Policy Change: An Advocacy Coalition Framework Perspective Notes

Policy Change: An Advocacy Coalition Framework Perspective

Introduction

The Advocacy Coalition Framework (ACF) seeks to explain policy change by analyzing 148 policy processes from 67 journal articles published between 2007 and 2014. The study identifies multiple pathways to policy change, examines the use of ACF concepts, and compares subsystems, especially in the environment and energy domain. It also addresses the need for clarity and transparency in applications of the ACF, explores associations between primary pathways and policy domains, as well as connections between primary pathways and secondary components. It further covers policy change and stasis, while pinpointing threats to the internal validity of core ACF concepts.

Policy process research aims to understand policy change and stasis. Frameworks and theories have been developed to explore these phenomena. One such framework is the Advocacy Coalition Framework (ACF). The purpose of policy process research and the ACF is to consider research implications beyond knowledge generation. ACF research should be clear and relevant for specific contexts, balancing the testing of concepts and theory with the understanding of particular contexts. This article explores this tension. It seeks to aid researchers and practitioners in understanding the conditions for policy change, expand upon past meta-analyses of the ACF, and identify areas for future research and theoretical development.

ACF and Policy Change

The ACF was created for diverse examination of policy foci to encourage comparability, replicability, and falsification. It is used to understand policy processes in various geographic and politically diverse areas. It models public policy as a conflict between actors competing to translate their beliefs into policy. It is useful for examining cases where conflicting goals and technical or scientific information influence policy processes.

The ACF is a framework that includes three theories: advocacy coalitions, policy-oriented learning, and policy change. This article focuses on the theory of policy change. Policy change reflects winning advocacy coalitions’ policy beliefs. ACF associates major policy change with alterations in policy core beliefs and minor policy change with alterations in secondary beliefs.

Primary Pathways to Policy Change

There are four primary pathways associated with bottom-up policy change in the ACF:

1. External perturbations or external events to the subsystem.

2. Internal events to the subsystem.

3. Policy learning.

4. Negotiated agreements.

These form a single hypothesis of necessary, but not sufficient, conditions for policy change. A fifth top-down primary pathway to policy change is change imposed by a hierarchically superior jurisdiction.

Policy Change Hypotheses:

• Policy Change Hypothesis 1, Bottom-Up Policy Change: Significant perturbations external to the subsystem, a significant perturbation internal to the subsystem, policy-oriented learning, negotiated agreement, or some combination thereof are necessary, but not sufficient, sources of change in the policy core attributes of a governmental program.

• Policy Change Hypothesis 2, Top-Down Policy Change: The policy core attributes of a government program in a specific jurisdiction will not be significantly revised as long as the subsystem advocacy coalition that instated the program remains in power within that jurisdiction—except when the change is imposed by a hierarchically superior jurisdiction.

Primary pathways may occur simultaneously or in sequence, leading to a policy change.

External Events

External events are a fundamental component of the ACF and a primary pathway to policy change. The policy subsystem exists in an ecosystem of relatively stable parameters (e.g., basic attributes of the problem area, distribution of natural resources, sociocultural values, social structure, and constitutional structure) and external subsystem events (e.g., changes in socioeconomic conditions, public opinion, systemic governing coalition, and other policy subsystems). Stable parameters influence both external events and long-term opportunity structures (e.g., degree of consensus needed, openness of political system, and societal cleavages). Opportunity structures also influence external events. Long-term opportunity structures and external events are mediated by short-term constraints and resources, offering opportunities for coalition exploitation, similar to Kingdon’s windows of opportunity. The external event pathway involves many intervening and mediating steps, making it a necessary yet insufficient pathway to policy change.

Superior Jurisdiction

Superior jurisdiction is another primary pathway to policy change, which may instigate new politics and another policy process within the policy subsystem or by becoming an external event for another policy subsystem. Like external events, superior jurisdictions may or may not change policy but are a function of stable parameters, long-term and short opportunity structures, and external events.

Internal Events

Important activity also takes place within the policy subsystem, where actors are simplified into coalitions with attributes like hierarchical beliefs, resources, strategies, and coordination, which they employ in their attempts to make policy. Policy change may result from internal subsystem events linked to these attributes and coalitional behaviors. Internal events include crises, policy failures or fiascos, and scandals. Such internal events provide an opportunity for the attributes of coalitions to change or be confirmed, especially strategies or beliefs.

Learning and Negotiation

Learning and negotiation are also primary pathways to policy change. Learning may occur between or within coalitions and may subsequently lead to policy change. Negotiations may also occur between two or more coalitions that may lead to policy change. Negotiations are more likely to occur when coalitions recognize the existence of a hurting stalemate initiated by a policy broker.

Secondary Components of Policy Change

In addition to the five primary pathways of policy change, ACF scholars note a variety of secondary components that may influence policy change. These secondary components are intermediate variables and causal mechanisms associated with one or multiple primary pathways to policy change. The specific mechanism by which secondary components alter policy is not clearly ensconced within framework hypotheses in comparison to the primary pathways.

Secondary components of policy change identified in the ACF include:

1. New dominant coalition.

2. Change in the distribution of resources.

3. Venue opening or closing.

4. Minority coalition mobilization.

5. Dominant coalition belief change.

6. Minority coalition belief changes.

7. Dominant coalition belief confirmation.

8. Minority coalition belief confirmation.

9. Dominant coalition strategy change.

10. Minority coalition strategy change.

11. A hurting stalemate.

12. Presence of a policy broker.

Some of these secondary components are ripe for further theoretical specification and empirical analysis, such as the policy broker. This study excludes two additional secondary components: heightened public and political attention and changes in the government agenda, as these are assumed to be present for policy changes.

Methods

Following Weible et al. (2009), a list of peer-reviewed journal articles applying the ACF was developed. This review follows more recent studies and generated a list of peer-reviewed journal articles citing at least one of the six foundational documents developing the ACF using the Web of Science database. This review was limited to English peer-reviewed journal articles published between 2007 and 2014, producing an initial data set of 1,067 peer-reviewed articles.

Content analysis of these articles proceeded in two rounds. First, five coders recorded the bibliographic information of each article. Four codes were utilized to determine whether an article was an application of the ACF. These codes identified keywords (coalition, learning, or advocacy) in the title, abstract, and keywords sections, and whether or not theoretical foundation citations were used at least twice. This practice identified 512 articles. However, relying only on the frequency of keywords and citations may lead to Type I errors. In addition to keyword and reference searches, coders examined articles to determine if they include data and/or a case study about a topic, and identify and analyze at least one theory of the ACF (coalitions, policy change, and/or learning). This process narrowed the pool to 161 articles identified as ACF applications. To mitigate subjectivity, intercoder reliability assessments for this coding were conducted on a random sample of 256 articles being reviewed by an intercoder.

In the second round of coding, seven coders analyzed the remaining 161 articles to identify if they focused on theory of advocacy coalitions, policy-oriented learning, or policy change. In this round, 87 articles or 54 percent were subject to inter-rater reliability. This led to the identification of 67 articles that focus on the theory of policy change. A third round of coding was conducted on these 67 articles to identify if they identified policy change or stasis, the policy identified, primary pathways, secondary components, and other theory-based codes using a detailed codebook. Of the 67 articles identified as focusing on the theory of policy change, inter-rater reliability was conducted on a random sample of 39 articles, or 58 percent of the sample. The codebook includes 35 codes.

Results

Policy Change

In total, the 67 articles identify 148 different policy processes using the ACF. Among those policy processes, 131 identify policy change and 17 identify policy stasis. Twenty-eight articles (42 percent) analyze multiple policies, among these the range is from 2 to 11 policies. Examples of policies analyzed include: UK badger culling; land use policy in Wisconsin; EU satellite programing; street sign language in Czech Republic, Hungary, and Wales (UK); and German defense policy and the War in Afghanistan. Further analysis of the 28 articles identifying multiple policies reveals 17 articles comparing policies across subsystems. In contrast, 11 articles compare multiple policies over time within a single subsystem. Major and minor policy changes are identified by a minority of applications. Among the 131 cases of policy change, major policy change is identified 28 times and minor policy change 20 times for a total of 48 (37 percent) policy changes. In total, 19 of 67 (28 percent) articles identify either major or minor policy changes.

According to Sabatier and Jenkins-Smith (1999), external events are often associated with major policy changes while policy-oriented learning is often associated with minor policy changes. There is support for these propositions as 64 percent (18) of major policy changes are associated with external events, and 60 percent (12) of minor policy changes are associated with policy-oriented learning.

Primary Pathways to Policy Change

The ACF posits that one or multiple of the following primary pathways to policy change are necessary, but not sufficient to cause policy change: (1) external events, (2) internal events, (3) policy-oriented learning, (4) negotiated agreement, and (5) changes imposed by a superior jurisdiction. These primary pathways may occur in combination with each other or in isolation. An examination of the 131 cases of policy change reveals eight (6 percent) do not identify any primary pathways to policy change. Among the 131 cases of policy change, 52 (40 percent) have only a single primary pathway, and 71 (54 percent) have multiple primary pathways to policy change. The most frequently isolated primary pathways are external events (20) and learning (16). Negotiation (13) as a single primary pathway occurs almost as frequently as learning. Studying policy change as a function of only superior jurisdiction (2) or internal events (1) are relatively rare, as these tend to be applied in combination with other primary pathways.

Further examination of the 71 policy changes identifying multiple primary pathways reveals associations. External events are frequently identified in association with other primary pathways, in fact 74 percent (58/78) of the time that they are identified it is in conjunction with other primary pathways. The majority of external event occurrences are identified with learning (46/78). External events are also identified with negotiation (18/78), superior jurisdiction (15/78), and internal events (9/78). Overall, external events are identified in 60 percent (78/131) of all policy changes, and 82 percent (58/71) among applications that identify multiple pathways to policy change.

Learning occurs in combination with other primary pathways 78 percent (57) of the time almost as frequently as external events (58). Of all applications of learning, it is identified with external events 63 percent (46/73) of the time. Learning associated with policy change appears in limited frequency in combination with negotiation (22/73) and superior jurisdiction (13/73). Learning is rarely identified with internal events (3/73). Overall, learning is identified in 56 percent (73/131) of all policy changes, and 80 percent (57/71) among applications that identify multiple pathways to policy change.

Negotiation is identified as a major pathway to policy change in 31 percent (41/131) of all policy changes. Similar to external events and learning, most of the time (68 percent or 28/41) negotiation is identified it is in conjunction with another primary pathway to policy change. A majority of the time that negotiation is identified as a primary pathway to policy change it is in conjunction with learning (22/41). It is also frequently identified with external events (18/41). Negotiation is identified only three times with superior jurisdiction or internal events. Overall, there may be an association between negotiation and learning, as well as a relationship with external events in terms of policy change.

Superior jurisdiction is identified 18 times with other primary pathways to policy change, but only twice as a single pathway to policy change. A vast majority of the time superior jurisdiction is identified with external events (15/20). Superior jurisdiction also frequently occurs with learning (13/20), but infrequently occurs with negotiation (3/20) or internal events (3/20) leading to policy change.

Internal events are identified just once as the only pathway to policy change. Like superior jurisdiction, internal events are almost always identified in conjunction with other primary pathways to policy change (12/13). Internal events are identified with external events a majority of the time (8/13), but are less frequently identified with learning (3), negotiation (3), and with superior jurisdiction (4).

Policy Domains and Primary Pathways

Among the 131 cases of policy change, the most frequent policy domain is environment and energy (54) and this represents a plurality of applications studying policy change. Other policy domains that identify a policy change include education (16), urban planning and transportation (13), economics and finance (12), social welfare (10), public health (9), emergency management (5), science and technology (4), foreign and defense policy (3), immigration (3), and media (2).

Associating the major pathways and policy change with policy domains reveals multiple trends. External events are always associated with policy change among emergency management (5/5). In addition, external events are found to be associated with at least half of the policy changes in urban planning and transportation (10/13), social welfare (7/10), education (11/16), environment and energy (29/54), and economics and finance (6/12). Overall, external events are most frequently identified among environment and energy policy changes, but occur with the greatest rate among emergency management policy changes.

Learning is identified among at least half of policy changes associated with environment and energy (34/54), education (8/16), public health (7/9), social welfare (6/ 10), and emergency management (3/5). Learning is most frequently identified with environment and energy, but occurs with the greatest rate among public health policy changes.

The combination of External Events (\land)} learning is identified a total of 46 times among all policy changes. This pattern occurs most frequently among environment and energy (20/54), education (7/16), social welfare (4/10), and emergency management (3/5). However, the pattern of external events (\land)} learning is not frequently identified among public health (2/9), economics and finance (3/12), or urban planning and transportation (2/13).

Negotiation is identified among a majority of policy changes within the domains of economics and finance (8/12), emergency management (3/5), and public health (5/9). However, negotiation is most frequently associated with environment and energy policy changes (17/54). A minority of the applications about education (3/ 16), urban planning (2/13), transportation (2/13), and social welfare (1/10) identify negotiation as a pathway to policy change. Negotiation is frequently identified either with external events or learning, or with both external events and learning. The combination of negotiation and external events is identified 18 times, and the combination of negotiation and learning is identified 22 times. The pattern of the three primary pathways of negotiation (\land)} external event (\land)} learning is identified a total of 13 times. In combination with learning, negotiation is most frequently associated with economics and finance (4/12) and public health (4/9). With external events, negotiation tends to be applied to emergency management (3/5), and with external events and learning, negotiation is most frequently identified with environment and energy (9/54). This demonstrates that negotiation is a pathway to policy change commonly identified across a wide range of policy domains.

Superior jurisdiction is not identified among a majority of any policy domains that identify policy change at least five times. It is most frequently identified in the environment and energy domain (11/54). Among all other policy domains, superior jurisdiction occurs two times or fewer. The pattern of superior jurisdiction (\land)} external event occurs a total of 15 times, and nine of these are environment and energy applications. The pattern of three primary pathways of superior jurisdiction (\land)} external event (\land)} learning is identified in seven environment and energy applications, meaning that the combination is identified 64 percent (7/11) of the time that superior jurisdiction is identified among environment and energy applications. This demonstrates that superior jurisdiction is most commonly identified in the environment and energy domain and frequently in combination with an external event and/or learning.

Internal events are the least frequently identified major pathway, and are most frequently associated with environment and energy (5/54), but are identified at the highest rate among economics and finance (4/12). The pattern of internal events with external events are identified eight times and appear in a wide range of policy domains including: environment and energy (2), economics and finance (2), public health (1), urban planning and transportation (1), emergency management (1), and foreign and defense policy (1).

Primary Pathways and Secondary Components

Secondary components are identified in the literature as sometimes being involved with policy change, either as a causal mechanism or intermediate variable, but are not explicit in policy change hypotheses. These concepts are important to ACF policy change studies and at least one is identified among 91 percent (119/131) of policy changes, which is almost as frequent as the identification of primary pathways (123/131). The most common and only secondary component identified in a majority of policy changes is minority coalition mobilization (81/131). The secondary components that occur at least among 20 percent of the policy changes are: dominant coalition belief confirmation (32/131) and dominant coalition belief change (29/131), opening or closing of venues (29/131), and change in the distribution of resources (27/131).

There are various patterns of association between secondary components and primary pathways to policy change. Among the external event primary pathway, the most frequent secondary component is minority coalition mobilization (44/78). Changes in the distribution of resources (20/78) or the opening or closing of venues (20/78) are only identified in about a quarter of policy changes associated with external events. However, these two components are the most frequently identified in association with external events after minority coalition mobilization.

The primary pathway of learning is associated with the altering of concepts and assumptions in relation to policy actors in a subsystem. Among policy changes associated with the primary pathway of learning, the most frequent secondary component is minority coalition mobilization (44/73). Changes in beliefs among the dominant coalition (24/73) and the minority coalition (12/73), and changes in strategy among the dominant coalition (11/73) and the minority coalition (18/73) are all identified among a minority of policy changes. However, belief changes among the dominant coalition (24/73) is the most frequent secondary component associated with learning after minority coalition mobilization. Other possible associations of interest with learning are the opening or closing of venues (21/73) and the identification of a policy broker (18/73).

Within the negotiation primary pathway, there is a hurting stalemate and a lack of alternative venues. The opening and closing of venues is identified among 15 percent (6/41) of negotiation primary pathway cases, and hurting stalemate and policy broker are both identified among 29 percent (12/41) of cases. Both hurting stalemate and policy broker are the most frequently identified secondary components in association with negotiation after minority coalition mobilization (31/41).

The theoretical literature identifying secondary components does not associate them with a superior jurisdiction leading to a policy change. Despite this, the secondary components most frequently associated with superior jurisdiction in practice include minority coalition mobilization (13/20), opening or closing of venues (9/20), and a new dominant coalition (8/20). The rarest primary pathway, internal events (13), exhibits the highest frequency of secondary components.

Overall, there is some support for changes in strategies and beliefs, or confirmation in association with internal events. Overall, outside of minority coalition mobilization, secondary components are generally identified among a minority of policy changes. This makes general statements about associations with primary pathways difficult. However, there is limited support for associations between external events and the opening or closing of venues (20/78) and change in the distribution of resources (20/78); learning and belief change among the dominant coalition (24/73) and strategy change among the minority coalition (18/73); negotiation and the identification of a hurting stalemate (12/41) and policy broker (12/41); superior jurisdiction and the opening or closing of venues (9/20) and new dominant coalition (8/20); and internal events and the opening or closing of venues (6/13) and change in the distribution of resources (6/13). Each of these associations is based on the theoretical literature and should be areas for future research.

Policy Stasis

In addition to the 131 cases of policy change, there are 17 cases of policy stasis among 13 articles. Policy stasis has also not been explicitly explored in previous ACF reviews. While infrequent, these cases reveal important observations. Policy stasis as identified by ACF scholars studying policy change accounts for only 11 percent (17/148) of all policy processes. These studies exemplify how policy stasis in the ACF can be studied.

Policy stasis applications also identify primary pathways and secondary components that could have led to policy change. In total, five policy stasis applications do not identify a primary pathway, but 12 identify at least one primary pathway to policy change. Of those, half identify a single primary pathway (6) and half multiple primary pathways (6). The most frequent primary pathways that did not lead to a policy change are external events (10/17), followed by learning (7/17). These are also the most frequent primary pathways to policy change. For policy stasis, negotiation is identified twice, and internal events are only identified once. The pathway of a superior jurisdiction imposing policy change is not associated with policy stasis.

Examining the relationship between policy domain and policy stasis is similar to policy domain and policy change, except for foreign and defense policy. Environment and energy (9/17) is the most frequent domain identified, but other domains including foreign and defense (4/17), public health (2/17), and education (2/17) are associated with policy stasis. Within the environment and energy domain articles, external events (8/9) and learning (6/9) are associated with policy stasis a majority of the time. In addition, half of the applications in relation to foreign and defense (2/ 4) identify external events with policy stasis. Overall, policy stasis accounts for 14 percent (9/63) of all environment and energy applications and is the most frequent. However, the highest rate of policy stasis is 57 percent (4/7) among foreign and defense.

Belief confirmation appears to be an obstacle to policy change whether among the dominant coalition (10/17) or the minority coalition (6/17). However, after minority coalition mobilization, belief confirmation among the dominant coalition is the most frequent secondary component among policy changes as well. Other secondary components occurring frequently include minority coalition mobilization (8/17), opening or closing of venues (6/17), and changes in distribution of resources (4/17). The remaining secondary components are only identified two times or fewer in association with policy stasis. In fact, several secondary components are never associated with policy stasis. They include new dominant coalition, dominant coalition belief change, minority coalition strategy change, hurting stalemate, and policy broker. This may indicate that these secondary components are associated with policy change but not with policy stasis.

Discussion

This article analyzes how the ACF is used to study policy change in 67 articles exploring 148 policy processes. This review differs from previous ones by exploring each policy process resulting in either change or stasis as a unit of analysis. This systematic focus on theoretical concepts has allowed for the identification of associations between primary pathways, policy domains, and secondary components in relation to multiple policy changes and stasis in the same article.

According to Sabatier and Jenkins-Smith (1999), a major strength of the ACF is a “clear-cut criterion” for distinguishing between major and minor policy changes. This distinction has been perpetuated in theoretical works about the ACF, however, the vast majority of articles do not distinguish major or minor policy change. Future applications of policy change should be clearer about identifying major and minor policy change as this might help uncover more general theoretical relationships between ACF variables, as Weible et al. (2009, p. 128) suggest. It may also be an indication of the necessity for further conceptual development about minor and major policy changes.

Based on this study, most policy changes and even most policy stasis applications identify multiple primary pathways. However, a major limitation for identifying these primary pathways is the permeability of boundaries and transformative features of policy subsystems, in particular those at the national level. About half of all ACF applications since 2007 are at the national level. This presents an obstacle for identifying the scope of subsystems. For the study of policy change at a national level, the ACF may have obstacles and limitations that need further research and theoretical development.

A limitation of ACF applications is the identification of secondary components in association with primary pathways to policy change or stasis. In theory, external events are associated with changes in the distribution of resources and the opening or closing of venues. These are the most frequent secondary components identified after minority coalition mobilization, but they are identified in only about 25 percent of the policy processes that external events are associated with policy change. The other primary pathways have similar issues. Overall, the lack of secondary components being explicitly identified with primary pathways and policy change or stasis is an obstacle to making general statements about their utility and possible necessity and is a challenge to comparability. Future applications should explicitly identify and state if the secondary components are present or absent, especially those that the ACF literature proposes are associated with a given primary pathway, such as change in distribution of resources with external events.

Conceptual validity is limited if the ACF seeks to be both empirically comparable and flexible enough for unique case specificity. For example, the ACF identifies public opinion as both a type of external event and as a resource for advocacy coalitions. Multiple secondary components are also problematic. The opening or closing of venues is identified as a secondary component, but may also be considered a superior jurisdiction. Another major issue is the identification of dominant and minority coalitions.

The ACF is primarily applied to understand environment and energy subsystems, as well as policy changes and stasis. This has both benefits and burdens. While this background may raise questions about the external validity of the theory to explain policy change in other policy domains, this research shows the ACF has been applied to explain policy change across 10 different policy domains. However, applications of the ACF to these policy domains are relatively infrequent, making comparison and generalization difficult. Future research should continue to expand the ACF study of policy change beyond the realm of environment and energy.

A limitation of many of the policy processes examined is the absence of timing and sequencing. Using cross-sectional data does not take into account the timing and sequence of variables. The ACF in general would benefit from greater theoretical development and empirical investigation into the timing and sequencing of the primary pathways and secondary components. For example, what is the sequence between learning and negotiation, between an external event and a superior jurisdiction? This ambiguity leaves many unanswered questions regarding policy change. ACF scholars studying policy change should consider event history analysis, process tracing, and other methods that include time and sequence as variables.

Conclusion

This article examines how the ACF is being applied in practice to study policy change. In particular, it addresses a central struggle within the ACF—the simultaneous need for clarity and structure to foster comparable empirical results and the need for the framework to be flexible enough to accommodate a variety of foci in particular contexts. Content analysis was conducted to identify the frequency and associations between ACF policy change primary pathways, policy domains, and secondary components. The most frequent policy change primary pathways are external events and learning. A majority of applications explaining policy change identify multiple primary pathways. A plurality of applications study environment and energy, but over 10 different policy domains are identified. There is evidence that some primary pathways to policy change are more frequently associated with certain policy domains and not others. Secondary components are infrequently identified, and some, such as those in relation to dominant and minority coalition, may have internal validity problems. In order to increase comparability scholars need to more frequently and explicitly use ACF concepts, and ACF scholars need to further develop and better define ACF concepts to better ensure their applicability to various contexts.

One way the ACF can be a better tool for practitioners is to help them better understand the necessary and sufficient conditions for policy change including the sequencing and timing of pathways and secondary components. This means explicit identification of the presence of primary pathways and secondary components during both policy change and stasis. In addition, scholars should expand the study of policy stasis. By expanding the number of policy stasis applications along with cases of policy change, ACF scholars can better understand necessary and sufficient conditions. Researching policy stasis and greater explicit identification of secondary components will help scholars better understand the processes that lead to policy change, and better serve those advocating for change.