. At the institutional level, the then existing formal and informal rules and norms were becoming outdated in dealing with Turkey's macroeconomic problems. Against these macro contexts, Kemal Dervis, then-vice president for poverty reduction and economic management at the World Bank, was appointed as a new (unelected) Minister for the Treasury and Economic Affairs. Dervis had strong social capital as a member of a transnational policy network, which elevated him to the strategic position holding a key role and power in domestic policy-making via enabling his to access political and financial support of these intergovernmental international organisations. In addition, he held a formal hierarchical position in the key ministry and steering bureaucracy at the national level. Dervis also represented 'a powerful agency with multiple identities that enabled him to operate in different ideational realms' as a decision-maker (e.g. minister), theorist (e.g. academic at Princeton University and published articles regarded as influential in economics), framer (e.g. spin doctor publicising favourable interpretations of some neoliberal ideas with effective use of communicative and coordinative discourses), as well as mediator (e.g. 'mediating various ideas and utilising discourse within and among domestic and transnational policy communities in a punctuated institutional equilibrium') (Bakir, 2009a, p. 587). These complementarities arising from the macro, meso and micro contextual environments reinforced policy and institutional entrepreneurship activity in the various stages of public policy-making processes by empowering Dervis with legitimacy, power and resources towards policy and institutional change during the various stages of public policy-making in 2001.
Additional empirical evidence illustrating these interactions comes from Australia, which made a radical decision to adopt a new financial regulatory model based on the Twin Peaks idea in 1997 (Bakir, 2003(Bakir, , 2009b(Bakir, , 2013. Australia was the first country to adopt the 'Twin Peaks' model of financial supervision and regulation. Shifting from an entity-based focus of supervisory and regulatory arrangements that concentrated on the legal status of corporations, Australia instead adopted a functionally based system. On 1 July 1998, the 10 supervisory positions covering the federal, state and territorial levels were abolished in favour of 2 new financial regulators that focused on prudential regulation (Australian Prudential Authority, APRA) and disclosure regulation (the Australian Securities and Investment Commission, ASIC). In addition to this, the bank supervision powers of the Reserve Bank of Australia were revoked. These institutional and organisational changes were revolutionary in terms of scope (i.e. breaking away from the existing entity-specific orientation in favour of function-specific regulations and bureaucracies) and pace (i.e. occurring rapidly and impacting all parts of the regulatory and supervisory apparatus simultaneously) (Bakir, 2013, p. 111).
It is puzzling that Australia adopted the 'Twin Peaks' model despite the lack of pressure for change by financial firms and regulators, absence of market or regulatory failure, non-existence of international normative, coercive or mimetic pressure nor crisis necessitating such adaptations. In the case of Australia, the institutional and policy changes of Twin Peaks were not a response to external shocks from a crisis, but rather to a new idea that was translated into policy via entrepreneurship of the Treasurer, enabled by complementarities at the structural, institutional and agency level (Bakir, 2013, pp. 116-134; see also Bakir, 2003Bakir, , 2009b. Systemic structural complementarities included financial globalisation processes and financial technical and technological innovations, which resulted in greater financial conglomeration and blurred 'distinctions between firms, sectors and products in the financial services industry (Bakir, 2013, p. 166). This, in turn, reinforced 'product-based rather than entity-based prudential regulation/supervision' (Bakir, 2013, p. 36). Domestically, the main reinforcing structural complementarity was the election of the Liberal Party, which won its first federal election on 2 March, 119, after 13 years of rule by the Australian Labor Party. It reinforced the drive for change as the Liberal Party government desired the establishment of a financial system inquiry (FSI) to review the Australian financial system (later known as the Wallis Inquiry); the previous Labor government, however, had indicated that no such review was under consideration (Bakir, 2005). Thus, this election result opened a 'window of opportunity' , or a critical juncture for policy and institutional change. Moreover, at the informal institutional level, the 'Twin Peaks' idea was aligned with the bureaucratic preferences of the Treasury. A 'Treasury consensus' for institutional change was the result of the Treasurer adopting the Treasury's regulatory agenda. As a policy entrepreneur, the Treasurer coupled problems, solutions and political opportunities. While the Treasury provided support in shaping the governmental agenda on this issue, the support of the government would be required to translate it into a public agenda item. The Wallis Committee was established by the government in 1996 with the goal of reviewing the financial system. Packed with sympathetic members by the government and guided by the 'Treasury consensus' , the Committee 'played a pivotal' role in setting the public agenda and building consensus within and outside parliament towards policy and institutional change in prudential regulation and supervision (Bakir, 2013, p. 58). The Treasurer at the agency level was enabled by the Treasury and the Wallis Committee, which facilitated translation of governmental agenda to the public agenda. Committee members' technical knowledge, authority and discursive skills in their responses to critiques of this institutional change in financial regulation enabled the promotion of the Treasury Consensus in the political and public arena by apparently neutral sources. The Wallis Inquiry was thus used 'to generate industry and political support for the 'Twin Peaks' idea' by the government (Bakir, 2013, p. 130). In the micro level context, this allowed the building of a network of alliances within and outside parliament. As this case demonstrates, the environment for institutional entrepreneurship is most favourable when complementarities arising from structures, institutions and agents mutually reinforce similar incentives impacting entrepreneurial actions, bringing about institutional change.

Conclusion
If, as structuration theory (Giddens, 1976(Giddens, , 1979 and institutional theory (Campbell, 2004;Campbell & Pedersen, 2001) show, the agential authority of agents is shaped by the structural and institutional contexts in which they are embedded, two central challenges for research on policy and institutional entrepreneurship emerge: how and why these contexts inform actions, and how entrepreneurs can initiate institutional change if their decisions and actions are 'conditioned by the very institution they wish to change' (i.e. the paradox of embedded action) (Bakir, 2013, p. 14). Our purpose in this article has been to provide an overview of interactions between various contexts and actor agency and how public sector entrepreneurship research can be better 'contextualized' to address these two main challenges.
Institutional entrepreneurs, like any other individual agency, both shape their context and are shaped by it. Thus, to enhance the research on policy entrepreneurship and institutional change, there is a need for researchers to consider a 'contextual view of policy entrepreneurship' . Not only should public policy scholars offer contextualisation (i.e. thick and detailed descriptions of actions within contexts), but also, they should move forward to theory construction and testing that relate to micro, meso and macro-level interactions that inform institutional entrepreneurship. Rigorous, transparent and inductive qualitative research design (for an application of Gioia methodology in political science and public policy, see Bakir, 2017), coupled with comprehensive understanding of the dynamic and complex interactions between various contexts and entrepreneurship in the public sector, are likely to remedy shortcomings repeatedly attributed to public policy and institutional theory: (1) their limited analysis of the agency of agents embedded in various contexts in the institutionalisation processes; and (2) when, how and why such interactions generate causal mechanisms linking actions to socio-economic outcomes. Such a multilevel, process-oriented investigation enables simultaneous consideration of the impact of interactions among structures, institutions and agents on policy and institutional entrepreneurship.
We recognise that the SIA framework is just one approach for contextualising entrepreneurship research in institutional change processes. Researchers should also examine alternative approaches and reap the benefits of interdisciplinary knowledge and integrative frameworks. They should also consider integrating inductive qualitative and/or mixed methods approaches that recognise the value of comparative perspectives to include contextual influences in their future theory development. In doing so, we will be able to bring ideas, power and political economy to institutionalisation processes through additional insights linking entrepreneurship and institutional change with context.

Disclosure statement
No potential conflict of interest was reported by the authors.

Notes on contributors
Caner Bakir is an associate professor of Political Science, with a special focus on International and Comparative Political Economy, and Public Policy and Administration at Koc University, Istanbul. He is a Co-director of Center for Globalisation, Peace and Democratic Governance (GLODEM) and visiting professor at University of Strathclyde. His work relates to political economy and public policy with special emphasis on comparative institutional analysis and policy change. He has published articles in leading journals such as Policy Sciences, Governance, Public Administration, Development and Change, and New Political Economy.