Policy Change in Comparative Contexts: Applying the Advocacy Coalition Framework Outside of Western Europe and North America

Abstract

The advocacy coalition framework (ACF) is one of the most frequently applied theories of the policy process. Most applications have been in Western Europe and North America. This article provides an overview of the ACF, summarizes existing applications outside of Western Europe and North America, and introduces the special issue that features applications of the ACF in the Philippines, China, India, and Kenya. This article concludes with an argument for the continued application of the ACF outside of Western Europe and North America and a research agenda for overcoming challenges in using the ACF in comparative public policy research.

Introduction

Comparative public policy research has faced numerous challenges and critiques over time. These challenges stem, in part, from the difficulty of developing common conceptualizations and the intractability of conducting studies across time and space (McDougal 1952; Feldman et al. 1978; Capano 2009; Capano and Howlett 2009; Howlett and Cashore 2009; Gupta 2012). For this reason, developing better theories of policy change is an important goal in comparative public policy research. The goal must be to understand why certain political choices or institutions emerge in some settings versus others, and to use theories that are generalizable yet flexible enough to account for differences across national or regional contexts.

One pathway to better theories of comparative policy change is to promote an effort among scholars to pursue questions involving policy processes not as independent researchers, but as a community of researchers operating within a research program (Laudan 1977). The advocacy coalition framework (ACF) exemplifies one such research program (Sabatier 1988; Sabatier and Jenkins-Smith 1993). The ACF is one theoretical framework, among many, that simplifies the great complexity of social and political systems with the goal of moving towards a positive understanding of the various drivers of policy change at multiple scales (Sabatier 1999).

In terms of a generalizable theory of policy change, the ACF has shown merit in its portability across different contexts, particularly across North America and Western Europe (Weible et al. 2009; Jenkins-Smith et al. 2014). However, the ACF research program includes relatively few applications outside of North America and Western Europe. The purpose of this special issue in the Journal of Comparative Policy Analysis is to explore the generalizability of the ACF as well as its capacity to contribute to the literature of comparative policy analysis in helping to describe and explain phenomena across vastly different political-institutional contexts.

A Synopsis of the Advocacy Coalition Framework

The ACF provides a means for scholars to better understand coalition formation and behavior, learning, and policy change (Sabatier and Jenkins-Smith 1993; Sabatier 1998; Weible et al. 2009; Jenkins-Smith et al. 2014). The ACF depicts the policy process as adversarial competition where actors form and maintain coalitions, engage in analytical debates with the potential for learning, and advocate for their preferred policy problems and alternatives. Following Sabatier and Jenkins-Smith (1993; Sabatier and Weible 2007), the appropriate scale for analyzing policy processes over time is a policy subsystem comprised of actors who regularly seek to influence policy choices surrounding a particular issue and within a particular geographic scope. Subsystems often contain advocacy coalitions, through which actors pursue common policy goals. These coalitions are made up of actors who have some degree of specialization in the policy area and include legislators, researchers, journalists, and actors drawn from government agencies, non-profits, businesses, and interest groups.

The ACF argues that policy-relevant beliefs are a principal motivator of individual behavior within policy subsystems. Policy-relevant beliefs include a mixture of normative beliefs (such as value systems or beliefs regarding appropriate roles of government in decision-making) and positive or empirically grounded beliefs (such as the likely impacts of a particular policy choice). In the ACF, belief systems have important consequences for the use of scientific information in the policy process (Sabatier and Jenkins-Smith 1993, 1999), the emergence and persistence of trust (Henry and Dietz 2011), and the structure of policy networks (Weible and Sabatier 2005; Henry 2011; Ingold 2011; Leifeld 2013). The ACF proposes an explicit model of belief systems, which are viewed as a three-tiered hierarchy of beliefs: deep core beliefs include normative and ontological axioms applicable to multiple subsystems; policy core beliefs include beliefs that support the achievement of deep core beliefs within a particular subsystem; and secondary aspects include a multitude of instrumental propositions to achieve policy goals within the subsystem (Sabatier 1998; Henry and Dietz 2012).

The ACF facilitates the study of change within policy processes that might include changes in beliefs through learning, changes in coalition members and their interconnections, and changes in policy. Policy change, for example, is hypothesized to occur through some combination of policy-oriented learning and belief change, negotiated agreements among members of rival coalitions, and exploitive activities of coalition members after major events, such as crises in conjunction with a coalition that exploits the opportunity (Nohrstedt and Weible 2010). These developments are likely to be followed by changes in governmental programs (Sabatier and Jenkins-Smith 1999). Similar to patterns of change examined in other policy theories (e.g. in punctuated equilibrium theory, see Jones and Baumgartner 2012), the ACF argues that resistance to policy change is the norm. For this reason, the most frequent type of observed change is minor adjustments to policies or their corresponding belief systems, with major policy changes occurring less frequently (Sabatier and Jenkins-Smith 1993). Certain contextual factors make major policy change more likely; for example, policy brokers can facilitate learning and policy change when they intervene in high-conflict situations to mediate solutions between opponents (Ingold and Varone 2011).

The ACF provides several advantages to scholars of comparative public policy. First, the framework identifies general concepts and provides theoretical and operational clarity in defining these concepts. Second, the ACF provides a lens to understand policy not as a single point in time but as a process without beginning or end, thereby sidestepping many of the limitations of viewing policy as a linear set of “stages” with a definitive beginning and ending point. Third, given its basic concepts and assumptions, the framework is somewhat malleable and open to experimentation in developing theoretical expectations and in using different forms of data collection and analysis.

Although the ACF offers a useful research platform to address puzzles of coalition formation and maintenance, learning, and policy change, it is far from providing all the answers. Many questions remain. Some hypotheses remain untested and underdeveloped (Weible et al. 2009; Jenkins-Smith et al. 2014). Instruments for coding texts, interviews, and surveys need refinement. And, most importantly from the perspective of comparative policy research, the ACF has rarely been applied to contexts outside of North America and Western Europe. As a result, critics claim the ACF is only applicable in open pluralistic political systems, where actors are free to mobilize and assemble in coalitions and to apply pressure on government for policy change through multiple venues over extended periods of time (Parsons 1995; Sabatier 1998). The purpose of this special issue is to help address this challenge by investigating and broadening empirical applications of the ACF in contexts outside of North America and Western Europe, by showcasing four applications with examples in Kenya, India, the Philippines, and China.

Is the ACF Applicable Outside of the US and Western Europe?

Since the inception of the framework, researchers have aspired to apply the ACF to different policy issues within different political systems (Sabatier and Jenkins-Smith 1999). Meanwhile, one of the early criticisms stated that the framework is founded on pluralistic assumptions and, therefore, that it would be restricted as an approach to explain policymaking (John 1998; Carter 2001). According to this view, the theoretical premises of the ACF recognized key features of pluralism, including factionalized, competitive, and adversarial groups involved in struggles over access to multiple venues for legislative influence (cf. Schattschneider 1960). Meanwhile, doubts were raised concerning its applicability in other systems characterized by democratic corporatist policy styles or authoritarian policy regimes (Parsons 1995).

In response, the ACF was gradually expanded (Sabatier 1998; Sabatier and Weible 2007; Jenkins-Smith et al. 2014) to include coalition opportunity structures, focusing specifically on the degree of openness of political systems (a function of the number of decision-making venues and the accessibility of each venue) and the degree of consensus needed for major policy change (Kübler 2001; Zafonte and Sabatier 2004; Sabatier and Weible 2007). These characteristics are assumed to affect the resources and constraints of subsystem participants. Yet, in practice, the nature of the relationship between these stable parameters and the behavior of subsystem participants remains an open question. Therefore, in order to explore the applicability of the ACF across different systems, researchers should engage in efforts to investigate empirically how processes of coalition formation, resource mobilization, and policy change may be enabled and constrained by basic coalition opportunity structures. On this basis, we agree with Parsons (1995: 200) that the bounds of applicability of the ACF should be assessed by empirical analysis: “We need to see a good deal more use of the framework outside and with case studies from other polyarchies if we are to come to any conclusions as to its general applicability.” Against this background, this special issue explores the theoretical, empirical and methodological challenges involved when studying policymaking in different political systems through the lens of the ACF.

Despite fundamental institutional differences across political systems, public policy scholarship recognizes the importance of “within-case variation”. Although the openness and venue access of political systems vary considerably, there are also significant variations between policy subsystems across time and space. Situated within the same constitutional structure, policy subsystems display different opportunities and constraints for policy actors to access and influence the policy process (Freeman 1985). Some of these properties may change over time within the same subsystem as well. What also speaks in favor of comparative empirical analysis across subsystems is the observation that the nature and practice of political systems are changing over time. Studies in corporatism, for example, suggest that traditionally corporatist systems have changed gradually towards more open, informal and conflictual policymaking modes (e.g. Kickert 2003; Rommetvedt 2005; Lindvall and Rothstein 2006). Also, when it comes to understanding the interaction between political institutions and policy change in developing countries, researchers have focused much attention to underlying social, economic, and political constraints (e.g. state–society relationships, corruption, financial crises, and political power relationships) whereas less attention has been devoted to understanding processes of policy change that result in revised or new programs with improved performance (Steinberg 2003). Comparative studies on policymaking across policy subsystems in different political systems provide an opportunity to document how these changes emerge in practice.

The ACF outlines three major theoretical emphases focusing on advocacy coalitions, policy-oriented learning, and policy change. Each one of these theoretical emphases narrows the scope of inquiry, links concepts by outlining observable implications, and establishes relationships among variables (Weible et al. 2011; Weible and Nohrstedt 2012). This special issue primarily focuses on one of these theoretical emphases – policy change – and it seeks to deepen our knowledge of how policy change occurs in different political systems. To meet this objective, the contributions address several overlapping questions. How do institutions, culture, policy-relevant beliefs, or problem attributes shape the policy change process? What are some of the intervening steps linking events to policy change? How might learning and negotiated agreements lead to policy change? How important are changes in the distribution of resources within and among coalitions as well as the role of policy brokers in policy change? The effort to answer these questions by comparative analysis relates to the continued development and refinement of the ACF (Weible et al. 2011; Scott 2012) as well as to the ongoing study of policy change in comparative contexts (Capano and Howlett 2009). Recent contributions to the study of policy change recognize the need to more carefully address the “dependent variable problem” (Howlett and Cashore 2009) and the explanatory variables and causal mechanisms that contribute to policy change over time (Capano 2009). In response to these challenges, each paper in this special issue specifically addresses the nature and scope of policy change and plausible explanations derived from the ACF. We hope that this special issue will contribute to policy change research in general and the ACF research program in particular.

Empirical Applications Outside of Western Europe and North America

Applications of the ACF outside of North America and Europe are a recent and relatively rare endeavor with only 27 applications to date out of 224 total (see Table 1), with Sato’s publication in 1999 being the first.¹ Most of the studies listed in Table 1 apply nearly all aspects of the ACF, including coalitions, learning, and policy change. This reflects the general tendency to apply the entire framework to understand the overarching policy context of a given policy subsystem over time, rather than to narrow the focus on one particular component of the framework in order to develop greater theoretical description and explanations. Such general applications also reflect the source of data. Most studies outside of North America and Western Europe use an informal analysis of documents with some document coding, and some studies complement their analysis with open-ended qualitative interview data. These applications generally lack large-N data coupled with quantitative analysis, which may reflect the myriad challenges of administering mail and online surveys outside of North America and Western Europe.

Table 1. A listing of ACF applications outside of Western Europe and North America

Author (year)	Subsystem topic	Country
Sato (1999)	Tobacco policy	Japan
Bulkeley (2000)	Climate change policy	Australia
Hogl (2000)	Forest policy	Australia
Carvalho (2001)	Metallurgical policy	Brazil
Elliott and Schlaepfer (2001)	Forest certification policy	Canada, Indonesia, and Sweden
Tewari (2001)	Forest policy	South Africa
Arnold (2003)	Forest policy	Chile
Chen (2003)	Computer regulations	Australia
Farquharson (2003)	Tobacco Policy	Transnational
Kim (2003)	Conservation policy	South Korea
Hsu (2005)	Nuclear power policy	Taiwan
Nagel (2006)	Labor policy	Australia
Villamor (2006)	Forest policy	Philippines
Beverwijk et al. (2008); Beverwijk (2005)	Education policy	Mozambique
Ainuson (2009)	Water policy	Ghana
Fang (2010)	Economic/Business policy	South Korea
Kingiri (2011)	Biotechnology	Kenya
Lahat (2011)	Poverty policy	Israel
Montefrio and Sonnenfeld (2011)	Forest policy	Philippines
Santa 2013)	Tourism policy	Philippines
Dressel (2012)	Economic/Business policy	Philippines
Li (2012)	Environmental policy	China
Francesch-Huidobro and Mai (2012)	Climate policy	China
Kim (2012)	Environmental policy	South Korea
Ortmann (2012)	Civil society policy	Singapore
Marfo and McKeown (2013)	Forest policy	Ghana
Runkle et al. (2013)	Health Policy	Liberia

The authors from Table 1 also commented on the strengths and weaknesses of applying the ACF in their context. The reported strengths are paraphrased below:

• The ACF generally provides a useful theoretical approach for understanding policy issues comparatively (Elliott and Schlaepfer 2001).

• The ACF’s focus on change in coalitions, in learning, power, and policy is useful for studying public policy over time (Carvalho 2001; Elliott and Schlaepfer 2001; Chen 2003; Kingiri 2011; Marfo and McKeown 2013).

• The ACF provides a useful means for identifying coalitions based on beliefs and perceptions (Kim 2003; Nagel 2006).

• The ACF provides a good way for understanding policy change as resulting from coalition politics (Sato 1999; Hsu 2005).

• The ACF’s perspective of subsystem politics and non-sequential change is useful for understanding policy processes (Carvalho 2001).

• Policy change within the ACF can be studied by examining changes in the relatively stable parameters (Beverwijk et al. 2008).

• The ACF offers useful principles for recommending political strategies regarding policy change (Runkle et al. 2013)

The authors reported several weaknesses as well:

• The ACF needs to do a better job of integrating into policy subsystems international organizations and the effects of different contexts (Elliott and Schlaepfer 2001; Farquharson 2003; Hsu 2005; Ainuson 2009; Santa 2013)

• The ACF assumes sovereigns are “neutral arbiters” (Chen 2003).

• The ACF needs to develop its hypotheses better for nascent policy subsystems (Beverwijk et al. 2008).

• The ACF could improve its description of coalition politics by dealing with internal coalition relations and divisions, provide better theoretical depictions of belief systems (including relations and depictions between policy core and secondary beliefs), and describe theoretically how coalitions change over time (Sato 1999; Arnold 2003; Kim 2003; Nagel 2006).

• The ACF provides limited explanations for processes of learning (Bulkeley 2000) and how coalitions translate their beliefs into actual policy (Sato 1999).

• The ACF undervalues the role of institutions (Nagel 2006).

• The ACF needs theoretical development to understand and explain the structure and stability of coalitions in different political systems (Li 2012; Ortmann 2012; Santa 2013).

• The ACF needs to account for fundamental values of a society that could drive subsystem activities and policy change (Kim 2012).

Overall, the authors describe beneficial uses of the ACF to countries outside of North America and Europe, but not without a number of accompanying challenges. Among those who question the applicability of applying the ACF in different political systems is the collection of applications in Asian countries (Scott 2012). The applications in Singapore and China, in particular, raise thought-provoking questions about the structure and stability of coalitions in hybrid and authoritarian regimes. Similarly, Beverwijk et al. (2008) commented strongly on the challenges of applying the ACF in different political settings. Ainuson (2009), among others, mentions how useful it was to apply the ACF in their country of study. However, Ainuson (2009) also directs authors towards understanding the factors shaping subsystem affairs.

An Overview of the Special Issue

This special issue includes four applications spanning the globe (see Table 2). The major aim of the contributions is to highlight drivers for major or minor policy change. Building from the assumptions and hypotheses of the ACF, the authors focus on factors internal and external to the investigated subsystems in order to explain what might induce change in different policy fields. Aside from being non-US and non-European applications, the four applications cover a wide range of political subsystems including hydropower in China; indigenous peoples’ rights in the Philippines; nuclear energy and forest management in India; and biotechnology policy in Kenya. For each study, we outline the research questions, major analytical elements, empirical and methodological approaches, and conclusions.

Table 2. A listing of ACF applications in the JCPA special issue

Author (year)	Subsystem topic	Country
Han et al. (2014)	Hydropower development project	China
Montefrio (2014)	Indigenous peoples rights	Philippines
Gupta (2014)	Nuclear energy and forest management	India
Kingiri (2014)	Biotechnology regulatory policy	Kenya

Han, Swedlow and Unger analyze factors that led to China’s suspension of its latest large-scale hydropower project on the Nu River. This paper is one of the first applications of the ACF in China and its authoritarian system. The authors convincingly demonstrate the added value of the ACF to explain policy change in China’s hydropower development subsystem in comparison to the fragmented authoritarianism model. Through primary and secondary literature research and additional open-ended interviews, the authors were able to identify actors’ belief systems, coalitions, resources, and mobilization. More concretely, China’s hydropower subsystem is characterized through various state and non-state entities that either coalesced into a development coalition or an environmental coalition which held incompatible policy beliefs regarding this hydropower project. Finally, the authors conclude that the following factors contributed to the project suspension: adversarial interactions between the two coalitions, policy learning within the subsystem, interventions by one policy broker (Premier Wen), and, finally, the intervention by the State Environmental Protection Administration (SEPA).

Montefrio analyzes the case of the Indigenous Peoples Rights Act (IPRA) in the Philippines. The aim of this case study is to explain regressive policy change and weak implementation, focusing on exogenous factors such as constitutional structures and conflicting subsystems. The identification of coalitions and their coordination is based on data gathered through content analysis of official documents, in-depth key informant interviews, and participant observation. The author explains the delays in the creation of the law, the law’s weak implementation, and incremental policy changes through the Regalian Doctrine provisions in the Philippine constitution and the competing environmental, conservation, and natural resource policy subsystems that overlap with the IPRA. One added value of this paper is therefore the in-depth investigation of subsystem interconnectedness and how exogenous drivers intervene with internal subsystem factors in producing policy change.

Gupta compares two subsystems: nuclear energy and forest management in India. The author asks how coalition opportunity structures impact on coalition strategies. Gupta outlines how subsystem-relevant coalition opportunity structures influence strategies in multiple ways, including providing opportunities or constraining access to new resources within subsystems, shaping opportunities for venue exploitation, and setting legal impediments that can impact on the range of available political strategies. One major added value of this study is its longitudinal and comparative investigation across two subsystems (nuclear and forest policy) applying a similar-systems approach. Through a content analysis of primary and secondary sources, including newspaper archives, Gupta was able to identify coalitions and their strategies in both cases and over time. She finds that highly centralized policy subsystems with restricted access to decision-making led to coalitions adopting strategies such as public protests, hunger strikes, and rallies designed to disrupt the subsystem status quo. By contrast, she concludes that decentralized policy subsystems with relatively open decision-making processes led coalitions to adopt strategies such as deliberation, appeals, and petitions designed to work within the subsystem status quo. This application provides substantial insights about the investigation of subsystem internal opportunity structures and shows how they vary, both across subsystems and within subsystems over time. The article is therefore an intriguing and significant addition to the growing literature within the ACF studying opportunity structures.

Kingiri, in her study of biotechnology policy in Kenya, asks how advocacy coalitions and policy brokers influence policy change. She combines analytical and theoretical elements from the advocacy coalition framework with theories of innovation, focusing mainly on the role of brokerage in policy processes. Data were gathered through interviews with key representatives of private and public organizations, and provide insights into the subsystem structure, coalition formation, the identification of key actors and brokers, and the factors that motivate actions of actors in the conflictive domain of Kenya’s biotechnology policy. This case study supports ACF propositions that disagreement over policy beliefs generate conflict and distrust between adversarial coalitions. This led to venue shopping strategies being employed by some subsystem actors in order to preserve the status quo in Kenya’s biotechnology subsystem. Kingiri concludes that coalition and broker activities are further explained by, among other reasons, interests, motivations, and opportunities.

Toward an International Research Program on the Policy Process

Despite the emergence of the ACF in the context of the United States, the applications presented in this special issue demonstrate how the ACF can be used as a common foundation to empirically study advocacy coalitions and processes of learning, and policy change across a range of contexts. The ACF serves as a common foundation among the researchers in this special issue by establishing the policy subsystem as a common scale for analysis, by guiding their analyses of beliefs among actors in coalitions, and by offering some theoretical rationale for processes of learning and policy change.

Such a foundation, however, cannot provide every descriptive factor or explanatory mechanism which is important for understanding coalition behavior, learning, and policy change. The reason for this is simple: the number of specific factors and interrelations between factors are uncountable. But a framework can provide a shared language and approach for studying similar phenomena and generating shared knowledge. To make such an enterprise work, attention needs to be given to developing strategies for conceptual clarity and theoretical expectations (Easton 1971; Weible and Nohrstedt 2012). For instance, as shown by the contributions in this special issue, the investigation of belief systems is one of the important initial steps for an ACF application. The ACF assumes a three-tiered belief system, thereby guiding the analyses across contexts of fundamental deep core beliefs, more subsystem-bounded policy core beliefs, and instrumental secondary beliefs. However, the exact beliefs to be measured within any tiered level are contingent upon the context of study (Henry and Dietz 2012), as illustrated by the beliefs measured in the study by Kingiri (2014).

Another common foundation for ACF applications is the policy subsystem as an appropriate scale for empirical analysis. As a subunit of a broader system, all policy subsystems are set within broader political, cultural, socio-economic, and physical conditions. The range of approaches to documenting the broader governing system includes Gupta’s (2014) study of coalition opportunity structures, which she found affected coalition strategies. Such attention to the external context of policy subsystems must continue to be accounted for when applying the ACF in any setting, including Western Europe, North America, and anywhere else in the world.

In addition to learning about added value and shortcomings when applying the ACF to cases outside of the US and Europe, the four studies also nicely contribute to central, current discussions in the recent literature on the ACF (see Weible et al. 2011; Weible and Nohrstedt 2012; Jenkins-Smith et al. 2014). One major question, for instance, is how different subsystems may influence one another. Montefrio focuses strongly on subsystem interconnectedness to explain changes in the Philippine Indigenous Rights Act, and highlights how constitutional factors on the national level may affect conflicts among different policy fields. Through the comparative study of India’s nuclear energy and forest management policies, Gupta contributes to the discussion of how subsystem specificities, and in subsystem-internal opportunity structures, are able to shape coalitions’ strategies. All of those studies point out the need for further research into vertical and horizontal subsystem interconnectedness.

Another issue of attention in recent ACF literature involves policy brokerage. Han et al. (2014) show, for instance, how the premier acted as a policy broker in shaping policy change in the Nu River subsystem. This application raises the important question of how to identify policy brokers based on the ACF’s definition of the term. Kingiri’s study of Kenya’s biotechnology policy takes on this question and asks, as a starting point, how policy brokers influence policy change. This application shows how the ACF can be combined with other theoretical frameworks. It therefore contributes to the discussion on how to conceptualize and operationalize brokers and brokerage in the policy processes (for a comparison, see also Christopoulos and Ingold 2011; Ingold and Varone 2011).

The evidence offered by this special issue should provoke further examination of competing theoretical approaches to policy change across political systems. One avenue for future research involves examination of the strengths and weaknesses of the ACF compared to alternative policy process approaches and frameworks. When widening the empirical basis for the ACF beyond political systems within the US and the EU, new questions are being raised regarding, for example, the ability of marginalized groups or “non-elites” to participate and influence the policy process (Holmes and Scoones 2001). This illustrates how, as applications of the ACF in diverse contexts continue to accumulate, we will be able to observe variation in important variables that tend to remain constant when one focuses solely on the US and the EU but are central to other theoretical perspectives. Another intriguing challenge involves contrasting the ACF with conventional approaches used to understand policy processes in specific contexts, such as the fragmented authoritarianism framework in China (Han et al. 2014). Studies that critically compare and contrast the theoretical premises of the ACF with other frameworks and theories are therefore an important basis for future development of the ACF (Weible et al. 2011).

Contributions to this special issue add new empirical insights that should inspire further analysis of the comparative strengths and weaknesses of the ACF as a basis for understanding the policy process across contexts. We expect these empirical efforts to be followed by future applications that continue to address the basic premises of the framework through critical theoretical analysis and methodological innovation. This will also support scholarship focused on testing the micro-level theories of human behavior embedded in the ACF, and the degree to which political or social contexts are likely to matter for processes such as the learning of new beliefs, shifting norms of behavior, or the formation of cooperative networks (Henry and Dietz 2012).

Finally, although limitations of applying the ACF outside of Europe and North America are evident in this special issue as well as through the applications listed in Table 1, we encourage the continued comparative development of the framework across these different contexts. Assumptions may need to be questioned, theory developed, context better incorporated, and hypotheses rejected and revised to enable comparisons and valid insights. We see no other way to gain traction in issues of coalitions, learning, and policy change other than to continue to apply the ACF outside of North America and Western Europe. The challenge, of course, is not to blindly apply the ACF without the incorporation of the context of study, and then to offer best practices for incorporating such contexts into the framework.

This introduction of the ACF suggests that the framework is somewhat malleable and can adapt to different contexts. In this regard, we agree with Cairney’s (2012: 19) assessment that “the ACF resembles one of its coalitions; its authors engage in learning to refine the ACF’s secondary aspects while protecting the core argument. This is no accident. Rather, it reflects a “‘Lakatosian’ approach to science that may be relevant to the overall study of public policy”. Frameworks and theories that cannot, in some way, adapt to scholarly mistakes or new information will probably be discarded too quickly to permit learning among the applicators, especially given the challenges in generating instruments, the context of our observations, and the difficulties in communication. Overcoming these challenges will happen through the application of the ACF in different contexts, which will provide important insights into the strength and weaknesses of the ACF’s logic and help propel the literature forward.

We adhere to Simon’s advice in his metaphorical description of how developing good theory is like designing a motor:

In a benign environment we would learn from the motor [framework] only what it had been called upon to do; in a taxing environment we would learn something about its internal structure – specially about those aspects of the internal structure that were chiefly instrument in limiting performance. (Simon 1996: 12)

Through the application of the ACF in the taxing contexts of China, Philippines, Kenya, and India, we are able to learn more about the strengths and limitations of the performance of the advocacy coalition framework.

Notes

^1. The 27 applications listed in Table 1 were collected using a key word search on the phrase “advocacy coalition framework” in ProQuest and Google Scholar. The results were then manually searched and affirmed (or not) to have used the ACF as a principal theoretical lens for the empirical application. See Jenkins-Smith et al. (2014) for a description of the 224 applications of the framework.