Policy process research and the causal mechanism movement: reinvigorating the field?

Schlager and Blomquist’s paper comparing three “emerging theories of the policy process.” It examined the ACF, Elinor Ostrom’s Institutional Rational Choice (IRC, which later became IAD), and Terry Moe’s politics of structural choice approach. Issues of mechanisms and causality were briefly discussed, with ACF touted as a more sophisticated incorporation of the roles of information and learning; it challenges the other frameworks to consider the “ideological filtering of information, and changes in individuals’ beliefs, as mechanisms promoting or inhibiting policy change” (p. 666).


INTRODUCTION
Over the last few decades, enormous strides have been made in developing theoretical frameworks that capture the complex, dynamic, inertial and punctuated change features of policy-making. Leading policy process scholars have developed a portfolio of frameworks and theories for addressing key challenges of policy-making, identifying key dynamics and institutional contexts, and explaining why change occurs or does not occur at different phases of the policy process and whether policy succeeds (Weible and Sabatier, 2018). In this chapter, we focus on seven popular approaches: the Multiple Streams Framework (MSF), the Advocacy Coalition Framework (ACF), Punctuated Equilibrium Theory (PET), Institutional Analysis and Design (IAD), Policy Diffusion (PD), Policy Feedback Theory (PFT), and Narrative Policy Framework (NPF). Although interest in other approaches has surged at times -for example, policy network or constructivist perspectives -these frameworks predominate, having been debated, modified and enduring as the leading approaches. They provide the concepts and propositions that many scholars and practitioners use to analyze and appraise various facets of policy-making. 1 What is surprising, though, is the extent to which these approaches overlap (which their progenitors have long recognized regardless of when each tradition gathered steam), and second, while appearing well articulated from a theoretical perspective, they do not seem well specified considering the granularity of policy-making, the challenges of conducting empirical research on specific policy problems and domains, and the recent interest in policy design and the practical implications of such research (Weible and Cairney, 2018). As the policy design movement has taken shape (e.g., Clarke and Craft, 2018;Howlett, 2018), a complementary wave of interest in better specifying Policy process peer-reviewed papers using the word "mechanism" the "mechanisms" movement has been gathering steam. Inspired by the philosophy of science, social science and political science (Tilly, 2001;Falleti and Lynch, 2009), it has important implications for the policy sciences and policy process theoretical frameworks and associated programs of research.
Over the past decade, the language of mechanisms has been seeping into the policy process literature. Figure 2.1 broadly illustrates this trend by tracking the use of the word "mechanism(s)" in titles or abstracts of 226 peer-reviewed papers using the seven policy process frameworks noted above. It shows that notions of mechanisms are proliferating in the policy literature but does not provide a nuanced sense of the extent to which and how mechanisms are applied in the theoretical literature. This chapter is a first attempt to assess the state of theorizing from a "mechanism" perspective associated with policy process frameworks.
We were asked by the book editors to appraise the assessments of contributors to the seminal Theories of the Policy Process collection (Weible and Sabatier, 2018), now in its fourth edition, devoted to describing and providing snapshots of the state of theorizing and empirical studies associated with seven well-known policy frameworks. 2 The progenitors of these frameworks explicitly set out theoretical assumptions and propositions for their approach, and encouraged empirical studies that test and improve the evolution of the frameworks. This chapter appraises these descriptions of the main policy process frameworks, provides a sense of which frameworks seem best developed from a policy mechanism perspective, and identifies the implications for research, Evert Lindquist and Adam Wellstead -9781788118194 Downloaded from Elgar Online at 05/06/2020 07:13:30AM via free access design practice, and pedagogy. We see the arrival of policy mechanism approaches -what we refer to in this chapter as the "policy mechanism movement" 3 (PMM) -as invigorating a field coalescing around a few theoretically informed high-level frameworks through a call for deeper theorizing and more focused empirical studies. The PMM, we suggest, seeks more detailed specification of frameworks and theories, not only to inform finer-grained empirical work but also to arrive at more nuanced and perhaps more robust explanations, with the potential to improve governance and policy design. While PMM challenges and provides a strong critique of existing policy theoretical frameworks, it constructively points the way forward and promises important implications for furthering knowledge and design practice, and better linking theoretical frameworks to teaching in policy and public administration schools. This chapter has five main sections. The first section provides a brief review of essential concepts and distinctions from the mechanism literature. The second section identifies the main theories of policy-making, largely by introducing a key book in the field, Theories of the Policy Process (TPP) which over four editions (spanning 1999-2018) has filtered out and described the central theoretical frameworks. The third section describes the approach we took to reviewing TPP, especially the fourth edition (Weible and Sabatier, 2018), while the fourth section contains our assessments of each chapter and its policy process framework. In the fifth section we provide observations about the extent of use of mechanisms in the policy literature as well as the state of theorizing. We conclude by suggesting that the PMM could lead to a new round of theoretical and empirical research building on the foundations of different theories of the policy process, but also serving to better work across and link these traditions.

THINKING MECHANISTICALLY: ESSENTIAL CONCEPTS AND DISTINCTIONS
Advancing the policy process literature to improve explanation and design requires deeper understanding of causality. This will involve policy scholars identifying and measuring key "causal mechanisms" responsible for policy outcomes. There have been calls for identifying causal mechanisms in policy theory (Yee, 1996;John, 2003;Steinberg, 2007;Nowlin, 2011;Kay and Baker, 2015). Quite independently, Weible (2018) has recently argued that "causal drivers" lie at the heart of the scientific assumptions underlying these theories. Throughout the literature, causation is often claimed or implied, but often only supported, if at all, by sketchy or shallow explanations (Nowlin, 2011). The renowned philosopher of science, Mario Bunge (1997), coined this incomplete theorizing as "gray box theory," where causality is assumed but the mechanisms are rarely described and hypotheses infrequently tested. Similarly, Morgan and Winship (2014) refer to such abstract relationships as "mechanism sketches," which stand in contrast to "mechanism schema," where the component parts or entities are known. A more detailed understanding of policy-making requires what Bunge (1997) calls a "translucent" box, more clearly identifying details about mechanisms or the cause-and-effect relationships between policy-makers' attention to policy problems and their receptivity to policy solutions. Once the nature of pertinent mechanisms are understood, it becomes possible to apply more rigorous empirical methods and to make inferences. Identifying causal mechanisms helps explain how and why some decisions have the desired effect, and why others are resisted. In other words, we seek to identify "what works" when actors engage in the policy process to use evidence and seek policy outcomes consistent with that evidence.
The large social science "mechanisms" literature has been influenced by insights from the natural sciences and philosophy of science. Mechanisms are sets of entities and activities organized to produce a regular series of changes from a beginning state to an ending state (McAdam, Tarrow and Tilly, 2008). Mechanisms usually "invoke some form of 'causal agent' that is assumed to have generated the relationship among the entities observed," serving as "analytical constructs that provide hypothetical links between observable events" (Hedström and Swedberg, 1998, p. 11). At first blush, mechanisms are often seen as unobservable or hidden phenomena, sensitive to variations in context, but in principle can be empirically traceable processes where a cause (or causes) can generate an outcome (Pawson and Tilley, 1997). Assessing the logic of association helps to open the black box of the limited X → Y causal inferences so prevalent in the social sciences (Falleti and Lynch, 2009). Mechanistic thinking brings to the surface deeper processes and variables at play.
Causality is not simply a functional description of a certain variable, but requires uncovering how X produces Y under specific conditions. It is a theoretical formulation that "adduces properties of the relationships among phenomena with the potential to recur, which helps explain why x causes y" (Hall, 2013, p. 21). Context can trigger or condition a relationship and the role it plays in determining outcomes. Initial conditions play a key role in determining how mechanisms are triggered and how they respond to certain contextual conditions. Identifying the context and the mechanism is important when formulating hypotheses. It is critical to understand under what conditions mechanisms are most likely to occur or produce a particular outcome (Pawson and Tilley, 1997). Various scholars have adopted a "context mechanism outcome" (CMO) approach: namely, the observed patterns of (un)intended outcomes can be explained by identifying the plausible causal set of mechanisms within the situational context of the process (Pawson and Tilley, 1997;Biesbroek, Dupuis and Wellstead, 2017) (Figure 2.2). The context influences when and   how certain mechanisms are triggered, and how they play out. Context is critical because similar initial conditions may lead to dissimilar outcomes (multifinality); conversely, outcomes can be reached from distinctly different developmental paths (equifinality) (Biesbroek et al., 2017).
This more robust understanding of causality permits the skeptical scientist to open up the black or gray boxes of policy-making to find a diversity of causal mechanisms that could affect policy outcomes. First, different categories of mechanism have been identified: structural (e.g., environment, institutions), cognitive (e.g., individual perceptions and ideas), and relational (e.g., network connections between people). Second, mechanisms can span across micro-level (individual) and macro-level (structural) phenomena (Bunge, 1997;Checkel, 2006). Given the multi-level nature of policy-making, such distinctions are important, as Figure 2.3 shows. 4 Situational mechanisms occur when social structures or environmental phenomena constrain individuals' action or shape beliefs. Action-formation mechanisms link individual micro-level activities or behavior to their actions. Transformational mechanisms are those where individuals, through their actions and interactions, generate intended and unintended outcomes.
The temporal nature of mechanisms is material, including the time horizons associated with the workings of a mechanism (which might involve a chain of effects) and the outcome (Pierson, 2003;Beach and Pedersen, 2013). For example, a slow-moving causal process may eventually result in a threshold event precipitating a sudden change. In the policy and social sciences, many mechanisms fit these broad categories. However, identifying high-levelalbeit somewhat abstract -mechanisms in policy theory often brackets detailed delineation of the causal factors associated with important components of policy theory in a gray causal box (Sartori, 1970;Falleti and Lynch, 2008). Drawing on Sartori's (1970) "ladder of abstraction," Falleti and Lynch (2008) argue that developing more compelling and measurable causal explanations can only happen if mechanisms are disaggregated from high-level abstractions -which they label as "processes," "mechanisms-as-type," and "mechanisms-as-example" -which are more conceptualizations of mechanisms than a measurement strategy. More importantly, they claim that "mechanisms-as-cause" and "mechanism-as-indicators" are critical to making a measurable causal claim and describing how things happen.
From a vertical perspective, Machamer, Darden and Craver (2000) argue that mechanisms are often nested hierarchies in systems that contain "lower level entities, properties, and activities" that "produce higher level phenomena" (p. 13). That is, "the components that are accepted as relatively fundamental or taken to be unproblematic as far as the observables in the data" (Morgan and Winship, 2014, p. 239). Machamer et al. (2000) borrow from molecular biology and find that mechanisms "bottom out in descriptions of the activities of macromolecules, smaller molecules, and ions" (p. 14).
This brief review of the causal mechanism literature distils the essence of the approach, and its relevance for the multi-level (macro, meso, micro) policy process frameworks that attempt to deal with multiple dependent and independent variables. This literature does not favor quantitative over qualitative empirical methods, but more generally calls for theoretical elaboration and empirical testing in the service of better specifying predictable relationships. This aspiration has long been shared with the policy sciences, which has always sought to improve the basis for advice and the quality or predictability of policy interventions. (1999,2007,2014,2018) The Theories of the Policy Process (TPP) collections have become a vital resource for policy scholars, providing snapshots of the state of theorizing and empirical work. 5 The four editions reviewed seven or eight policy theories or frameworks. After the first edition, the editors of the subsequent editions flagged the importance of causality. In TPP (1999), Sabatier mentions mechanisms and states that, "to think carefully about the steps in the causal process is one of the principal steps in going from general frameworks to denser, more logical interconnected theories" (Sabatier, 1999, p. 268). Several years later, Sabatier (2007, p. 328) argued that, "[t]he failure to develop clear chains of causal relationships" was the contributing factor for the extinction of earlier policy process frameworks (e.g., Dye-Sharkansky-Hofferbert -DSH framework). Sabatier and Weible reference the mechanisms scholarship in TPP (2014). Finally, Tosun and Workman (2018) suggest that approaches "facilitating a more demanding empirical test of the causal mechanisms underlying policy process theories" will "help to increase confidence in their analytical merits" (p. 330).

THE CAUSAL PREREQUISITES IN THEORIES OF THE POLICY PROCESS
A challenge to applying mechanisms is raised in Schlager's chapter in TPP (2007). Citing Ostrom, Schlager states that "frameworks organize inquiry, but they cannot in or themselves provide explanations for, or predictions of, behavior and outcomes. Explanation and prediction lie in the realm of theories and models. These approaches are often a collection of theories from which mechanisms can be derived" (Schlager, 2007, p. 293). In TPP (2014), and updated in TPP (2018), Cairney and Heikkila (Cairney and Heikkila, 2014;Heikkila and Cairney, 2018), undertook the first extensive comparison of the approaches. In both editions, the elements, research program, and the emphasis of the approaches are raised in a series of three tables comparing the seven approaches. However, no comparison was ventured with respect to causality. By identifying the role of causality and mechanisms in TPP (2018) comparative fashion, we aim to construct a " Table 4." These collections have never sought to provide detailed reviews of each policy framework, but rather, described the essential features and evolution of each approach, surveyed empirical contributions, and identified trends and avenues for future research. Given its centrality in our field, and that Sabatier and/or Weible stress the importance of causality and mechanism in the policy process literature, a review of TPP through the lens of mechanisms should provide a useful, if high-level, sense of the extent to which the approach has been taken up in the field. Policy process frameworks, theories, and models reviewed in four editions of "TPP" TPP1 (1999) TPP2 ( A two-pronged textual analysis was undertaken (Lindkvist, 1981). First, we sought to get an indication of the extent to which the authors invoked mecha- nisms or causality terminology in the text. QSR NVivo 11, a qualitative data analysis computer software package, was employed. References relating to the terms "causality" and "mechanism" were coded as nodes. From this qualitative effort, the context of the term's use was accessed. However, frequencies may not necessarily reflect each chapter's treatment of the terms. Second, we reviewed each chapter and assessed whether and the extent to which it applied mechanistic analysis based on our survey of the mechanisms literature, and regardless of whether it used such terminology, whether it undertook or was disposed to such analysis. Table 2.2 reports the frequency of causality-and mechanism-related terms in TPP4. From the quantitative accounting of approaches undertaken in QSR NVivo, the most frequent mention of mechanism was found in IDM and PFT chapters. The NPF chapter invoked causality the most often. MSF had very few references to either term. However, as we discuss below, because the terms "causality" and "mechanism" were mentioned in the chapters does not mean that they were invoked in the way called for by the policy mechanisms movement; conversely, some chapters demonstrate approaches consistent with mechanistic analysis but not invoking mechanistic language, though grappling in varying degrees with the challenge of working across levels of analysis (individual, meso, macro) when describing the genesis and evolution of the frameworks as well as the empirical progress made in the respective domains.

RESULTS
What follows provides an overview of each policy framework's essential approach, the extent to which the authors of the respective chapters relied on a "mechanism" approach (explicitly or implicitly), and whether the empirical literature and prospective research agendas as described in the chapters (usually in very high-level ways) seem consistent with the "policy mechanism" approach advocated for in this collection.

Multiple Streams Framework (MSF): Limited Causal Claims, High Potential of Mechanisms Research
First introduced by Kingdon (1984), the MSF is one of the oldest policy process approaches. Inspired by Cohen, March and Olsen's (1972) organizational garbage can model, the MSF assumes that policy-making is beset by ambiguity, time constraints, incomplete policy preferences, and constant fluidity of policy actors. The basic structural make-up consists of three streams (problems, policies, and political), policy windows, and policy entrepreneurs and processes (attention, search, and selection). Whether policy proposals become part of a government's agenda is the function of the three streams coupling at an opportunistic time (which "opens" a policy window), often promoted or taken advantage of by a policy entrepreneur. Mechanisms and causality receive limited discussion in this chapter by Herweg, Zahariadis and Zohlnhoefer (2018). Only the idea of types of coupling of the three streams is presented as a type of mechanism. The authors state that future MSF applications should specify causal mechanisms dependent on the context (policy stage). Some of the MSF's elements and process reference identifiable high-level mechanisms such as prospect theory and feedbacks, which can be readily disaggregated. On a positive note, several hypotheses are listed, which implies the potential of mechanism-related research.

Punctuated Equilibrium Theory (PET): Policy Mechanism Research in Practice, Not Name?
Inspired by early work on bounded rationality and notions of episodic and rapid speciation from the natural sciences, Punctuated Equilibrium Theory (PET) seeks to describe and explain how public policy regimes break from patterns of "marginal and incremental" policy change towards more dramatic change at certain junctures. Baumgartner, Jones and Mortensen (2018) describe how policy monopolies, buttressed by "policy images" (problem definitions and claims for certain policy approaches), can shape and constrain change in specific policy domains, along with the limited attentiveness of governments and policy-making systems at the macro level (flowing from bounded rationality of individuals, organizations and systems), and can combine to produce negative feedback loops, constraining forces and calls for change. Conversely, significant policy change can occur when events and facts burst through the claims of policy monopolies, and issues move high on the policy agenda, leading governments at the macro level to take an interest in a specific policy domain. This Evert Lindquist and Adam Wellstead -9781788118194 Downloaded from Elgar Online at 05/06/2020 07:13:30AM via free access may not necessarily lead to significant policy change, but the authors argue that it is a necessary condition, and constitutes a positive feedback loop. The authors survey the empirical work that they and colleagues have undertaken with PET, focusing on studies of budgeting and other policy domains in the US and noting the expanding comparative literature in these areas, and relying on stochastic studies of budget outcomes over many years. The authors use the term "mechanism" a few times, but not in the way advocated by the policy mechanisms movement literature described earlier.
Mechanism is variously invoked to describe policy subsystems allowing the large system to "engage in parallel processing" (p. 59), processes associated with human and organizational "cognitive architecture" (p. 65), "discontent" as a means for communicating problems (pp. 66-8), policy images, issue expansion, and policy development (respectively, pp. 69, 85-6), and "continuous dynamic adjustment" as "the primary decision mechanism" (p. 74). These references neither delve into the more detailed processes through which the process or function would achieve its effect, nor the conditions. They can be seen as pointing to "gray box" processes.
However, while not using the mechanism language, Baumgartner and Jones's multi-decade program of research on agenda-setting and punctuated equilibria nevertheless exhibits many of the tenets of "mechanistic thinking." It is shown in how carefully specified and elaborated the assumptions and theoretical underpinnings of their multi-level framework are, and a systematic empirical research agenda of comparative research within the US, particularly with respect to budget outcomes. In recent years, they secured funding for cross-jurisdictional studies using similar research methods, which allowed them to explore the effects of different governance systems and political contexts, and evaluate the effects of top-down system-level dynamics with more bottom-up dynamics associated with specific issues and policy domains (Jones, Baumgartner and True, 1998;Baumgartner et al., 2009).

Policy Feedback Theory (PFT): Unpacking and Disaggregating Mechanism Sketches
Mettler and SoRelle (2018) state that Policy Feedback Theory (PFT) has led to the investigation of the specific mechanisms and pathways through which policies affect political attitudes and behaviors among mass publics. Like the MSF, the PFT employs broad "streams" whereby current public policies affect meaning of citizenship, forms of governance, the power of groups, and political agendas, and the definition of policy, which in turn constrain future public policies.
Despite the causal arrows from public policies to political outcomes, discussion of causality is completely absent. For example, it is unclear what the cause Evert Lindquist and Adam Wellstead -9781788118194 Downloaded from Elgar Online at 05/06/2020 07:13:30AM via free access of that effect from public policy (in Time 1) happens to be. Also absent are causal mechanism(s) relating to "meaning of citizenship" and its effect on a public policy in Time 2. Mettler and SoRelle, to their credit, hone in on the specific factors affecting meaning of citizenship (e.g., resource effects, interpretative effects, civic capacity, and civic predisposition). By doing so, there is a rudimentary form of theory-testing process tracing. Later, they argue that it is important to open up the "black box of how public policies impact political behavior of ordinary citizens" (p. 118). The need to test hypotheses in more rigorous ways and identify mechanisms with greater specificity is highlighted. Without using the term mechanism, for each of the major streams of policy feedback inquiry, several mechanisms (e.g., stratification, learning, power of groups, mobilization, free-rider effects) are discussed as possible approaches to refine PFT.

Advocacy Coalition Framework (ACF): Delineating a Policy Mechanism Research Agenda?
In the mid-to-late 1980s, the progenitors of the Advocacy Coalition Framework (Sabatier and Jenkins-Smith, 1993) developed a framework to explain the effects of changes in external variables and system-level shocks, as well as new flows of scientific and technical information, on the behavior and strategies of contending advocacy coalitions (comprising individuals with similar beliefs and varying degrees of cohesion) on policy subsystems and policy change. Animated by threat and conflict, advocacy coalitions learn in order to develop new strategies and in response to new information and changing conditions, sometimes informed by more or less authoritative policy brokers. Like PET, it is a multi-level theoretical framework -focusing on specific policy domains, taking into account macro-level or system-wide influences, while rooted in assumptions of individual beliefs and policy preferences. Jenkins-Smith et al. (2018) refer to mechanisms, but in a gray box fashion or as "causal pathways," asserting effects and linkages rather than demonstrating them. The term mechanism is variously used to refer to belief systems "simplifying and interpreting the world" (p. 142), "heightened public and political attention, agenda change, and most importantly redistribution of coalition resources and opening and closing of policy venues" as enabling policy change (p. 145), internal and external shocks that "mediate the effect from external shocks" (p. 146), and learning (p. 157). However, the ways in which these effects might obtain are not specified.
The chapter points to an extensive empirical literature using the ACF, comprising cases, surveys, and experiments. What is striking, though, is the extent to which important processes are underspecified from a "policy mechanism" perspective. How do deep values, policy core, and secondary aspects of belief systems of individuals and advocacy coalitions change? How does Evert Lindquist and Adam Wellstead -9781788118194 Downloaded from Elgar Online at 05/06/2020 07:13:30AM via free access policy learning occur? How do policy brokers and forums further learning and changes of view? Such questions -and the list could go on -are not answered. But the authors act as if they appreciate this gap, with a long list of researchable topics. Many of these could easily be cast as seeking more detailed delineation of causal linkages at a finer-grained level along with empirical studies for understanding belief systems, learning, coalition dynamics, venues and forums, and conflict and role of science and policy analysis (see pp. 154-9).

Narrative Policy Framework (NPF): Embryonic and Component of Other Frameworks?
This theoretical framework is a relatively new addition to the TPP collection, featured in the last two editions. Situated, styled, and elaborated as a post-positive approach -as a contrast to the Multiple Streams, Punctuated Equilibrium, and Advocacy Coalition Frameworks -it focuses on the structure, uses, and impact of contending narratives on policy-making, largely from a poststructuralist orientation. The framework is still being elaborated but mimics many of the features of ACF theorizing and hypothesis development, including working at three levels (micro or individual, meso or sector, macro or system). Empirical studies rely on textual and interpretive analysis, surveys, social media data, and experiments. While empirical work has been carried out, a considerable amount of space is devoted to justifying the need for the framework and describing theoretical underpinnings. The chapter by Shanahan et al. (2018) refers to causal mechanisms in several places, but they seem to refer to internal elements or components of narratives, incentives or strategies of those using narratives, and different high-level pathways (pp. 177-8, 193) and empirical relationships (p. 186). Thus, NPF is ontologically interpretivist rather than causal. But these assertions of causal mechanisms fall short of what policy mechanism advocates call for in terms of delineating how variations in the components, incentives, or strategies achieve their effects, even if the effects were observed in certain empirical studies. Much of the literature's appeal that points to causal effects are, in fact, those theories from very different disciplines and subdisciplines (very much like ACF). Most of the theorizing and causal chains are not directly connected to policy change -there is no theory or stylized circumstances delineated concerning when new narratives, different kinds of narratives, or differently structured narratives might have more of an impact on policy outcomes or in combination with other variables.
Narratives are surely important as a political and policy tool, but even with deeper specification the potential of this approach to explain policy changes and outcomes could be overstated. It is surprising that NPF is not more explicitly linked to the ACF's belief systems that define and structure advocacy coalitions. Indeed, evolving narratives could be an important indicator and means of policy-oriented learning, as well as how they become adapted in response to external and internal perturbations and new scientific and technical information, and even more closely linked to the "policy images" associated with PET and policy monopolies. This points to the overlap and interlocking quality of many of the frameworks and their components.

Institutional Analysis and Development (IAD): A Broad Framework in Need of Mechanisms?
The IAD is largely a highly context-specific evaluative framework that has been extensively used to analyze and address collective action problems, self-governance arrangements, and the management of shared resources. All governance arrangements, regardless of their scale, are determined by institutional rules. Specifically, institutions are defined as sets of rules used by individuals to organize repetitive activities. This approach involves individuals seeking agreements with each other that could be enshrined in a set of meaningful rules. At the center of the IAD is the "action situation" where various actors interact to produce various patterns of interaction, which in turn leads to various outcomes. Ostrom identifies seven "rules-in-use" (boundary, position, choice, payoff, scope, aggregation, and information) that are exogenous, context-specific, and affect or structure action situations (Schlager and Cox, 2018, pp. 219-221).
This chapter by Schlager and Cox (2018) has several mechanism-type references (e.g., design principles identifying institutional variables, congruence between appropriation and provision rules and local conditions, low-cost conflict resolution mechanisms interacting to support users of common pool resources). Ostrom states that an analyst must make assumptions about how and what participants value, their information-processing capabilities and internal mechanisms they use to decide upon strategies. Moreover, there are implicit causal assumptions within and between operational, collective, and constitutional "situations." Economic theory, game theory, transaction-cost theory, social choice theory, theories of public goods and common-pool resources are all compatible with the IAD and lend themselves to mechanism-based research. The IAD is, by design, by far the broadest framework that has allowed scholars from fields beyond public policy to address collective action problems. Given its portability, IAD policy scholars might be well served by borrowing policy-specific mechanisms developed in other policy process frameworks.

Innovation and Diffusion Model (IDM): Ontological and Methodological Reconsiderations
The adoption of new programs by governments according to the IDM approach is explained by "internal determinants" and diffusion models. The internal determinants model postulates that political, economic, or social causal factors leading a jurisdiction to innovate are internal to the state. A number of internal determinant hypotheses are illustrated, including examples from organizational theory (e.g., resource slack). Diffusion models examine the adoptions of policies as emulations of previous adoptions by other governments. There are three contexts: when governments learn from one another by borrowing successful innovations from other jurisdictions, compete with others, or respond to public pressure to adopt policies. In their chapter, Stokes Berry and Stokes (2018) identify five alternative mechanisms for diffusion: learning, imitation, normative pressure, competition, and coercion. They draw attention to multiple mechanisms, which can change over time due to contextual factors such as the characteristic of the state. The focus of IDM research, however, has been on developing three testable models: national interaction, regional diffusion, and the leader-laggard. Unlike the other TPP approaches that rely on case studies, its empirical insights have largely come from probabilistic regression analysis from national and regional datasets. Despite operationalizing independent and dependent variables employed in their models, Stokes Berry and Berry (2018, p. 282) note that there has been inattention to isolating the mechanisms for a policy's diffusion. Quite forcefully, they advocate a shift to a mechanisms approach: A key challenge for researchers in the decade ahead is to continue to conduct empirical research seeking to identify the mechanism(s) underlying policy diffusion but to focus attention on developing better empirical indicators for the presence of specific mechanisms. For each diffusion mechanism -learning, imitation, competition, coercion, and normative pressure -the goal should be to construct indicators for the presence of the mechanism that can not only successfully detect this mechanism when it is present (avoiding false negatives) but also fail to detect the mechanism when it is not present (avoiding false positives). (2018) acknowledge that for policy process theorists to shift from a general process framework to a mechanistic approach it requires disaggregation and greater specificity of mechanisms. Moreover, they recognize the challenge of empirically identifying mechanisms.

MECHANISMS AND POLICY FRAMEWORKS: TAKE-UP TO DATE AND POTENTIAL
What follows looks across the high-level findings presented in the previous section. We consider to what extent, as covered in the latest TTP collection, the policy process studies field has embraced mechanisms to guide theoretical and empirical work. We consider the challenge of theorizing about complex policy systems, when that very complexity and reality bandwidth pushes scholars to embrace more rather than fewer variables and causal factors and pathways. This also leads to considerable overlap across these frameworks, and we suggest that mechanism perspectives could further and usefully link parallel investments in theorizing more deeply and precisely within and across these process frameworks.

Policy Mechanism Nomenclature and Specification Not Yet Evident in TTP
Efforts to apply mechanisms has arrived selectively in certain policy journals (Policy Studies Journal, Journal of European Public Policy). Our review of the chapters in the most recent edition of TPP suggests that the policy mechanisms movement has yet to establish a strong beachhead across the waterfront of policy process theorizing, and, despite the advocacy of progenitors such as Sabatier for more causal elaboration, no connection is made with the mechanism approach as a cross-cutting guide to the next generation of theorizing. To be sure, some TPP authors do invoke the term "mechanism" but typically not in ways consistent with the spirit and approaches suggested by the policy mechanism movement. This reflects the state of policy process theorizing and empirical work that, although it has made enormous strides over the last 30 or 40 years, proceeds at a fairly high level of abstraction and aggregation. Our sense, though, is that many of these frameworks -while intuitive at the system or macro level, and typically designed to study the dynamics of policy domains at the meso level -have less empirical traction with respect to specific cases at the micro level. Whether for the purpose of more detailed explanation of policy processes or, instead, moving towards a design orientation, we believe that the next wave of theoretical elaboration, along with empirical research, will require finer-grained theories that move from identifying higher-level processes, functions, and causal pathways to more precise delineation of causal connections and conditions under which they obtain.

Some Policy Theory Frameworks Lean Towards Policy Mechanism Approaches
Some chapters, while not using policy mechanism nomenclature to assess the causal depth and richness of the theoretical framework in question, nevertheless seem disposed to and rely on approaches congruent with a mechanism perspective (e.g., PET). In other words, taking up the challenge put on the table by the policy mechanisms movement ought to be consistent with many of the frameworks. For example, the hypothesis-testing posture of other traditions (ACT, NPF, etc.) suggests that, regardless of the empirical methods employed by their colleagues, they would embrace the key messages of PMM and more precisely elaborate how mechanisms work rather than assert the dynamics and effects (likewise, with regard to PET they might start using other methods rather than stochastic approaches). We think this will be an exciting agenda because, in part, to give credit where it is certainly due, the progenitors and contributors to the literature associated with each framework have laid important foundations as a point of departure. The essential research questions and accompanying distinctive theoretical approaches will neither be diluted nor go away; rather, they stand to be strengthened. Table 2.3 represents our first cut effort to identify what might constitute the mechanisms, working at different levels of analysis, for each policy theoretic framework. It provides an overview of the broad characteristics of the mechanisms in seven frameworks featured in TPP (2018). However, to further policy mechanisms research, issues of mechanism specificity and disaggregation discussed by Falleti and Lynch (2008) need to examined.

The Challenge of Theorizing About Policy Systems Across Levels of Analysis
In reviewing the chapters comprising TPP, we were constantly reminded of the challenges of theorizing across levels of analysis of complex systems (micro, meso, or macro perspectives; or, put differently, the individual, policy subsystem, and system levels of analysis). First, each of the traditions, in vary degrees, point to causal pathways moving up, down, and across these levels of analysis. Second, many of the frameworks comprise different combinations of theories and bundles of assumptions assembled to factor in local or broader environments. This is similar to inserting pre-built subroutines or component parts, but for mechanism seekers they still function as opaque "black box" or at best "gray box" convertors. These multiple levels of analysis and reliance on theoretical components make it difficult to ascertain what is presumed and what actually gets measured, tested, and proven.
We find this state of affairs intriguing because each of the frameworks are attractive, plausible, and compelling. However, when one scratches the theoret- Differentiation between national and regional innovation, but that is about it -no micro level of analysis ical surface in a process-tracing manner (regardless of whether those working in the field rely on quantitative and qualitative methods), the evidence appears increasingly more tenuous and the assumptions loom larger. This suggests that we do not have strong foundations for even rough order-of-magnitude predictions of when certain outcomes might obtain, what the relative impact or explanatory value of different variables or causal pathways or functions might be, or what ought to be the crucial design elements for good policy, process or governance designs.

Implications of Mechanism Approaches for Significantly Overlapping Frameworks
Our review of TPP reinforced an obvious point: each theoretical framework is animated by distinctly different research questions, although they often share many of the same variables and causal links, which most contributors acknowledge in varying degrees. The precise explanatory focus means that researchers will identify different variables or effects as dependent, intervening, independent, or more proximate and material. However, when all the frameworks are lined up together, the extent of overlap is considerable. It is difficult to keep track of all the frameworks, let alone fully appreciate where one starts and the others leave off, particularly since, as noted above, many of the frameworks import or point to similar components from other frameworks. This state of affairs can lead to confusion or the perception that the frameworks agree in the main about many things, but disagree or diverge in order to address particular dynamics and issues in policy systems. But bringing a mechanistic perspective to bear on this challenge raises the intriguing possibility that delineating causal chains and animating conditions for one policy framework might lead to progress in another. Likewise, it suggests that the perhaps difficult decisions of scholars and graduate students to invest time working with one framework and associated methodological approaches may not constitute a pure trade-off: that continuing work across the framework domains may lead to increasing theoretical returns. While not suggesting the prospect of a "unified policy process field theory" (because what variables a scholar invokes depends on the research questions asked), we are suggesting that much more progress can be made in thinking about the equivalent to the causal "wiring" or vascular network fanning out vertically, horizontally, and diagonally in the representations we have created to capture policy systems. Deepening our understanding of these connections with mechanism approaches may well sharpen our appreciation of what each framework does and cannot do.

INVIGORATING POLICY PROCESS INQUIRY: IMPLICATIONS FOR RESEARCH, DESIGN, PEDAGOGY
In our first endnote of this chapter, we alluded to an "end of policy theorizing" hypothesis, which suggests that the leading theories of public policy theories have attained a level of maturity, while greatly respecting the groundwork that has been laid. There is considerable agreement on what constitutes the most plausible frameworks invoked and relied on by established and new scholars in thousands of articles and dissertations around the world. The TPP collections have provided a useful compilation, crossroads, and appraisal of the state of theorizing and empirical work associated with each of these traditions, adding new ones as they gather momentum. However, these frameworks have considerable overlap in terms of identifying the variables, components, and functions at play in policy-making systems. At one level this is reassuring, and at another level, concerning, because it seems increasingly difficult to distinguish where one framework ends and another begins. This chapter suggests that, from a policy mechanism perspective, the frameworks as articulated in the TPP4 collection have, going forward, different levels of maturity and promise. Many authors seem more interested in documenting the take-up and citation rates of the tradition they have invested in than more carefully appraising whether they have become theoretically better developed, and whether and how empirical studies could test and deepen theoretical assumptions and propositions.
At the end of TPP4 (2018), Weible (2018) asks how scholars can improve the quality of theoretical and empirical work. He suggests that more attention should be directed to empirical studies and testing, and less to more theory and propositions. We think that delineating policy mechanisms that link theoretical assumptions and empirical investigation offers an exciting opportunity to infuse policy process theorizing just as it has become a bit staid and reified, and yet it does not purport to challenge what each of the traditions seeks to accomplish. Focusing on mechanisms may assist those working with different TPP traditions: a more granular approach can encourage further parsing out and deepening theories associated with broad frameworks with the goal of identifying more specific empirically grounded studies. Even more exciting is the possibility that, by working in more detail with policy mechanism thinking and process testing, policy scholars can better work across the frameworks to develop theoretical and empirical insight into how their components link together, potentially enabling progress in different functional domains to buttress development in other domains. Indeed, the policy mechanism approach will not rival established and emerging policy process frameworks -rather, it collectively challenges them and points to a way forward.
Our study has reviewed the latest version of the TPP collection, but this is not a survey of the scholarship associated with each theoretical framework. We need deeper reviews of the empirical literature associated with each of the seven frameworks reviewed in the TPP to more definitively ascertain the state of the art (i.e., to what extent is mechanism thinking used explicitly or implicitly by scholars?). This would be no small task, and could perhaps focus on exemplar studies that might serve as benchmarks of sorts, pointing to new avenues for empirical and theoretical research. This is a significant research agenda, engaging all TPP traditions, but would provide spillover benefits due to the overlap in theories.
Finally, given the design orientation of this collection, we would be remiss not to consider the promise of policy-mechanistic perspectives for design practice and pedagogy. Seeking out policy mechanisms, and the conditions under which they obtain, will not just be an exercise in theory specification and testing; it should also lead to better understanding of the circumstances and conditions under which initiatives have promised or predicted effects, always allowing for the complexity, inertia, and multifaceted nature of policy-making and governance. We also see important pedagogical implications: teaching students about how the policy process works, how to navigate it, and how to undertake policy analysis, has always been a design enterprise, even if scholars working in different traditions variously cast this as "craft" or "science." There is great potential for the bottom-up practical design thinking to meet on empirical middle ground (at a variety of different policy domain sites) the top-down theorizing from multiple frameworks. These frameworks have always sought to capture the complexity and essential features of policy-making as context and something to be shaped, but have not always reached their full potential as reliable diagnostic and design tools. We think that the research and theoretical research agenda stimulated by policy mechanism thinking can bridge these divides. NOTES 1. This does not quite amount to the "end of policy theorizing," but the stability in the number of approaches given the complexity and diversity of policy challenges and governance arrangements is worthy of note. 2. A "second cut" has proceeded on a collaborative basis with our colleagues Jeroen van der Heijden and Johanna Schulman (Van der Heijden et al., forthcoming), based on a systematic review of literature using MSF, ACF, PET, NPF and IAD. 3. This terminology will clarify things later when we evaluate how TPP authors use "mechanism." 4. Since the time of presenting this work, Howlett and Capano (in the current volume) have made a distinction between "first-order" and "second-order" mechanisms, which we do not take up in this chapter. 5. The TPP collections have not been the only efforts to compare and contrast leading theoretical frameworks on policy-making. The first significant overview was Schlager and Blomquist's (1996) review paper comparing three "emerging theories of the policy process." It examined the ACF, Elinor Ostrom's Institutional Rational Choice (IRC, which later became IAD), and Terry Moe's politics of structural choice approach. Issues of mechanisms and causality were briefly discussed, with ACF touted as a more sophisticated incorporation of the roles of information and learning; it challenges the other frameworks to consider the "ideological filtering of information, and changes in individuals' beliefs, as mechanisms promoting or inhibiting policy change" (p. 666).