Draft not to be cited. RC30 Comparative Public Policy: Panel on Policy Entrepreneurs and Governance: New Perspectives, 9 July 2012, Madrid

Transcription

1 Jan Seifert PhD candidate Lee Kuan Yew School of Public Policy (LKYSPP) National University of Singapore (NUS) Draft not to be cited Paper submitted to the 22 nd IPSA World Congress of Political Science 2012 RC30 Comparative Public Policy: Panel on Policy Entrepreneurs and Governance: New Perspectives, 9 July 2012, Madrid Policy change and policy entrepreneurship 1. Introduction A popular theme in the evolution of the policy sciences is the interest in policy change: What is it that explains policy u-turns, policy reforms and policy change? Surely, no one theory is sufficient to explain every single phenomenon. But while there is no one common mechanism to all stories of policy change, the debates in the literature have, at different times, focused, to varying degree, on factors such as institutions, agents, paradigms, ideas, groups, networks, external events or history. Since the early 1990s three frameworks have gained most attraction for our conceptualization and understanding of policy change: multiple streams (MS) (Kingdon 1984), the advocacy coalition framework (ACF) (Sabatier 1988; Sabatier and Jenkins-Smith 1993) and the punctuated equilibrium theory (PET) (Baumgartner and Jones 1991, 1993, 2009). Interestingly, none of these three pre-dominant frameworks for understanding policy change sees a key role in the behavior or action of (individual) agents. Even though the three frameworks are helpful to explain many cases of policy change, it remains counter-intuitive why there should not be instances in which individual agents, 1

2 or policy entrepreneurs as they shall be called in this paper, play a decisive role in bringing about policy change - the process of change needs to be led by someone (Capano 2009, 3). Besides, it might well be possible that agents can also play a role within these three frameworks. This paper seeks to discuss the role of agents as drivers of policy change. It suggests that there are instances in which policy entrepreneurs may have a crucial function to bring about policy change. With this in mind the paper is set to address three objectives which structure the following discussion. Firstly, it will review how the mainstream frameworks explain policy change. Secondly, the existing literature and allusion to policy entrepreneurship within and beyond the dominant theories will be explored for evidence and characteristics of agent behavior. Thirdly, a theory of policy entrepreneurship towards explaining policy change will be sketched out. The research question guiding this paper is: Can policy entrepreneurship explain policy change? And if so, how? 2. Policy entrepreneurship as contingent or complement to policy change in the main theories It has been argued in the previous section that the three theories of Multiple Streams, ACF and PET suggest some role for policy entrepreneurs in the process of policy change. This section will critically evaluate how the role of policy entrepreneurs is crucial to change in the causal framework of each theory and it will be argued that an absence of entrepreneurial agents would most likely not lead to change i.e. making the presence of entrepreneurs a necessary condition for bringing about change within each of the frameworks. This is a strong claim because it would challenge the broader and often more structuralist perspective of the three approaches while emphasizing that without the active involvement of agents, policy change is unlikely under any serious framework. 2

3 a. Policy entrepreneurs in Multiple Streams (MS) Policy change according to MS follows the opening of policy windows, which in turn is affected primarily by exogenous factors (special issues arising on the agenda or political and institutional changes). However, the theory does not suggest that there is any automatism between stream convergence and consequent policy change, or even stream convergence alone. It might well happen that the streams converge, i.e. a problem comes to the agenda, solutions and policies are suggested, there are potential majorities for these but they are either not adopted or not implemented later on. This suggests that even if conditions are right (i.e. the streams converge and the policy window is open), someone needs to take decisions accordingly or facilitate and push for decisions to be taken. This is where agents play their role. Agents and specifically the policy entrepreneur as envisaged by Kingdon can firstly facilitate the convergence of the streams through entrepreneurial behavior by prompting important people to pay attention, [ ] coupling solutions to problems and [ ] coupling both problems and solutions to politics (Kingdon 1995, 20). This would be agent behavior in the stricter sense and it may allow for a missing part in the causal chain of this argument. Policy entrepreneurs appear as facilitators but not as conditional in his model. However, it appears intuitive to think that in many cases the action of individuals is essential in helping the convergence of at least one stream towards the others. Exogenous factors can change decision-makers and they can alleviate topics and problems on the agenda. If more than incremental solutions are sought (and this is the interest of Kingdon s work), this requires some sense of leadership or simply boldness to bring to decision what everyone might be expecting anyways. It is here where agents are likely to play a decisive role. Only few political decisions (incl. entitlements or initial disaster and emergency response) follow quasi-automatic decision streams. In other cases like the impact of the Japanese Fukushima nuclear accident, very similar policy windows for policy change have opened in very similar societies like those of Western Europe. While controlling for the similarity in broad economic background, party affiliation and the same exogenous event, some governments chose to reverse their countries nuclear policy (e.g. 3

4 Germany, Switzerland) while others have not (e.g. France, Poland, UK). This is one exemplary suggestion for the potential role of policy entrepreneurship or leadership to explain why a convergence of streams is not enough for policy change. All this shows that the Multiple Streams model is somewhat coherent when it comes to explaining the macro-factors of policy change. There might also be cases where policies are changed more or less automatically or very consequentially following shifts evoked by exogenous factors. However, the theory falls short of theorizing sufficiently about the causal links on the micro-level about who does it and how change is brought about. It appears unlikely that, firstly, the streams are made to converge and secondly decisions in favor of policy change are actually taken without active agent involvement. b. Policy entrepreneurs in the Advocacy Coalition Framework (ACF) Of the theories discussed in more detail here the ACF is the most accommodating of agents. From its outset it bundles groups of agents with the same belief system into coalitions in their respective policy subsystems: actors can be aggregated into a number of advocacy coalitions (Sabatier and Jenkins-Smith 1993, 18). Competing coalitions are mediated by policy brokers in order to reduce intense conflict (ibid, 18f) and find solutions. The ACF proponents see these brokers as neutral or mediators; they appear in contrast to self-interested actors such as those in Riker s (1962) coalition theories or more recently Tsebelis (2002) veto-players. Typical brokers would be e.g. some elected officials or high civil servants (Sabatier and Jenkins-Smith 1993, 27). Despite accounting for agents as aggregated in coalitions the ACF does not offer more specifics or theorizing about the role and features of brokers and more entrepreneurial agents. A first critical remark should be addressed at the supposedly neutral role of the policy broker. While it might be that some brokers can mediate in problem-solving 1, it seems almost naïve to think that brokers would not pursue their own agenda. One striking 1 Such brokers (or mediators) are increasingly sought for high-profile public conflicts e.g. over infrastructure projects in Germany or elsewhere. 4

5 example might be EU decision-making. While formally neutral and acting as the guardian of the (EU) treaties, the European Commission is unlikely to be a truly neutral arbiter between national governments or the European Parliament and Council. Practically any decision will somehow touch personal or organizational interest and it is hard to remain neutral for individuals when it comes to that as Niskanen (1971) suggested in his classic work. In that sense brokers might not be partisan activists but it is unlikely that under all circumstances they are fully neutral. This suggests that the line between allegedly neutral policy brokers and more activist and self-interest driven policy entrepreneurs is one of degree. The second critical remark goes beyond the policy broker and is directed at strategic coalition behavior. The ACF suggests that coalitions are formed and held together by a shared belief system of core values. Their composition might change and also the belief system might adjust over time. However, there is little suggestion in the initial theory about how coalitions are formed or how they are led. While it is conceivable that some groups coalesce around obvious interests or beliefs like religion or protected professions, a majority of coalitions might depend on strategic leadership to survive and succeed over the long-run. The ACF says surprisingly little about strategic behavior of agents who advance the agenda of their coalition in particular in cases where an emerging advocacy coalition challenges and replaces a dominant one. Is it likely that this happens without some sort of strategic behavior by (an) agent(s)? In their study of education policy reform in Michigan (USA) Mintrom and Vergari (1996) suggest that a policy entrepreneurship model is compatible with and can be complementary to the ACF. After analyzing reform from these two different narratives, they argue that it is a matter of perspective which determines the analytical role of policy entrepreneurship in relation to advocacy coalitions. The ACF is better to conceptualize long-term trends leading to policy change while they argue that policy entrepreneurship holds better explanations for the details of the process by which [ ] change actually happened (Mintrom and Vergari 1996, 425). Mintrom and Vergari make a strong point and valid criticism of the ACF here. The framework seems incomplete in explaining change comprehensively without stricter attention to the role agents play. While it is 5

6 possible to use the framework to explain the broad lines behind policy change and the main groups, the role strategic individuals play is certainly ignored. c. Policy entrepreneurs in Punctuated Equilibrium Theory (PET) Baumgartner and Jones propose the most explicitly strategic role for policy entrepreneurs but fall short of recognizing the centrality of these agents in policy change. For them the origin of change is (again) in exogenous factors and, if anything, agents might make use of externally induced instability within a subsystem to push their issue(s) on the agenda. Baumgartner and Jones (2009) suggest that policy entrepreneurs perform two primary functions. They take new ideas on the agenda (or prevent it) and they attempt to strategically use voting situations which are most favorable to their cause. Although this sounds rather crucial to the evolution of new policies, the focus of the theory really is on the broader patterns of change. In this sense the theory puts less emphasis on details of how externally-induced perturbations are translated into action and, instead, is more focused on how attention to issues translates into policy change and showing that such change is radical when it happens. Just like the MS and the ACF also the PET finds the origin of change primarily in exogenous factors. While many such instances surely enable a re-focus of attention, and large-scale crises surely evoke policy change (Birkland 2006), how much of an external event is crucial to provoke changes within a subsystem may well be a matter of degree. Besides large-scale disasters and economic shocks it is up to policy entrepreneurs to turn external events into focusing events and lead a change of ideas that ultimately allows for policy change. Law-and-order issues come to mind where nasty instances of child abuse, brutality or stigmatizable culprit backgrounds can skillfully be turned into justifications for any degree of toughening of legislation or not. But how high events are pushed on the agenda and how problems are tailored to solutions is a matter of choice and this leaves space for policy entrepreneurship. The theory might leave that development to chance, and in fact it acknowledges that it does not claim to predict that policy change ever needs to happen in any given policy 6

7 subsystem. Nonetheless, the theory falls short of emphasizing the causal role policy entrepreneurs might play in the lead up to any policy change. If there is no strategic action, even the most suitable (external) event may not be exploited to bring about policy change. d. Agent behavior a gap in the theories of policy change The review here has shown that the primary factor for policy change for these three theories is to be found in exogenous factors and not in agent behavior. However, all three theories account for some degree of entrepreneurial action. The MS and PET mention policy entrepreneurs explicitly and make some suggestions about their role and character in the policy process. The role is arguably more limited in the ACF where a comparable agent (the policy broker) is primarily presented as a mediator. One contribution of the three theories is what they propose in terms of characteristics of the policy entrepreneur. Kingdon (1995) suggests that entrepreneurs invest their personal resources and act strategically to offer solutions to problems, bringing these on the agenda, network with decision-makers. Baumgartner and Jones (2009) see policy entrepreneurs as creators and promoters of new ideas and strategic manipulators of the decision-making process. Sabatier and Jenkins-Smith (1993) see a less strategic role for their policy brokers who primarily mediate between diverging interests. It has been argued above that such pro-forma mediating roles, that are often held by civil servants, can in cases be relatively important and turn brokers into entrepreneurs if they so wish. It remains surprising that these three theories, none of which accepts a decisive role for entrepreneurial action, are the predominant approaches for analyzing the policy process. No doubt particularly the ACF and PET have provoked a deep and substantial research body over the last years which often confirm and work within the frames of these theories. But it has also been argued in this section that each theory appears to have a missing causal link by not including a clear role for agent action in their theoretical outline. Change might be theoretically possible without agent involvement but it is highly unlikely that such patterns are predominant within each of the frameworks 7

8 over time and across policies. To the contrary, entrepreneurs should be assumed to play a central and strategic role in bringing the streams together, leading and mediating coalitions as well as in translating events into attention and agenda change. Without accounting for these possibilities, the three theories fall short of giving the full picture of policy change. Each of them leaves space for an alternative or complementary (Mintrom and Vergari 1996) approach that puts the policy entrepreneur as the central link into a causal chain of change. 3. Policy entrepreneurship as an approach in the broader literature A good number of studies (Mintrom and Vergari 1996; Arnold 1989; Mintrom 1997; Moon 1995; Thompson 1994; Christopoulos 2006; Crow 2010; Mackenzie 2004) have used policy entrepreneurship approaches to analyze policy processes and policy change. Interestingly, they most often appear to suggest that policy entrepreneurship is a well-developed framework and there is often little critical review of its theoretical underpinnings. These studies (most often case studies) refer to the notions brought forward by Kingdon as well as Baumgartner and Jones but authors seem to lack a critical understanding of the propositions in MS or PET. As has been argued above, on the one hand the proponents of these two theories lack a coherent theoretical approach to the causal role of agents in policy change. They also fall short of proposing a detailed set of features or hypotheses to be applied and used for studying policy entrepreneurship. This leads to a situation in which some of the authors of specific studies risk using the notion of policy entrepreneurship just as it pleases the story of their case at hand. Their story lacks a developed theoretical framework or testable hypotheses. Alternative frameworks are not used nor even considered in their discussions. However, case studies of policy entrepreneurship may hold for some suggestions with regards to the environment, characteristics and causal logic in which policy entrepreneurship happens. Some of these suggestions shall be reviewed in the 8

9 following to establish their contribution to a more genuine theory of policy entrepreneurship. It should be noted that there is a diversity of terms being applied to describe entrepreneurial behavior by agents. While change agents are often referred to as policy entrepreneurs (the notion applied in this paper), their action is also discussed under terms such as leadership (Capano 2009; Moon 1995), political entrepreneurship (Simmons, Yonk, and Thomas 2011), institutional entrepreneurship (North 1990), public entrepreneurs (Roberts and King 1991; DeLeon 1996) or policy brokers (Sabatier 1988; Sabatier and Jenkins-Smith 1993). As long as the action analyzed is concerned with policy change and based on agent-centered action, relevant findings will be included for the purpose of this work. a. Background and characteristics of policy entrepreneurs Studies referring to policy entrepreneurship analyze the behavior or different kinds of actors from local activists up to prime ministers. Roberts and Kings (1991, 152) suggest four types of personal backgrounds of entrepreneurs: elected or appointed government leaders, or governmental non-leadership or government-outsiders. Unfortunately, they do not make clear if government in their understanding includes all parts of the system of government (incl. legislative or administrative jobs) or strictly executive positions. Crow (2010), in contrast, classifies entrepreneurs as citizens, experts and those elected. She is also the only one comparing the distinct features of entrepreneurs by their positional background, inter alia, in view of their democratic legitimacy (see further below). What stands out from the case studies applying some sort of policy entrepreneurship approach is that most of these agents do indeed have a background or function somewhere in the government (executive or senior administrative level) or legislatures. This might be little surprising given the access and opportunities actors at these levels have to influence politics. It also shows that future research might want to address the challenges or experiences of e.g. expert and citizen entrepreneurs. 9

10 In describing or suggesting the characteristics of policy entrepreneurs, researchers (Simmons, Yonk, and Thomas 2011; Mackenzie 2004; Thompson 1994; Christopoulos 2006; Mintrom and Norman 2009) have come up with an exhausting (idealized) catalogue of skills. They can possibly be classified around five groups, namely a) drive and commitment, being able to b) seek opportunities and take risks, c) rely on powerful networks and bring the right people together, d) argue and (even) manipulate, and e) find innovative strategies and solutions. Naturally, the policy entrepreneur would be a top performer on most or all of these skills. Unlike many senior policy-makers entrepreneurs should distinguish themselves through forceful application of these skills in the operation of policy change. In the research available there is unfortunately no comparison to non-entrepreneurial actors. It might well be possible that at least those in senior positions have a more balanced set of skills around those described here and further research would have to work out how entrepreneurs particularly from within the decision-making system distinguish themselves from their peers. b. Strategies and actions applied by entrepreneurs The skill-set successful entrepreneurs are supposed to possess suggests that the actions and strategies employed by them are directly related. Not surprisingly, a number of authors suggest a set of actions and strategies which mostly mirror the skills mentioned above but are also more concrete in terms of operationalization. For Mintrom and Vergari (1996) the setting is straight-forward: all entrepreneurs perform three functions [ ][:] they discover unfulfilled needs and suggest innovative means to satisfy them [ ][,] bear the reputational [ ] risks [ ] [and], they serve to resolve collective action problems (Mintrom and Vergari 1996, 422). In his review of the literature Capano (2009) finds that what entrepreneurs are doing is identifying problems and finding solutions [ ][,] submitting new ideas to [ ] actors, [ ] mobilizing public opinion [ ] [,] solving the emergent problem of collective coordination [, and] find new venues for policy-making (Capano 2009, 4). In other words policy entrepreneurs are the ultimate fixers in the policy-process. Roberts and Kings (1991) further argue that the role of entrepreneurs is particularly in the earlier stages of policy-making up until 10

11 implementation. This is a somewhat longer perspective than that suggested in the MS and PET frameworks which are primarily concerned with agenda-setting. Surely, entrepreneurs are relevant in agenda-setting and possibly decision-making but there is practically no research on entrepreneurship over implementation. So far the interest in the literature is less focused on something like obstructive entrepreneurship (e.g. by Lipsky s (1980) street-level bureaucrat) when it comes to individuals and their role in policy implementation. In a more Machiavellian (or game theoretical) perspective Simmons et al. (2011) propose four strategies employed by policy entrepreneurs. Firstly, he separates cost and benefit considerations so as to find ways to make others pay for his clients, voters, or cause s lunch. Secondly, by engaging in rentseeking he announces the intention [ ] to introduce legislation or regulation that will threaten someone s interests. He would also, thirdly, use a policy of symbolism: making grand policy statements whose objectives cannot be achieved and fourthly magnify threats to their [own] group(s) in order to keep [ ] group members involved and contributing to the cause (Simmons, Yonk, and Thomas 2011, 375f). Many of these suggestions are reflected in the findings or descriptions of the empirical literature. Mintrom (1997) as well as Mintrom and Vergari (1996) confirm that entrepreneurs work through ideas and attach them to the relevant problems. They exploit the view that a crisis is at hand (Mintrom and Vergari 1996, 425) and also do their part of coalition building. An important element with regard to the latter is to unify the forces behind the agent. Thompson found that giving a sense of purpose (Thompson 1994, 405) is an important management activity in this respect. Depending on circumstances a conscious low-profile strategy can achieve the desired results as Arnold (1989) finds. In his work on homeless support in the US under Reagan he argues that For the homeless programs to be adopted, they must be removed from public scrutiny (ibid, 61). Surprisingly, the protagonists have also refrained from grandstanding about their achievements, arguably in order not to receive a backlash and cuts in the funds they had appropriated for the measures. In line with the suggested symbolic actions Zott and Huy (2007) find in their study of business start-ups that entrepreneurs who engage in the performance of symbolic 11

12 actions are more successful. They group such actions into four categories: conveying the entrepreneur s personal credibility, professional organizing, organizational achievement, and the quality of stakeholder relationships (Zott and Huy 2007, 70). It appears intuitive that their findings are equally valid for the political context but this requires further study. Finally, Christopoulos (2006) looks into a possible difference between incremental and opportunistic entrepreneurialism. His study reveals evidence for a dominance of opportunistic entrepreneurial behavior among political actors with low levels of political capital. And he goes on to argue that political entrepreneurship is contingent upon actor status and authority (ibid, 772). In other words actors with low political capital can only hope to attain prominence by engaging in high-risk opportunistic actions; while actors with high political capital can be more circumspect and invest their more extensive resources in low-risk incremental ventures (ibid, 773). This is an interesting finding for two reasons. Firstly, it acknowledges that policy entrepreneurs might start off with sub-perfect levels of skills and yet achieve results. Secondly, the strategies they employ or researchers should expect need to be measured against agents personal background, skills and position. c. Contextual factors Many authors warn of a perception in which agency alone can account for and explain policy change ( the actions of the policy entrepreneur will not always be enough to ensure approval of policy innovations (Mintrom 1997, 765). Rather, contextual factors matter (Mackenzie 2004; Mintrom 1997; Mintrom and Norman 2009). Roberts and King (1991, 172) argue that even the perception of crisis is necessary. Whereas context obviously matters for entrepreneurial opportunities and strategies pursued, this poses a deductive challenge for policy entrepreneurship approaches towards the study of policy change. But the question of how much impact macro developments have shall be dealt with in the next section. There are important contextual variables to be considered even in the more immediate environment of the entrepreneur and different authors find diverging ways to classify them. Moon (1995) argues that the innovation policy entrepreneurs aim to achieve is constraint by three kinds of factors social (public 12

13 opinion and lobby work), governmental as well as economic and secular (economic situation and societal trends) ones. Christopoulos distinguishes between [p]ersonal and positional attributes as well as structural and institutional constraints (Christopoulos 2006, 758). Capano (2009) broadly agrees with this dichotomy. He sees the entrepreneur s position among his followers and the institutional or historical context of his action as the relevant background. MacKenzie (2004) structures his empirical work into two levels whereas the first looks at skills and strategies while the second includes two contextual variables, namely Institutions and Positional Power (ibid, 380). d. Drawbacks of entrepreneurship Policy change through agency may go hand in hand with a less consultative or inclusive policy process. Indeed its very nature and the strategies discussed above suggest that often entrepreneurial success comes at the expense of democratic discourse and deliberation. It has even been suggested (in the case of homeless program funding) that behind-the-scene dealing might be the only way to achieve an objective because broader public attention may kill an initiative. By definition most entrepreneurial activities challenge the status quo and in that way decisions or programs which have been previously adopted, are likely to follow due deliberation and process. Given that successful entrepreneurship much depends on the capabilities of agents, and their capabilities are likely to enhance as they find themselves closer to power, there appears to be an inherent status quo bias: entrepreneurs with the willingness to upset the status quo are likely to have disproportionately less resources than those potential entrepreneurs who are closer to power and therefore keener in defending the status quo or enhancing their influence. One motivation for policy entrepreneurship to start with is personal benefit. Simmons et al. rightly point out that agents can also extract rent for themselves (Simmons, Yonk, and Thomas 2011, 379) for good or for bad. Crow (2010) is therefore cautious about the moral story of agency and cautions that there can be a downside to policy entrepreneurship [because they could] [ ] abuse their power, misuse and misguide people and policies, and can succumb to ethical challenges (Crow 2010, 301). DeLeon is equally alert and warns that Entrepreneurship 13

14 is not the appropriate response to all social problems (DeLeon 1996, 508) even though political or institutional barriers to change continue to appear high. Another concern is that background may limit entrepreneurial action space. In her case study Crow (2010) finds that experts are more likely than citizens or elected officials to act as entrepreneurs. This might inhibit the democratic process because When experts influence policy change [ ], decisions are not necessarily transparent and open to public debate (Crow 2010, 313). 4. Suggestions for a generic policy entrepreneur theory After Roberts and King complained in the early 1990s that there was no guiding substantive theory and a viable operational model (Roberts and King 1991, 153) to study policy entrepreneurship, Mintrom and Vergari claimed only a few years later that several scholars have begun to develop a policy entrepreneurship model (Mintrom and Vergari 1996, 422). Unfortunately, none of these attempts appears to have convinced so far. A few more case studies do not necessarily advance a substantive theory and Mintrom remains the only one who has studied the phenomenon of policy entrepreneurship more seriously. But even he has refrained from proposing a genuine theory based on agency. Instead he appears to prefer the study of policy entrepreneurship mingled with established theories of policy change. Together with Norman he suggests that the concept of policy entrepreneurship fits well with other explanations of policy change (Mintrom and Norman 2009, 650) like MS, ACF or PET. In such a complementary function it can serve to focus [on the] [ ] microlevel political activity (Mintrom and Vergari 1996, 431). It is regarded as useful for the study of relatively short time spans (ibid, 424). Over the previous sections it has been argued that neither the predominant theories of the policy process nor the more specific studies on policy entrepreneurship have come up with a coherent framework and testable theory of policy entrepreneurship. Michael Mintrom, who has published the most thorough works employing policy entrepreneurship concepts and reviewing the overall idea over its position in the 14

15 literature, merely suggests using an agency approach as complementary to other theories. With all this two questions remain. Firstly, can a focus on agency be sufficiently consistent to provide for a theory without taking structural or contextual matters as a precondition; and secondly, if that is the case or to test such a proposition, how could a genuine policy entrepreneurship theory look like? a. Policy entrepreneurship and context Both empirical and theoretical studies of policy entrepreneurship have pointed to the fact that agency can hardly be understood without the prevailing conditions surrounding it. No doubt it is true that entrepreneurs need enabling conditions to pursue their actions and eventually deliver. To what extent such externalities are a given is a matter of degree though. It is true that a single person possesses only limited capacity to change policies and programs and no matter how democratic a political system is (not), institutions limit individual action anywhere. Any theory of policy entrepreneurship therefore needs to account for context. One way to deal with this is to follow Mintrom s suggestion and accept the macro-level developments as suggested in the ACF, MS or PET as the dominant forces leading to change. The study of policy entrepreneurship will remain a micro-level attachment but no necessary condition for change. It has been argued above that this is not a very convincing approach because agency is required to ultimately make change happen in many cases. Entrepreneurship is likely to be found in many cases of policy change. Where it has been pursued strategically and with measurable impact, it deserves more attention namely to be accepted as a causal part of change. Having the study of entrepreneurship as a complementary, but non-contingent, aspect of these theories, does not solve the gap in the causal chain of the predominant frameworks. Instead it makes much more sense to see context as a matter of degree. Supradevelopments like the convergence of streams, the evolution of advocacy coalitions and outbreaks of attention can have significant impacts on policy change. In particular when these are provoked by extreme external shocks, they are likely to permit significant 15

16 policy change. In such cases there might not be the need for much agency, because responses are more or less obvious. However, when there is an absence of significant exogenous influences (and already the determination of their severity is a matter of judgment), then there is scope to accept contextual impact as a matter of degree. - It needs to be taken into account, but it will hardly tell the whole story. When some indications of agency exist, then the study of policy change through a genuine policy entrepreneur approach becomes relevant (if not necessary). Taking context into consideration is sufficient but its impact on change would not need to be necessary. The literature surveyed in the previous section has made some useful suggestions for how and where context matters. The first level of context touches on the entrepreneur s skills and background. The second is concerned with the more immediate environment he operates in what are the institutions and social or economic conditions determining his risk and opportunity structures. A third level might concern the broader context, i.e. the general nature of the policy subsystem or the evolution of the streams. Studies of policy entrepreneurship become relevant when they can control for or minimize an impact of the third level. If that is not possible, respective phenomena would be better explained by other theories. A more elaborate theory would then propose and test hypotheses with regards to the more immediate contextual factors and their interplay with the personal background of the entrepreneur. One such hypothesis would be that resource-poor entrepreneurs take higher risks. Another hypothesis can be that expert or bureaucrat behavior is less inclusive or democratic than that pursued by citizen or elected entrepreneurs. More hypotheses here? A final concern touches on the acceptance of fixed constraints. Context matters but context is equally fluid. The very assumption of policy entrepreneurship is that they change policy and thereby context. In that sense context may seem to limit operational flexibility but the whole point about entrepreneurship is that, at some point, they will go against and surmount some of the external constraints in order to achieve their objectives. Therefore, the study of entrepreneurship is also an assessment of the relativity of context. If context is all about fixed constraints, then theories of policy 16

17 change end up finding the only cause of change exclusively in exogenous shocks - which is the key weakness of the predominant theories. b. Towards a policy entrepreneurship theory (PolEnThe) Given the absence of a genuine framework that puts policy entrepreneurship at its heart to explain the policy process and specifically agent-induced change, some rudimentary ideas for a new theory shall be outlined in this sub-section. Their structure will be based on Capano s (2009) critique of the major theories of policy change and his proposition for some theoretical choices for a framework to study policy change. Capano argues that five aspects need to be tied together to establish epistemological coherence in a (new) theory: the definition of policy development and change (what is the real object?); the type of change (incremental or radical?); the output of change (is it reversible or irreversible?); the level of abstraction and the structure/agency dilemma; the causal mechanisms, the explanatory variables, and the configurative dimensions (ibid, 13). On the way towards PolEnThe some indicative suggestions shall be made for each of these five elements. The object of change This first dimension is concerned with what policy change means. In the PolEnThe policy change would mean a deviation from the business-as-usual (BAU) path of a given policy or program either in its legal form or the way it is implemented i.e. its outcome as perceived by the subjects affected or involved. The type of change Capano s second dimension touches on the nature of change as well as its abstraction, speed, progression and scope (Capano 2009, 15). Policy entrepreneurship is a phenomenon that is less likely to be associated with incremental change. Cases of 17

18 agent-driven incremental change might exist but if or when they do, they might experience the file-drawer effect because they are less likely to be of interest for study. It also lies in the nature of human life and capacity that efforts to affect change and achieve it through individual agency are more likely to be found relevant in cases of more significant deviations. The studies surveyed in previous sections have shown that such instances do not need to involve only the full economic transformation of a country as in the case of Margaret Thatcher (Moon 1995). They might touch on the introduction of a bold new foreign-language scheme across a country like Australia (Mackenzie 2004) or the establishment of firstly modest programs to support the homeless in the US (Arnold 1989). It is also perceivable that particularly knowledge-based activities might eventually lead to paradigmatic changes as was arguably the case with Keynesianism, monetarism or the institutional turn in economics following the work of individual (or small groups of) academics. The level of abstraction is a very important point because it alludes to the phenomena surveyed in most of the empirical studies cited in this work. Often change is surely not affecting the macro perspective in the form of the country s overall economy or even national debate but has significant effects on a subsystem or the micro-level when it comes to specific parts of the population or stakeholders. While a macro perspective should not be excluded from its analysis, policy entrepreneurship is more likely to benefit the analysis of (rather radical) micro-level developments. The distinction between speed and progression in Capano s outline is not very clear (or relevant) with regards to policy entrepreneurship. What might be expected is that it takes a period of time over which the entrepreneur argues his case and builds his coalition for change. Once that is established, the act of policy change may happen very quickly. The entrepreneur will undoubtedly build his argument and support around a core of existing resources which he needs to expand to succeed. This makes his progress cumulative. Such a phase will be less speedy and might in fact be running over years in which the agent gathers support to wait for his chance. However, after some crucial moment the agent will have gathered sufficient support or will make use of externally induced opportunities and will have to act rather quickly to deliver his 18

19 objective. In that sense there is no logical sense of speed and the actual act of bringing about change might appear mostly like the punctuations suggested in the PET. Finally, the scope of change is as much a relative phenomenon as it is in the predominant frameworks. Given their limited capacity most entrepreneurs will act within the specific policy subsystem in which they are competent. Some initial changes may not receive any initial attention but they might set the path towards paradigmatic change some time - even decades - later. In many cases change should concern subsystemspecific issues or even specifics below and within a policy field. The output of change The fourth dimension is concerned with the reversibility of change. This dimension is not particularly critical for policy entrepreneurship models. In contrast to approaches like path dependency or evolutionary theories, agent-induced change may be reversed in the medium- or long-term but even if that happens, it would not affect the validity of PolEnThe. The abstraction level and structure vs agency With its agent focus the PolEnThe would clearly be reductionist in that it focuses on the likely micro processes around the activities of the entrepreneur. As has been argued before, context needs to be taken into account but this might not extend to all or the broader macro structures and in that sense reduces the scope of abstraction. With its clear focus on the policy entrepreneur the PolEnThe is clearly agency focused but this does not exclude some attention to structural considerations (in the form of context). The causal logic 19

20 This fifth dimension concerns the logical chain and this is where more in-depth research is most critical. Existing studies of policy entrepreneurship unfortunately lack clear logical propositions to form a resilient base. A big-picture perspective on policy entrepreneurship would take a linear-additive view of causality (Capano 2009, 17) with policy change as the dependent variable and the policy entrepreneur and his features as the independent variable(s). Context would need to be controlled for. Such a model would expose why the agent achieves change. A study with a more strategic interest might want to focus on how the or a (specific) entrepreneur brings about change. It would use factors like background, skills and strategies employed by the policy entrepreneur as the independent variables (section three and four provide some testable input). Context can be controlled for or if it is assumed to interact, it can form part of an even more complex combined theory in which context-dependent strategies and traits would be analyzed. It would make most sense to require at least the study of the initial mechanism to establish that policy entrepreneurship is indeed responsible for policy change. Strategic considerations might be exploited in a second step. Illustraton table with overview along Capano epistemology here PolEnThe reviewed A simple form of the PolEnThe could be summarized as follows. Policy entrepreneurs are individuals with a certain standing in a policy subsystem. They bring about policy change, i.e. a deviation from the BAU path by drawing on their personal background, employing adequate strategies and operating within a given context. Change is rather radical (than gradual) and is witnessed more on micro- than macro-level. Its progression might vary but it would often look like a gradual build-up over the medium-term which is ended with a punctuation of change (much like the PET). Such changes may in fact not be systematic in scope and often touch on subsystem-specific issues. Change may be reversed or enhanced over the medium term, just like policy entrepreneurs potentially challenging and changing any pre-existing condition. The focus of any such study is 20

21 mostly on the micro-level surrounding the entrepreneur and in that sense reductionist as it cannot take account of all contextual factors at the macro-level. Finally, the causal logic in its simplest form is linear-additive. It would put the policy entrepreneur and his features and strategies as the determining factor of policy change while controlling for context. More advanced models might compare the actions and success of different kinds of entrepreneurs or focus more on how they achieve policy change. 5. Conclusion This work has started by showing that none of the three major public policy frameworks accounts for a decisive role of agency in the process of change. Policy entrepreneurs, as metaphors for change agents, are referred to in some capacity within the Multiple Streams, Advocacy Coalition and Punctuated Equilibrium theories but all three fall short of proposing either a necessary role for their involvement in policy change or a broader set of characteristics to measure their action. In the meantime a growing body of literature has evolved which analyzes policy entrepreneurship often referring to some of the predominant frameworks as informing their approach of agency behavior. These predominantly empirical studies as well as the leading theories have been reviewed with regards to their propositions about how policy entrepreneurship affects policy change. Many empirical studies show convincingly that policy entrepreneurship has played a decisive role in bringing about policy change but these studies lack a clear theoretical framework to be measured against and rooted in. One of the possible reasons for a lack of such a theory is the relationship between agency and the environment (or structures) it occurs in. Key protagonists in the study of policy entrepreneurship like Michael Mintrom maintain that its application is contingent on the use of other theories to explain the broader issues or context - surrounding entrepreneurial action. However, the argument pursued in this paper is that context needs to be accounted for but it might well be incorporated into a genuine agency model in a reductionist fashion. This, then, leaves space for the evolution of a specific policy entrepreneurship theory (PolEnThe). A rough sketch of some of its features has 21

22 been outlined in the last section following the epistemological framework suggested by Capano (2009). The initial research question of this essay was if policy entrepreneurship can explain policy change. No ultimate answer to that question has been found but the argument put forward here is that there are very strong indications for such a possibility. One reason is that all existing frameworks ignore that factors of structural convergence still need to be tied and translated into policy change by actors within the system. - Policy change is most often not inevitable but needs to be realized through conscious action. Where this is the case, degrees of policy entrepreneurships should be detectable and deserve study. If policy entrepreneurs play a decisive role and how exactly they pursue their goals, however, leaves room for further research. A clear chain of logic to substantiate a genuine theory of policy change still needs to be established but the material discussed in this work suggest that there is ample foundation for such an undertaking. Some of the existing empirical studies have also shown great depth in explaining individual cases of agency-led change even in the absence of a broader theoretical framework. The findings from these works should now be used to answer the second part of the research question, namely how a PolEnThe might look like. For now, PolEnThe might enjoy the benefit of the doubt and it should at least be acknowledged, as has been persistently argued in this work, that existing theories are not comprehensive enough to fully explain policy change as long as they touch on agency only in passing. 22

23 6. Bibliography Arnold, Craig Anthony BEYOND SELF-INTEREST: POLICY ENTREPRENEURS AND AID TO THE HOMELESS. Policy Studies Journal 18 (1): Baumgartner, Frank R., and Bryan D. Jones Agenda Dynamics and Policy Subsystems. The Journal of Politics 53 (04): Baumgartner, Frank R., and Bryan D. Jones Agendas and instability in American politics. Chicago: University of Chicago Press. Baumgartner, Frank R., and Bryan D. Jones Agendas and instability in American politics. 2nd ed. Chicago: University of Chicago Press. Birkland, Thomas A Lessons of disaster : policy change after catastrophic events, American governance and public policy series. Washington, D.C.: Georgetown University Press. Capano, Giliberto Political change and Policy change: some notes on the role of leadership as a theoretical and empirical problem. In XXI IPSA World Congress of Political Science. Santiago de Chile. Capano, Giliberto Understanding Policy Change as an Epistemological and Theoretical Problem. Journal of Comparative Policy Analysis: Research and Practice 11 (1):7-31. Christopoulos, Dimitrios C Relational attributes of political entrepreneurs: a network perspective. Journal of European Public Policy 13 (5): Crow, Deserai Anderson Policy Entrepreneurs, Issue Experts, and Water Rights Policy Change in Colorado. Review of Policy Research 27 (3): DeLeon, Linda Ethics and Entrepreneurship. Policy Studies Journal 24 (3): Kingdon, John W Agendas, alternatives, and public policies. Boston: Little. Kingdon, John W Agendas, alternatives, and public policies. 2nd ed. New York: HarperCollins College Publishers. Lipsky, Michael Street-level bureaucracy : dilemmas of the individual in public services. New York: Russell Sage Foundation. Mackenzie, Chris Policy entrepreneurship in Australia: a conceptual review and application. Australian Journal of Political Science 39 (2): Mintrom, Michael Policy Entrepreneurs and the Diffusion of Innovation. American Journal of Political Science 41 (3): Mintrom, Michael, and Phillipa Norman Policy Entrepreneurship and Policy Change. Policy Studies Journal 37 (4): Mintrom, Michael, and Sandra Vergari Advocacy Coalitions, Policy Entrepreneurs, and Policy Change. Policy Studies Journal 24 (3): Moon, Jeremy Innovative Leadership and Policy Change: Lessons From Thatcher. Governance 8 (1):1-25. Niskanen, William A Bureaucracy and representative government. Chicago: Aldine. North, Douglass Cecil Institutions, institutional change, and economic performance, Political economy of institutions and decisions. Cambridge ; New York: : Cambridge University Press. Riker, William H The theory of political coalitions. New Haven:, Yale University Press. Roberts, Nancy C., and Paula J. King Policy Entrepreneurs: Their Activity Structure and Function in the Policy Process. Journal of Public Administration Research and Theory 1 (2): Sabatier, Paul A An advocacy coalition framework of policy change and the role of policy-oriented learning therein. Policy Sciences 21 (2):