Skip to main content
Sign In

University of Colorado Denver

School of Public Affairs

Advocacy Coalition Framework Overview

The ACF rests on five foundational premises (Sabatier and Jenkins-Smith, 1999).  First, scientific and technical information are given a central role in the policy process.  Second, a time perspective of 10 years or more is required to understand policy change.  Third, the policy subsystem (defined by policy topic, geographic scope, and influencing actors) is the primary unit of analysis.  Fourth, the set of policy subsystem actors is expanded beyond the traditional members of the iron triangle to include officials from all levels of government, consultants, scientists, and members of the media.  Fifth, policies and programs can be viewed as translations of beliefs (Sabatier and Jenkins-Smith 1999). 

The theoretical construct of the individual posited by the ACF comes from Simon’s theory of bounded rationality, which states that due to limitations in resources and ability to process stimuli, an individual typically relies on heuristics for decision making, rather than behaving in a classical “rational” sense (Simon 1985).  The ACF’s construct of the individual also borrows from prospect theory (Quattrone and Tversky, 1988), which suggests that actors tend to remember losses more than gains (termed loss aversion in prospect theory).  The tendency to distort stimuli and remember losses more than gains increases the likelihood that actors will exaggerate power and maliciousness of their opponents, in what the ACF refers to as the devil shift (Sabatier et al., 1987).  The devil shift is important because it can amplify the severity of losses to a rival coalition and mobilize coordinated action among policy allies (Sabatier et al., 1987; Sabatier & Weible, 2007). 

The ACF explicitly identifies belief systems as the primary heuristic on which individuals rely for political decision making.  According to the ACF, the most central of these belief systems are deep core beliefs, which are fundamental to individuals, normative, largely a part of childhood socialization, and are thus quite difficult to change.  Examples include individual views on the role of government, beliefs about human nature, or priorities regarding who should participate in government.  The second component of the ACF’s belief system is policy core beliefs.  Policy core beliefs tend to be subsystem-wide in scope and are the foundation for forming coalitions, establishing alliances, and coordinating activities among subsystem members.  Policy core beliefs are resistant to change, but are more malleable than deep core beliefs.  The third component is secondary beliefs, which are narrower in scope, more empirically based, and the most likely of the components of the belief system to change over time due to new information and learning.  Examples of secondary beliefs might include detailed rules and budgetary decisions.

One of the major elements of the ACF is the distinction between policy subsystems and the broader political environment.  The ACF specifies the subsystem rather than the policy system as the major unit of analysis since broader political systems cover a vast number of topics over a geographically diverse area.  Therefore, policy actors must act within the policy subsystem in order to understand the technical aspects of a topic, its context, and its geographical application.  On the level of the subsystem, the policy actor has the greatest chance to produce change.  However, policy subsystems are not immune to the effects of the political environment in which they operate.  Figure 1 presents a flow diagram of the ACF, and shows that subsystems can be influenced by two factors, relatively stable parameters and external system events. 

The concept of belief systems is critical to understanding the logic of the ACF.  Since the ACF considers policies translations of belief systems (Sabatier and Jenkins-Smith, 1993), belief systems of policy actors can be used to enhance policy understanding, evaluate the formation, maintenance and structure of coalitions, generate expectations regarding coalition composition, and predict future policy changes.  The ACF defines major policy change as a change in the policy core aspects of a policy subsystem and minor policy change as a change in the secondary aspects of a policy subsystem (such as a shift in resources among sub-programs).  The ACF identifies four paths for major policy change to occur: external events or shocks; policy oriented learning; internal shocks; and negotiated agreements. 

One of the fundamental criticisms of the ACF has been the largely tacit assumptions that coincide with American pluralism (Sabatier, 2007).  Specifically, the ACF assumes “well organized interest groups, mission-oriented agencies, weak political parties, multiple decision making venues, and the need for supermajorities to enact and implement policy change” (Sabatier, 2007; p. 199).  Obviously this assumption does not fit well with the realities of European governance, thus limiting the applicability of the framework abroad.  In order to ease the application of the ACF outside of the United States, a new category of variables, known as “coalition opportunity structures” was developed.  Drawing strongly from the work of European policy scholars, political opportunity structures refer to factors such as resources and/or other constraints that affect the behavior of advocacy coalitions (Kriesi et al., 1996; Lijphart, 1999; Kubler, 2001).  This revision to the ACF contains two additional sets of variables: the degree of consensus needed for major policy change, and openness of political system. 

The idea behind the degree of consensus needed for major policy change is that, in open political systems (such as the United States), major policy changes are subjected to multiple veto points.  However, in Westminster systems, such as the United Kingdom and New Zealand, decision making is relatively centralized, with parliamentary majorities typically at numbers below 50%.  The implication is that in situations where a high degree of consent is required, there is an incentive for coalitions to work together, compromise, and minimize devil shift (Sabatier, 2007). 

The second concept, degree of openness of political system, depends on two things: the number of decision making venues, and the accessibility of those venues.  For example, in the United States, due to the system of checks and balances as well as federalist structure, there are multiple decision making venues (federal, state, and local as well as presidential, legislative, and judicial) that provide multiple access points and even incentivize participation.  In contrast, corporatist regimes will tend to have fewer access points as well as decision making venues due to the small number of participants involved in central governance (Sabatier, 2007).

This is an admittedly brief overview of the ACF, however several publications are available that provide in depth descriptions of the framework’s assumptions, hypotheses, concepts, and causal logic (Sabatier, 1986; Jenkins-Smith, 1990; Sabatier and Jenkins-Smith, 1993; 1999; Sabatier, 1988; Weible and Sabatier, 2006; Sabatier and Weible, 2007).  Additional information can also be obtained from the website authors upon request.

Researchers, who have applied the ACF, have commented on its strengths and limitations.  Highlights of the framework’s strengths include the following points.

  • The ACF helps explain policy change through policy-oriented learning and external events (e.g., Andersson, 1999; Brenton et al., 2006; Elliott and Schlaepfer, 2001b; Eberg, 1997; Kübler, 2001; Lertzman et al., 1996) 
  • The ACF helps contrast and simplify hundreds of actors into coalitions based on shared beliefs (e.g., Abrar et al., 2000; Farquharson, 2003; Kübler, 1999; Leschine et al., 2003).
  • The ACF helps explain the policy role of scientific and technical information (e.g., Jordon and Greenway, 1998; Radaelli and Martini, 1998).
  • The ACF helps explain political behavior (e.g., Burnett and Davis, 1996; Bischoff, 2001; Farquharson, 2003; Green and Houlihan, 2006; Sobeck, 2003).

The ACF has also been criticized.  The most notable criticism, which has already been discussed, is the potential bias of the ACF toward pluralistic political systems (Kübler, 1999; Eberg, 1997).  In response, the ACF has been modified to include coalition opportunity structures. Other limitations include the following points.

  • The ACF does not explicitly account for, or is ambiguous about, the role of ideas and self-interest in the policy process (e.g., Kübler, 1999; Compston and Madsen, 2001). 
  • The ACF is unclear in its depiction of collection action and coalition formation, strategies, and maintenance (e.g., Eberg, 1997; Larson, et al., 2006; Parrish, 2003; Schlager, 1995).
  • Major policy change is better conceptualized as multiple cascading external events than a single external event (e.g., Smith, 2000).
  • The ACF overlooks the stages of the policy cycle (e.g., Marzotto et al. 2000; Olson et al., 1999).
  • The ACF overlooks the role of public opinion in public policy (e.g., Jordan and Greenway, 1998).

*This overview was adapted from Sabatier and Weible, 2007; Weible, 2008; and Sabatier, Weible, and Flowers, 2008 | Physical Address: 1380 Lawrence St., Ste. 500  Denver, CO  80204 | Mailing Address: University of Colorado Denver, SPA, Campus Box 142, P.O. Box 173364, Denver, CO 80217-3364 | ph: 303.315.2228 | fax: 303.315.2229 | SPA Website Feedback

University of Colorado Denver

© The Regents of the University of Colorado, a body corporate. All rights reserved.

All trademarks are registered property of the University. Used by permission only.