Comparative Policy Implementation Research
Policy implementation refers to a phase of the strategic planning process, when actions are taken to realize policy objectives. The implementation stage consists of identifying policy actors who are responsible for policy implementation, allocating funds, developing infrastructure, building strategic support among actors who deliver services, and other actions taken to administer political decisions.
Comparative implementation research is a special theme within the field of public management and administration research. It gained momentum due to the increased pressure for evidence-based practices in the policymaking process. The early stage in the development of policy implementation research (1960s–1970s) has been triggered by the exploration of topics such as healthcare, privatization, human resource management, civil service, and budget reforms. Subsequent stages witnessed the development of a new scholarly perspective aiming to describe implementation as a distinct topic of political science scholarship (Allison 1971; Easton 1979; Pressman and Wildavsky 1973). The new perspective offered to isolate the implementation from other stages and dimensions of policy process in relation to a multitude of causal factors explaining policy outcomes (Bardach 1977; Brewer 1974; Mazmanian and Sabatier 1983; Pressman and Wildavsky 1973). Thus, instead of taking policy field, subsystem, or agency as the unit of analysis, researchers in this stream focused on the criteria of policy implementation process that would explain the causes of policy success and failure across various policy fields. During the 1980s and 1990s, significant efforts were made to develop a set of generalizable principles behind a uniform model of policy implementation (DeLeon 1999; Goggin et al. 1990; O’Toole 2000). These efforts produced a body of research emphasizing the role of multiactor and multilevel perspectives of policy implementation processes.
The study of implementation process is continuously challenged by the fact that the field is theoretically and conceptually underdeveloped. It lacks clearly defined concepts for comparing policy implementation processes across nations and placing them into separate classification systems. One of the problems challenging our ability to study comparative state government reform effectively is that scholars have failed to provide a plausible definition of successful implementation process. Cross-country analyses point out that a policy field followed over many years can change so radically that it bears little resemblance to its initial form (Matland 1995, 152). In this context, it appears to be quite difficult to assess the progress of reforms that are based in conventional means, such as goal attainment.
Explanations of what constitutes ‘good implementation’ may be divided into two categories, “control” and “learning” options (Hoppe et al. 1985). In the former, successful implementation is confined to 100% compliance with policy planners’ design (it presupposes an accountability strategy that ensures the compliance of implementers and target groups). In the latter option, obeying the rules is not crucially important. By contrast, finding an optimal learning strategy is what guarantees the overall success of policy implementation (Hoppe et al., 588).
when agencies strictly follow directives and statutes issues by the national or subnational governments.
when agencies meet specific indicators of success, achieve goals of the statute; they are also held accountable for reaching their goals.
when improvements are identifiable in the improvement of the political climate around the program (in this context, an outcome may be considered successful when it produces learning) (Matland 1995).
Overall, these are only technical criteria of policy implementation success, and the failure to identify what is meant by successful implementation may serve as source of considerable confusion in public policy analysis.
Evolution of Policy Implementation Research
The development of policy implementation research may be outlined as a process consisting of at least three “generations” of interdisciplinary scholarship.
The first generation of implementation research (1960s–1970s) discussed the reasons for the successful and unsuccessful policy outcomes across various sectors, aiming to explain to what extent agencies comply with policy objectives. Questions covered by scholars working within this stream concerned implementation as a distinct stage of policy process. Many of the studies produced during this time served the goal of gathering data on the ways in which things get done at the local level rather than on the goal theory-building process (for more information, see Van Meter and Van Horn 1975; Mazmanian and Sabatier 1981).
The second stage (1970s–1980s) consisted in an even stronger empirical focus on implementation phenomena, which aimed to overcome the flaws of a-theoretical case study research through the creation of theoretical constructs, testable hypotheses, and relevant statistical techniques. This stage brought about the increased number of comparative small-n case studies, where the number of cases has been more than one, which allowed arriving at testable theoretical propositions (for more information, see Lester et al. 1987, Sabatier 1988).
Finally, the third stage of policy implementation scholarship (1990s–present) embraced studies aiming to create rigorous research designs, focusing on variables, hypotheses, case selection techniques, and the use of mixed-method techniques in acquiring and evaluating relevant data. This wave of research delved into the cross-national comparisons (small-n and large-n case study analyses) and relied on comparative research design as an important feature of political science scholarship (Goggin et al. 1990, 19). A series of studies produced during this stage discussed the utility of single, small-n and large-n case-studies and the ways in which their explanatory power could be improved through a careful selection of crucial and critical cases for the purpose of testing a theory. Generally, research on public policy implementation was vibrant and moving into a new century in qualitative and quantitative terms. Over time, it relied more heavily on multilevel and multiactor perspectives.
It is noteworthy that the development of implementation research has never been a steady process moving from one generation to another in a linear fashion. In fact, there has been a growing amount of scholarship that relied on a variety of perspectives. In this context, the term of policy implementation “generations” (Goggin 1986; Goggin et al. 1990) has been developed to cope with the growing amount of academic discussion focusing on the problem of managing change through policy implementation process.
O’Toole (2000) suggests that there have been various “strands” of policy implementation research beyond the stages of policy implementation scholarship, which span a few ‘generations’ of relevant policy implementation studies. Among these strands are: (1) institutional analysis, which included rational choice institutionalism and actor-centered institutionalism; (2) a governance approach, as well as political and policy regime theory, focusing on a process of governing rather than the organizational features of governments; (3) networks and management; (4) formal rational choice models (e.g., and deductive approaches); and (5) policy design and instrument approaches (273–282).
The observed strands did not necessarily make the distinction between the stages of policy formulation and policy implementation. However, research produced within relevant subunits of analysis contributed greatly to the existing theories and models of policy implementation process.
Institutional analysis, for example, has been successfully applied to the study of long-term solutions to problems of common-pool resource management. The governance approach extended the boundaries of policy implementation research by focusing on a broader set of theoretical propositions from game theory and regime framework. Finally, network and network management analysis scrutinized the forms through which most implementation develops (structural forms, multiactor networked patterns, etc.), whereas instrument approach built on a smaller number of policy implementation variables linking theory with the design of public policy and implementation mechanisms (O’Toole 2000).
Problems and Challenges of Policy Implementation Research
While the number of policy implementation variables is growing on a case-by-case basis, most of the existing analytical insights offer methodological guidelines rather than a set of theoretical propositions about policy implementation dynamics.
The observed guidelines, e.g., research strategies of policy implementation process, may be divided into at least two distinct categories, e.g., “top-down” and “bottom-up” models, which center around two distinct approaches toward the study of policy process, and accordingly on distinct categories of the actors involved in policy implementation stage.
Proponents of a top-down model assume that clarity of goals and control by the policy makers lead to more effective implementation and greater success in addressing problems. The implementation analysis that is located in this model tends to focus on factors that could be easily manipulated by policy makers at the central level (see Hoppe et al. 1985; Mazmanian and Sabatier 1983).
By contrast, supporters of the bottom-up approach emphasize target groups and service deliverers, arguing that policy change is in fact made at the local level. They start from a policy problem and then examine the strategies that are employed by relevant participants at different levels of the government as they attempt to deal with the issues that are consistent with their objectives (see Lipsky 1980).
In recent years, actors and processes, entrenched in various stages of reform, have re-aligned in a growing body of research that, deals with issues of policy implementation dynamism. The first and major step in this process occurred as a result of the difficulty observed in efforts to conceptualize the interaction between policy and setting or to construct any single policy theory that was “context free” (Matland 1995, 149). In view of the observed analytical difficulty, scholars experienced the increased pressure to bridge a natural division between the top-down and bottom-up units of analysis; and even though the means of doing so were never clear, some interesting and theoretically useful efforts transpired in establishing a dialogue between the design and implementation processes.
Hoppe et al. (1985), for example, asserted that the major problems at the stage of policy implementation emerge due to the disconnected cognitive levels and maps of reform designers, implementers, and target groups. In this respect, implementation research promised to significantly improve “high-level policy-makers’ cognitive maps” of the developments leading to both expected and unexpected policy outcomes (582). The process of cognitive mapping, which is quite curious on its own, was based on the idea that policy-making is a continuous learning process, which starts from attempts on behalf of various actors of the reform to interpret data and transform it into action-oriented knowledge (or empirical and normative knowledge). The process of learning has been defined as “simultaneous problem identification, problem definition, and problem solving, which are in turn interrupted by all sorts of discontinuities,” when programs are unrealistic and do not match the resource capacity of implementers (Ibid.).
A limited number of policy implementation studies have argued that top-down and bottom-up perspectives could be used in reference to different cases or otherwise combined depending on the issue at question. Some of the most often-cited perspectives within the area of policy implementation dynamism include Elmore’s concept of “forward” and “backward” mapping (Elmore 1985), which helps identifying potential problems by comparing cognitive maps of policy formulators and policy implementers; Sabatier’s policy cycle perspective (1988), which observes the development of policy as a function of political change; and Goggin et al. 1990 communications model, which places state implementers “at the nexus of a series of communication channels” (inducements and constraints from the federal, regional, and local levels) susceptible to distortions as a result of signal perception bias (Matland 1995, 151–152).
The usefulness of the above-mentioned frameworks depends on working assumptions, such as the role of ideas, interests, and institutions in policy process, or the distinction between the stages of policy formulation and policy implementation in policy-making process. It is well known, for example, that a number of competing policy models were developed prior to the emergence of policy implementation scholarship to explain how events occur to produce policy change. Some of these models are linear, while others are more complex. The linear model of policy developed by Lasswell (1951) and modified by scholars, such as Brewer (1974), Meier (1991), DeLeon (1999), included four steps taken in policymaking, e.g., agenda stage, decision stage, implementation, and evaluation stages). Subsequently, these ideas transformed into a theory of policy cycle, suggesting that the relationship among the observed stages appear to be locked in a circular system. Each stage requires significant resource inputs for the reform to take place.
Mainstream approaches, such as the policy streams approach developed by Kingdon (1984), do not focus on stages of policy process as such. Instead, they view the stages and processes of policy-making as being interconnected. Specifically, Kingdon argues that policy change comes about when streams – problems, politics, and policies – connect. This model suggests that the areas of expertise, politics, policy-making exist independently of one another until they come together under the pressure of policy advocates (policy entrepreneurs) entering the system during a critical moment of system opening described as “policy window.” The streams do not merge, but rather interconnect, shifting policy agendas or changing the entire mechanisms of policy implementation.
The current stage of policy implementation research consists in the development of theoretical insights that aim to move the analytical focus of political science scholars beyond a single issue or a single level of analysis. Most of these insights rest upon various versions of institutionalism theory that are applied to the field of public policy and administration (among these versions are rational choice institutionalism, historical institutionalism, and sociological institutionalism). Matland (1995), for example, demonstrated that implementation stage results from actors’ preferences and the level of conflict around policy goals. Based on these assumptions, he proposed that policy implementation stage results from (the choice of policy instruments, which corresponds with the level of ambiguity and conflict in policy implementation stage.
Based on these assumptions, the following types of policy implementation were identified: administrative implementation (low conflict-low ambiguity), political implementation (high conflict-low ambiguity), symbolic implementation (high conflict-high ambiguity), and experimental implementation (low conflict-high ambiguity) (145).
Similar to Matland, Chackerian and Mavima (2001) assert that different reform components have to be properly assessed and combined in order to ensure successful policy implementation. Building on criteria such as resource inputs (i.e., their size and similarity) and the time frame of policy implementation process, the author identified three possible effects that emerge from interaction among reform components, i.e., “synergy,” “trade-off,” and “avoidance.”(Ibid.)
The above-mentioned findings of scholars are useful for understanding how the process affects policy outcome, as well as the relationship between the reform design and implementation is. However, these studies ignore the reasons for the adoption of specific policies, or the difference between intended and unintended effects of policy-making process (the role of policy actors vis-à-vis structural conditions of policy change).
It is noteworthy that some policies are not designed to fulfill their official goals, and many public sector initiatives are nothing more than symbolic attempts to appease various segments of the populace rather than improve the quality of public administration (Edelman 1964). Therefore, whenever political leadership is missing, there is a distinct possibility that a set of reform measures will end up going nowhere in terms of policy implementation. Thus, regardless of whether there is a problem with the degree of consensus among the top-level public officials (“the unified nature of political will to reform”), or the quality of decision-making at the top level (it is autonomy and insulation from societal pressures), policy leadership dilemmas are highly consequential for the stages that following the process of policy enactment. In this respect, implementation difficulties may emerge for a number of reasons: politicians depending on a single group (or a few powerful groups) that promote diverse agendas; leaders diverting their attention to other areas in an attempt to pursue alternative or hidden agendas, or as a result of hectic decision-making style. One of the biggest flaws of the dynamic policy implementation models is, therefore, that they ignore the difference between intended and unintended effects of policy implementation, and as such, they do not fully contribute to the role of agency in policy implementation process.
Thomas and Grindle (1994) emphasize the importance of policy leadership in policy implementation process and suggest that the extent to which political actors (described as policy elites) should be involved in the process of reform implementation is influenced by the existing government’s degree of legitimacy and autonomy of the existing government. First, if regime is stable, it is more likely to generate wide-ranging support for the program. Second, if regime depends on a few extremely powerful interest groups, it would be more difficult to ensure consensus among the top-level political elites concerning the goals of reform.
The assessment of the political resource needed to support reform program requires consideration of who supports and opposes reform program, as well as the degree of support required for reforms to take place. In addition to the political support, bureaucratic, financial, and technical resources must be evaluated.
The field of institutional economics provides valuable insights into the type of resources required in the policy implementation stage. Rodrick (1996), for example, distinguished between the processes of policy initiation and the process of reform consolidation, and suggested that “successfully initiating reform and sustaining reform may depend on different approaches” (9–41). In this perspective, “the process of policy initiation requires independence or autonomy for the executive, while consolidation of reform necessitates [the] building of legislative and interest-group bases of support” (in Williamson 1994, 468). Legislative support means that “all necessary regulations are taken in order to develop reform initiatives.” Interest-group support means that “there is a consensus on the aims and the means of reforms, and major actors take decisions in line with their consensus” (Ibid.).
Interestingly, policy-making process nearly always results from a combination of actors’ preferences and political context, including “selective” problem definition and implementation based on a perceived salience of an issue. In this respect, one of the most important gaps observed in the literature concerns the system of “institutional capacities” of a state undergoing political, social, and economic transition. The concept of state capacity refers to the ability of a state to achieve its own goals; therefore, it may serve as both structural and contextual variable of policy implementation context. Institutional factors define and underpin state capacity (in this case, rational legality associated with bureaucratic and administrative capacity). The particular types of institutional capacities in question overlap to some extent with the policy network dynamics noted above, particularly, in relation to the institutional attributes of the state.
Each of the existing models captures just a minor component of policy process operating within a dynamic policy context. Institutional analysis shifts the discussion toward the analysis of policy choice and the accompanying structural conditions thus extending the boundaries of microlevel analyses employed by policy implementation scholars.
This entry has provided an overview of comparative implementation research as a special theme in public management and administration research. It began with the discussion of challenges that the field of comparative implementation research faces, such as the difficulty of providing a clear definition of successful implementation, and in identifying concepts and categories for grouping diverse phenomena into separate classification systems. It was observed that the unit of analysis of policy implementation scholarship has changed over time, moving the analytical focus from the nature of policy field to the organizational prerequisites of change process, actors, and strategies accounting for the success and failure of policy implementation.
This entry has also provided an overview of major strands of policy implementation research, as well as themes and questions raised by political science scholars regarding the process of policy implementation on a cross-country basis. Analysis indicates that there have been a few overlapping stages of policy implementation research coinciding with at least three generations of political science scholarship. The first stage consisted of a-theoretical descriptive case study analyses that focused on policy implementation variables (conditions for successful policy implementation). The second stage included efforts to formulate and test hypotheses that were focused on comparative case study analyses. Finally, the third stage moved the entire focus of political science scholarship toward the development of rigorous research designs, variables, hypotheses, case selection techniques, and the use of mixed-method techniques for acquiring and evaluating relevant data. All of the above generations were influenced by diverse perspectives, such as the tradition of rational choice institutionalism, governance approach, networks and management, formal rational choice models, and, finally, policy design and instrument approaches.
The current stage of policy implementation research consists in the development of theoretical insights that aim to move the analytical focus of political science scholars beyond a single issue or a single level of analysis. However, while the number of policy implementation variables is growing on a case-by-case basis, most of the existing analytical insights offer methodological guidelines rather than a set of theoretical propositions about policy implementation dynamics. One of the key gaps observed in the literature concerns the lack of systematic effort to link the discussion of policy implementation to the discussion of institutional prerequisites of policy process. The analytical and predictive capacity of policy implementation research will significantly increase as long as it relies more heavily on multilevel and multiactor models situating the discussion within the broader context of political science scholarship.
- Allison GT (1971) Essence of decision: explaining the cuban missile crisis. Little, Brown, BostonGoogle Scholar
- Bardach E (1977) The implementation game: what happens after a bill becomes a law. MIT Press, Cambridge, MAGoogle Scholar
- Easton D (1979) A systems analysis of political life. University of Chicago Press, ChicagoGoogle Scholar
- Edelman M (1964) The symbolic uses of politics. University of Illinois Press, UrbanaGoogle Scholar
- Elmore RF (1985) Forward and backward mapping. In: Hanf K, Toonen T (eds) Policy implementation in federal and unitary systems. Martinus Nijhoff, DordrechtGoogle Scholar
- Goggin ML, Bowman AOM, Lester JP, O’Toole JLJ (1990) Implementation theory and practice: toward a third generation. Scott Foresman/Little, Brown and Company, GlenviewGoogle Scholar
- Kingdon JW (1984) Agendas, alternatives and public policies. Little, Brown, BostonGoogle Scholar
- Lasswell H (1951) The policy orientation. In: Lerner D, Lasswell H (eds) The policy sciences: recent developments in scope and method. Stanford University Press, Stanford, pp 3–15Google Scholar
- Lipsky M (1980) Street-level bureaucracy: dilemmas of the individual in public service. Russel Sage, New YorkGoogle Scholar
- Matland R (1995) Synthesizing policy implementation literature: the ambiguity- conflict model of policy implementation. J Public Adm Res and Theory 5(2): 145–174Google Scholar
- Mazmanian D, Sabatier P (1981) Effective policy implementation. Lexington Books/D.C. Health, LexingtonGoogle Scholar
- Mazmanian DA, Sabatier P (1983) Implementation and Public Policy (Glenview, IL: Scott, Foresman)Google Scholar
- Meier GM (1991) Politics and policymaking in developing countries: perspectives on the new political economy. International Center for Economic Growth Press, San FranciscoGoogle Scholar
- Pressman JL, Wildavsky AB (1973) Implementation: how great expectations in Washington are dashed in Oakland: or, why it’s amazing that federal programs work at all, this being a saga of the economic development administration as told by two sympathetic observers who seek to build morals on a foundation of ruined hopes. University of California Press, BerkeleyGoogle Scholar
- Rodrick D (1996) Understanding economic policy reform. J Econ Lit 34(1):9–41Google Scholar
- Williamson J (ed) (1994) The political economy of policy reform. Institute for International Economics, Washington, DCGoogle Scholar