Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, POLITICS ( (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 26 March 2017

Implementation and the Policy Process

Summary and Keywords

As a part of the policy process, implementation follows policy as formulated and decided upon. Three aspects can be distinguished as inherent to the term implementation. The first one regards the temporal order in which implementation in a policy process takes place. The second aspect concerns the causal logic, while the third one is about the form of authority. Policy implementation is looked at and talked about from two fundamentally contrasting perspectives. One can be called an “ideal” perspective, the other a realistic one. “Ideal” stands here for a use of the term implementation without further reflection; the phenomena the term refers to are taken for granted. By contrast, the alternative perspective can be labeled as a realistic one. This perspective is a construction as well, but instead of taking things for granted it invites for empirical observation and testing.

From an ideal perspective implementation is viewed as following instructions. Implementation is seen as a separate stage, identifiable as such. Inputs are supposed to determine outputs, while authority is exercised in a hierarchical relationship. From a realistic perspective implementation is seen as practice. It is approached as a multilevel phenomenon. Results of a policy process are explained by a variety of factors and social mechanisms. Authority is exercised as based on various sources. Both the view on implementation as following instructions and its realistic opposite shed a relevant light on implementation and its place in the policy process. Each view can be found in the practice as well as the study of the policy process. In the expectations about national politics the view on implementation as following instructions may be more observed than the alternative view, while at the street level of government the opposite can be supposed. However, these concern empirical questions.

As far as implementation research is concerned, the normative appeal of the assumptions underlying the view on implementation as following instructions makes that view still occurring. At the same time, these assumptions have been challenged rather fundamentally, both at a theoretical and empirical level. The opposite character of the two views has consequences for the ways implementation and its place within the policy process are understood, but also for the ways in which variation in the results of policy processes is explained. Ultimately, understanding and explaining those results are enhanced when an approach is adopted in which elements from both views have been incorporated.

Keywords: implementation, implementation theory, implementation research, policy process, public policy, policy formation, policy design, decision-making, street-level bureaucracy, discretion, operational governance


Complete or execute (a contract), fulfill (an undertaking), put (a decision or plan) into effect. Those are the primary linguistic meanings of the verb “to implement” (Trumble & Brown, 2002, p. 1330). In French the Anglo-Saxon verb is translated as mettre-en-œvre, or accomplir (Larousse, 2010, p. 523). This meaning refers back to accomplishing: bringing intentions from decision-making toward their realization. In the context of politics and government what needs to be implemented usually is a public policy. Implementation then stands for the realization of the policy goals legitimately decided upon.

As such the word implies a threefold set of assumptions, concerning a particular temporal order, causal logic, and form of authority. The ways in which policy implementation generally is looked at and talked about, reflect two fundamentally contrasting perspectives. One can be called an “ideal” perspective, the other a realistic one. “Ideal” stands here for a use of the term implementation without further reflection; the phenomena the term refers to are taken for granted. This is the standard way of looking at what is seen as something instrumental. By contrast, the alternative perspective can be labeled as a realistic one. This perspective, more fit for the understanding and explanation of what happens in practice, is a construction as well. However, instead of taking things for granted it invites for empirical observation and testing.

The contrast between these two perspectives is central. The view on implementation stemming from the ideal perspective will be characterized under the label: “Implementation as following instructions.” The view on implementation from a realistic perspective is labeled “Implementation as practice.” The implications of the two views for implementation research and policy practice are considered.1

Implementation as Following Instructions

Public policy as a concept is connected with the modern state. It is often seen to be about addressing problems in society. Shelter (housing), food and clothing (social security), literacy (education), and so many more human needs are seen as to be brought about, one way or another, by some role of government. As Samuel Finer has shown in his history of government, the more elementary tasks of territorial protection and maintaining law and order are older, and actually stem from far before the modern nation-state (Finer, 1999). The establishment of the latter was accompanied in the 19th century by a role concerning the creation of a public infrastructure. It was particularly in reaction to the Great Depression, and after World War II, that government intervention took hold in the development of the welfare state. Also in the neo-interventionism of contemporary government, concerned with regulation and behavior change, public policy regards the substance of that role.

When hence a relation between public policy and modernity is apparent, the characteristics of the latter can be retraced in the thinking about the former. This is particularly so as far as the notion of the policy process expresses a problem-solving model (Lodge & Wegrich, 2014). The metaphor of a cycle refers to the subsequent stages in which a particular societal problem is identified (agenda-setting), next is being framed while action is decided upon (policy formation as a combination of policy design and decision-making), before such action is put into effect (implementation). The cycle continues when results are evidenced and assessed (evaluation) with feedback consequences for the public policy concerned. Conceiving the range of clusters of successive activities aimed at societal problem-solving as a form of problem processing Deborah Stone characterizes the stages model of the policy process as an assembly line (2012).

Hence the concept of the policy process as such has intrinsic features related to the rational ideal of the homo economicus, as in multiple ways embedded in the modern era. Consequently, in its paradigmatic form policy implementation is connected with the extension of the welfare state in the 1970s (Hill & Hupe, 2014). Jeffrey Pressman and Aaron Wildavsky (1973) coined the term with their book that would become emblematic for the policy-implementation paradigm. Although the heydays of the grand policy programs of the welfare state seem over, the view on implementation so well-articulated in that classic of public administration has survived the specific historical era in which it was published. Three assumptions underlie the ideal (or rational) view on implementation as axiomatically expressed in implementation. They are normatively appealing to such an extent that the persistence of this view is guaranteed. Those assumptions shed a specific light on the way implementation is positioned in time (implementation as a separate stage), on the implied causal logic (inputs determine outputs), and on which form of authority is at stake (a hierarchical relationship).

Implementation as a Separate Stage

However—even still—appealing, the stagist view on the policy process is a particular one. Elinor Ostrom (2007) distinguished between models, theories, and frameworks. Models are the least comprehensive, but the most specified. They entail particular sets of expected relationships between a limited number of variables. Comprehensiveness, empirical openness, and enhancing middle-range theory formation can be specified as the criteria of an analytical framework. The “stages model” of the policy process meets the criteria of a framework, in the sense that it aspires to be comprehensive, while each of the “stages” refers to what have become scholarly themes of their own. It is also robust, substantively rather empty, and is functional from a didactic perspective. However, it does not meet the requirement of empirical openness. In research, using the stagist model of the policy process does not explain much.2

The study of policy implementation itself suggests that implementation can be identified as something separate from what needs to be implemented. The latter is assumed to be a well-defined result of a preceding (sub) process. Such policy formation, as an approach to the societal problem involved, is seen as a linear chain of related, primarily mental, activities, interrupted by moments of decision-making. The perspective is one of design; knowledge-informed desk work is central. That kind of work, done by officials called policy analysts, involves identifying relationships between causes and consequences, problems and solutions, costs and benefits, instruments and goals. There is politics, of course, but it is located outside the “intellectual cogitation” aimed at designing a policy as the optimal response to the problem at hand, as Aaron Wildavsky (1979) has pointed out. Political authorities are also in practical terms supposed to be in the lead, in the sense that they control the policy process from the top.

Inputs Determine Outputs

When the causes of the societal problem at hand have been traced, the direction of the policy to be pursued becomes obvious. The policy objectives, politically expressed, are supposed to be decisive next for the policy instruments to be chosen. Accordingly, in implementation these instruments are put into effect. In the subsequent view on implementation, the latter is seen as a technical matter. “The bureaucracy” is viewed as an instrumental device. Christopher Hood, and, respectively, Lewis Gunn and Brian Hogwood, formulated the conditions for a perfect implementation (Hood, 1976; Gunn, 1978; Hogwood & Gunn, 1984). They point to action, practiced on all public-administrative layers, as following in line with the objectives and guidelines as stated: complete obedience of all actors involved, a flawless communication, and an absence of time constraints. It is clear these authors used “perfect” here in the sense of an ideal-typical construct.

The policy as laid down in a white paper or other official document is considered as guiding: in the public domain, but in the first place for implementers. Such a policy document is seen as entailing a clearly defined prescriptive input. As a next step in the policy process, implementation is a matter of applying instructions. Besides, there is a normative directive: when the policy goals have been formulated, and appropriate instruments have been legitimately decided upon, democracy demands a corresponding implementation. In some cases policy makers in ministries show an awareness of the ambiguity and conflict a policy may embody, as Richard Matland (1995) has pointed out. Then they may anticipate what Annemarije Oosterwaal and René Torenvlied have called “policy divergence” (2012). From a control perspective those policymakers then may circumscribe the degree of permitted divergence.

A Hierarchical Relationship

Implementation is thus conceived as following policy—both in terms of a temporal order and causal logic. These two assumptions are accompanied, and strengthened, by the grounding principle that, normatively, a hierarchy is prevailing here. As far as the exercise of authority is concerned it is politics that determines what administrators subsequently have to do. And the former should, next, keep control but without interfering too directly in the daily work administration entails. That is what Woodrow Wilson claimed in his famous article published in 1887. The claim offered a justification for the study of executive government as a separate object of research. The separation of administration from politics can also be found in the usual view of the policy process and the relationship between the parts it consists of. Policy formation is deemed more important than policy implementation.

This explains why, in general, cabinet ministers and other political authorities tend to give limited attention to the implementation part of a policy process. First, on both the legislative and executive side of the separated powers configuration a command-and-control form of authority is presupposed, exercised by the political principal. Within the executive branch civil servants are hidden behind their political superiors; constitutionally, they do not exist. The institutionalized norm of the political primacy implies that political representatives only do business with political authorities.

Second, in terms of political-administrative relations a matter of labor division is concerned. When policy advisers (policy analysts) at a ministry identify the means-end relationships, and the minister makes the decisions, the rest is matter of implementation. Given the policy agenda of an average department it is simply impossible for a cabinet minister to have—at most—a substantial role in more than the stages of agenda-setting and policy formation. Third, given a cabinet minister’s diary, participating in matters of high politics is more rewarding than in those of low politics. Sketching the outline of a system revision in health care, for instance, and giving direction to an intended reform, are politically more worthwhile than focusing on the nitty-gritty of such an operation.

Fourth, given the politics/administration dichotomy, within administration the self-conception of a “primacy” is reproduced. Vis-à-vis actors expected to implement the public policies they have formulated, policy advisers in the national civil service see themselves as principals. Within bureaucracy, the public servants working in so called “implementation organizations” (police districts, hospitals, schools, and the like) are supposed as agents to merely execute policy objectives. Fifth, the ideology of New Public Management, widespread as it is, implied that, where deemed possible, implementation was to be privatized, or in other ways contracted out. As long as the targets agreed upon are reached, the principal is not interested in how the agent performs his task—and that goes whether those principals are ministers, or their policy advisers at—cabinet level.

Until things in an unmistaken manner are perceived to go wrong—one could add. Then there is disappointment, but especially an inclination to show toughness. What usually follows are more stringent rules, more tight oversight on compliance, and stricter sanctions on noncompliance. This is the standard reaction observable in case great expectations are dashed (Pressman & Wildavsky, 1973). The effects of such a reaction remain claimed rather than evidenced. Nevertheless, implementation is supposed to follow policy—until the opposite becomes obvious. Then the black box of what happens on the ground is being opened.

As expressed in the view characterized here, policy implementation is addressed as something that is determined by the result of policy formation, presupposed as a residual toward goal achievement, and is normatively subordinate. Such a view on implementation as a phenomenon that comes next, is programmed, and of secondary relevance, can still often be recognized where and when policy goals are being formulated and decided upon. This view on implementation, stemming from an “ideal” perspective, has become generally known as the typical top-down view on implementation. As such it may be observable anywhere, but not in the least with top-level policymakers in national government. That this view also in research cannot be written-off is evidenced by the ongoing stream of single case implementation studies (Sætren, 2014). Despite its normative attractiveness in terms of democracy, the view on implementation as following instructions may prevent researchers from observing and understanding what actually happens and explaining why, before expressing judgements on the legitimacy of what is recorded.

Implementation as Practice

Ongoing Policymaking

From an ideal perspective, within the policy process as a whole, policy formation is the stage directly preceding, and therefore most formative to, implementation. The term policy formation stands for a combination of formulation and decision-making. Given that combination, policy formation is primarily seen as an intellectual exercise. Policy is a question of rational design. Its crux is instrument choice, expressed in a clear direction-giving policy document.

At the same time, however, there are some classic authors in public administration who have shed light on the importance of nonrational—not to be confused with irrational—factors shaping the substance of policy formation. Doing so, these authors have pointed to the relevance of a class of public servants, whose task it is as policy advisers to formulate policy documents of all sorts. Addressing the insights of those authors is relevant here, because the result of policy formation as a “stage” is a circumscription of what, next, needs to be implemented.

Herbert Simon (1945) pointed out how individual decision-making is inherently bound to cognitive limitations. Talking about “the science of muddling through” Charles Lindblom (1959) referred to the iterative, tentative, and most of all incremental ways policy advisers in the civil service formulate policy measures. In his threefold analysis of the Cuban missile crisis Graham Allison (1971) added, along with the rational actor perspective, two ways of looking at policy formation in national government. He showed how important standard operating procedures are, and how organization-specific repertoires may influence instrument choice. Moreover, Allison gave an account of the ways in which parts of government struggle with each other, in order to get the final say. This bureaucratic politics is not only normal—as far as it is embedded in procedures of political decision-making—it is functional as well, because it enables a multiple look at the problem at hand and the incorporation of a plurality of angles and interests.

Policy formation hence can be characterized as a process of social interaction, punctuated by moments of reflection—to paraphrase Aaron Wildavsky’s characterization (1979). No wonder then that the policy document resulting from such a process may contain something other than clear-cut policy objectives and appropriately chosen instruments. Often the policy that needs to be implemented has traits of a compromise. Power politics, idea-politics, bureaucratic politics, party politics—they all, to a greater or smaller extent, are involved. In cases of “co-production” with citizens, corporations, or organizations of civil society, policy formation extends beyond the walls of departments. Therefore, the policy document sooner or later agreed upon is the result of multiple negotiations. Its formulation may show omissions, flaws, ambiguities, and even inconsistencies. Hence, for those supposed to implement the policy involved the implemendum is anything but an instruction. Besides, those actors are seldom tasked with the implementation of only one public policy.

In a policy process, policymaking continues after the “stage” of policy formation. In a federal system, for instance, the implementation of a policy by a “lower” layer of public administration may be assumed, while in fact legitimate policy co-formation there is at stake. Also apart from this, public policies are inherently political, in the sense that they are about who gets what, when, and how, to use Harold Lasswell’s definition (1958). Politics does not stop when the policy design has been politically approved. During the entire policy process there are values involved, as well as stakes represented. Besides, all texts need interpretation, and policy-on-paper not in the least. As Michael Lipsky discovered, so called street-level bureaucrats are policymakers in their own right (2010). They de facto fulfill the role of “ministers on the ground.”

Outputs Explained by Multiple Factors

In reality, policy implementation is something else than putting an instrumental choice into effect. On all administrative layers other than those where top-level decision-making takes place, implementation concerns daily practice, rather than something that “comes next.” For police officers, health inspectors, and teachers, for instance, respectively, the law is to be enforced, the objectives stated in the ministerial white paper are to be addressed, the mission formulated by the school board is to be reckoned with—of course. Nevertheless, however the goals stated-on-paper may give orientation, they do not determine actual behavior. For often those policy goals are ambiguous, or even contradictory. Organizational and other stakes are involved. And only the need to adapt to the circumstances at hand implies that dilemmas, interpretation, and judgement are inherent to street-level practice.

All these facts challenge the norm of perfect implementation. Between system level and street level various factors and social mechanisms play a role. In principle, they all can be expected to influence variation in policy outputs, let alone outcomes. In general, most public servants can be expected to do their work properly, and in a way loyal to the policies they are supposed to implement.3 Even then, however, those public employees—whatever their profession and level of education and training—cannot be considered as puppets on a string.

Where implementation takes place numerous influences play a role, adding up to, strengthening, compensating, or eliminating other influences. Phenomena are occurring, developments unfold, without a design from a rational actor underlying it. To a certain extent the action logic in the realm of implementation has a dynamic of its own—not foreseen from any desk, and hard to control, from whatever position of power. For the examination of this implementation politics, empirical research needs to go beyond the evaluation of goal achievement. The latter qualification is used by Søren Winter to characterize the nature of mainstream, single case, implementation studies with a view from the top (2012).

A Relative Autonomy

Police officers, health inspectors, teachers in a local school: in their daily contacts with, respectively, car drivers, patients, and pupils, they all—one way or another, whether or not wearing a uniform—represent government. At the same time they have a certain room to fulfill their tasks in ways they individually deem appropriate. The usage of discretion is implied by the very nature of their task, and may be constrained as well as enhanced by the circumstances they do their work in, and the kind of organization they are members of.4 Apart from the “vertical” perspective of state agency, there is the “horizontal” orientation of citizen agency, as Steven Maynard-Moody and Michael Musheno have pointed out (2000).

Like discretion as used may vary, the discretion granted to these street-level bureaucrats may differ as well. Sometimes policymakers in national government, or even the legislator involved, are aware of the fact that the variety of situations in which the policy or law concerned has to be implemented is impossible to foresee. Consequently, they may formally grant a specified, greater or smaller, discretionary authority to “lower” public-administrative layers, and the public officials doing their work there.

Even if such a formal mandate explicitly is lacking, there is not immediately reason to fear that things will run out of control. At the street level the demands are multiple, but the accountability practiced is as well. Potential shortcomings in “vertical” instruction may be “horizontally” compensated by counterweights provided in the social interaction with professional and societal co-actors. There is the obvious accountability toward supervisors, public managers, and political authorities (political-administrative accountability). Apart from this hierarchy, street-level bureaucrats justify their acts toward their peers and colleagues in their occupation (professional accountability). Furthermore, accountability toward society is practiced. This happens at various levels; at the micro-level, for instance, when parents ask a teacher to explain the treatment of their child (participatory accountability) (Hupe & Hill, 2007). Hence checks and balances are working.

Public officials on the ground floor of government are daily confronted with multiple demands and action prescriptions stemming from a range of sources. The very fact of the variety of forms of accountability provides them ways to make balanced judgements. The degree of professionalization and institutional socialization, but also of individual internalization and self-binding, will vary. The usage of freedom in the implementation part of the policy process therefore may be beyond direct control. However, the occurrence of limitless freedom is an exception—not the rule.

Implementation: Authority in Practice

From an ideal perspective, implementation is presupposed as following instructions, in a residual toward goal achievement. The view could be characterized as the standard meaning of implementation. The alternative view, implementation exercised as practice in settings that vary in multidimensional ways, could be seen as a more “political” way of looking. From a historical point of view it seemed as if the competition between the two opposite views was resolved when Malcolm Goggin and his colleagues in 1990 made a plea for “synthesizing” approaches (Goggin, Bowman, Lester, & O’Toole, 1990; Sabatier, 1986). The number of studies with a more or less rigorous research design has grown (Sætren, 2014). The aim to structure the large number of potentially explanatory factors has led to some approaches with a more comprehensive character. Examples are the policy-implementation process approach of Donald Van Meter and Carl Van Horn, the communications model of intergovernmental policy implementation of Malcolm Goggin and his co-authors, the advocacy coalition framework Paul Sabatier developed with Hank Jenkins-Smith, Christopher Weible, and others, the political regime approach of Peter May and his colleagues, and the integrated implementation model of Søren Winter (Van Meter & Van Horn, 1975; Goggin, Bowman, Lester, & O’Toole, 1990; Sabatier, 1986; May, 2015; May & Jochim, 2013; Winter, 2012). These approaches share the aim of explaining the results of policy processes from implementation factors in combination with factors located in “preceding” parts of the policy process concerned, as well as in the “horizontal” environment of the implementation locus.

Despite all this, an ongoing stream of implementation studies still regards single cases, looked at from a top-down point of view. In such studies usually the identification of an “implementation deficit” is central, while sources of noncompliance are being sought. Bo Rothstein once characterized this kind of study as “misery research” (1998). On the other hand, currently a global movement aimed at establishing an “implementation science” can be observed (Nilsen, Stähl, Roback, & Cairney, 2013). The latter concerns an intervention-focused kind of research in health, medicine, and related life sciences. In a way, this initiative embraces the quest for rigor as central in the case made for a synthesis in policy implementation research. However, the claim of a “science” also seems to bring us back to the adoption of the principles of rationality. The claim appears to represent a sustained attractiveness of a technocratic ideal: the desire to rule out implementation politics.5 The latter can be conceived as everything that has not been designed from a desk but nevertheless has a substantial influence on what happens. It is the incorporation of ambiguity, conflict and other “political” dimensions of social interaction that may provide a greater potential for the explanation of variation in the results of policy processes than just the objectives stated in a policy-on-paper.


When assessing the two contrasting views on implementation, from both views implications for research can be distinguished. In the view on implementation as following instructions, politics stays reserved to the stage of policy formation. As a consequence, the political dimensions inherent in the continuation of policymaking throughout the policy process are neglected. At most, the occurrence of bureaucratic politics is acknowledged, but high politics is leading. Second, implementing actors—actually, any actor other than those involved in policy formation—a priori are presupposed to act at least in congruence with the policy objectives. While forms of policy obstruction tend to be exceptions, the actual relevance of discretionary capacities of legitimate co-policymaking actors remains under-assessed. Third, the hierarchical character of the relationship in which a public policy is being implemented functions as a general and widely adopted norm. However, the purity of what should happen cannot explain the messy character of happens. By confusing the norm with reality, the view of implementation as following instructions provides insufficient potential to explain empirical variation in what happens with policy objectives.

The research implications of the view of implementation as practice are different, but identifiable as well. When there is a focus on the street level, research tends to neglect the institutional dimensions inherent in policymaking and its environment. Second, the relevance of discretionary authority as granted from “higher” political-administrative layers is under-assessed. Third, acknowledging the relative autonomy of street-level bureaucrats empirically as a fact is one thing. Something else, however, is “putting the bottom on top” by normatively championing the street as opposite to the state or system. However “horizontally” oriented individual, or even categories of, street-level bureaucrats may be to the citizens they are in contact with, they remain part of what, fundamentally, is “vertical” administration.

In summary: in the view of implementation as following instructions, organizational- and individual-level actions are subsumed, while management dimensions are neglected. By contrast, in the alternative view structural dimensions may be overlooked, while organizational-level factors tend to get secondary attention. At the same time, from both views elements can be drawn that can contribute to future implementation research. The realistic perspective most of all draws attention to action (if going beyond perceptions, as a proxy for that action). What happens with policy-on-paper in a practice that is messy, only partly designed, and hard to control? Second, operational dimensions get attention. Who are, on the various layers, discretionary actors (cf. first-line public managers), and how do they use their discretion? Third, the human factor is incorporated. Even a flawlessly designed policy ultimately is dependent on individuals when putting it into effect. Ongoing innovations in office- and communication technologies suggest a development toward “infocracies” (Zuurmond, 2012; Bovens & Zouridis, 2002). However, when a teacher has to weigh the ambiguous results of an obligatory test taken by one of her pupils, she goes to her peer in the classroom next door for advice.

The view on implementation as following instructions contributes a few “top-down” elements potentially improving implementation research. First, it is acknowledged that high politics and top-level policy formation frame the authoritative parameters of a public policy. How have mandates been formulated? What party political struggle, bureaucratic politics, ideological debate, and power coalition are hidden behind the policy goals as stated? Second, technological and sociocultural developments of a global nature tend to get “path dependent” local translations. How particular is the given institutional setting? What are similar and what are varying findings when the fulfillment of the same public task in various contexts is being compared? Third, in a policy process public officials as well as other actors on various layers fulfill public roles. They do so in empirically varying degrees of shown professionalism. How are both hierarchical control and horizontal accountability being organized and practiced?

The two views on implementation have practical implications too for the activities a policy process as ongoing policymaking consists of. First, in policy formation it is wise to be aware of the fundamentally political character of a public policy per se. Preventing pollution may be expected to be less controversial than euthanasia, but each of these policies concern public matters—that is their raison d’être. This means that they have a normative basis; values are at stake. There is a political drive; usually the stakes are high. At the street level, interpretations of what needs to be implemented are inevitable, while there is an action imperative. As an alternative to leaving things open with an eye on the possibility of shifting the blame, in policy formation micro-management seems an option. Instead of both, expressing some political principles underlying the policy initiative at hand can be functional.

Laurence J’ O’Toole Jr once described the “top-down perspective’s conventional wisdom” (1986). Keep the number of actors involved in the implementation of your policy low, is one of the recommendations implied in that wisdom. As the realistic view learns, the idea that such would be possible is a fiction. Instead, it seems both sensible and feasible to give attention to arena formation—the second implication. Crucial is the organization of checks and balances. Welcoming opposition is a matter of using what Charles Lindblom called “the intelligence of democracy.” Countervailing powers enable the evolution of insights via social interaction and mutual adjustment (Lindblom, 1965).

Third, practically any actor in a policy process has, to an empirically varying extent, a relative autonomy. Some of it is based on a formally granted position, a different part on reputation: an actor is allowed room to act, because relevant others trust the actor to use that room in an appropriate way (Carpenter, 1999). When the possibilities of direct control are limited, institutional settings become important. Given the constraints on the designability of the latter, enhancing professionalization is crucial, on all layers of vertical administration.


Implementation is what is supposed to follow after a decision has been made. When the term implementation is used in the discourse of national government, it usually refers to a residual—something of a secondary nature. Once the policy goals have been formulated and the choice of instruments has been decided upon, the rest is a matter of implementation. This usage of the term then implies, first, a chronological order in which expressions of will precede action. Second, a linear causal logic is assumed: goals determine instruments, instruments determine results. Third, a hierarchy is inherent: policy formation is more important than policy implementation. The latter is seen as a matter of following instructions: an apolitical, merely administrative, activity.

Actually, these substantive assumptions underlying the view on implementation stemming from an ideal perspective are specific. In the study of politics and government the implemendum, that which needs to be executed or realized, generally concerns a law or public policy. Such a government document seldom is self-evident enough to function as a clear and consistent set of uniform prescriptions. Goals and means are intertwined, while certain formulations of the policy-on-paper may be results of negotiation between, within, and beyond governments. Therefore, the letter of the law, white paper, or whatever form an official document gets, rarely supplies clear instructions to the public officials who are supposed to implement the objectives stated.

The two contrasting views on implementation as part of the policy process refer back to a classic topic in political theory and epistemology, the relationship between Reason and Power. In a rational view, goals and means, causes and consequences, costs and benefits are central. In what can be called an action-logical view the focus, in an opposite way, is on the dynamics of human behavior and the mechanisms of social interaction. Both ways of looking at implementation as part of the policy process appear to remain competing. Actually, however, a co-existence rather than a competition is at stake here.

Both views on implementation as presented in this account are constructions; they are not to be reified. As shown, they each have different implications for research, suggesting points to learn from each other. In research the view on implementation as practice stands for the state of the art. At the same time, in several respects elements from the other view could be incorporated. Moreover, the adoption of the view on implementation as practice, proclaimed in handbooks on the subject, does not prevent hundreds of mainstream implementation studies being added yearly. As long as the limits of administration will have an impact on what happens with good intentions formulated in a public policy, the stream of implementation studies concluding with “dashed expectations” will continue (Pressman & Wildavsky, 1973). A guarantee for this continuity is the normative appeal of the assumptions inherent in the ideal perspective, matching as the view on implementation as following instructions does with a sustained quest for control (van Gunsteren, 1976). However, to understand what happens and to explain why, more is needed than adopting a view that actually has a more normative than empirical character.


Allison, G. T. (1971). Essence of decision: Explaining the Cuban Missile Crisis. Boston: Little, Brown.Find this resource:

Bovens, M. A. P., & Zouridis, S. (2002). From street-level to system-level bureaucracies: How information and communication technology is transforming administrative discretion and constitutional control. Public Administration Review, 62(2), 174–184.Find this resource:

Brodkin, E. Z. (1990). Implementation as policy politics. In D. Palumbo & D. Calista (Eds.), Implementation and the policy process: Opening up the black box (pp. 107–131). Westport, CT: Greenwood Press.Find this resource:

Carpenter, D. P. (1999). The forging of bureaucratic autonomy: Reputations, networks, and policy innovation in executive agencies, 1862–1928. Princeton, NJ: Princeton University Press.Find this resource:

Finer, S. E. (1999). The history of government (3 vols.). Oxford: Oxford University Press.Find this resource:

Goggin, M. L., Bowman, A. O’M., Lester, J. P., & O’Toole, L. J. (1990). Implementation theory and practice: Toward a third generation. Glenview, IL: Scott Foresman.Find this resource:

Gunn, L. (1978). Why is implementation so difficult? Management Services in Government, 33(4), 169–176.Find this resource:

Gunsteren, H. R. van. (1976). The quest for control: A critique of the rational-central rule approach in public affairs. London: John Wiley.Find this resource:

Hill, M. J., & Hupe, P. L. (2014). Implementing public policy: An introduction to the study of operational governance (3d ed.). London: SAGE.Find this resource:

Hogwood, B. W., & Gunn, L. (1984). Policy analysis for the real world. Oxford: Oxford University Press.Find this resource:

Hood, C. C. (1976). The limits of administration. London: John Wiley.Find this resource:

Hupe, P. L., & Hill, M. J. (2007). Street-level bureaucracy and public accountability. Public Administration, 85(2), 279–300.Find this resource:

Hupe, P. L., Hill, M. J., & Buffat, A. (Eds.). (2015). Understanding street-level bureaucracy. Bristol, U.K.: Policy Press.Find this resource:

Jenkins-Smith, H. C., Nohrstedt, D., Weible, C. M., & Sabatier, P. A. (2014). The advocacy coalition framework: Foundations, evolution, and ongoing research. In P. A. Sabatier & C. M. Weible (Eds.), Theories of policy process (3d ed., pp. 183–223). Boulder, CO: Westview Press.Find this resource:

Larousse. (2010). Le petit Larousse. Paris: Larousse.Find this resource:

Lasswell, H. D. (1958). Politics: Who gets what, when, how. New York: Meridian Books.Find this resource:

Lindblom, C. E. (1959). The science of muddling through. Public Administration Review, 19(2), 79–88.Find this resource:

Lindblom, C. E. (1965). The intelligence of democracy: Decision making through mutual adjustment. New York: Free Press.Find this resource:

Lipsky, M. (2010). Street-level bureaucracy: Dilemmas of the individual in public services (30th anniversary expanded ed.). New York: Russell Sage Foundation.Find this resource:

Lodge, M., & Wegrich, K. (Eds.). (2014). The problem-solving capacity of the modern state: Governance challenges and administrative capacities. Oxford: Oxford University Press.Find this resource:

Maynard-Moody, S., & Musheno, M. (2000). State agent or citizen agent: Two narratives of discretion. Journal of Public Administration Research and Theory, 2(2), 329–358.Find this resource:

Matland, R. E. (1995). Synthesizing the implementation literature: The ambiguity-conflict model of policy implementation. Journal of Public Administration Research and Theory, 5(2), 145–174.Find this resource:

May, P. J. (2015). Implementation failures revisited: Policy regime perspectives. Public Policy and Administration, 30(3–4), 277–299.Find this resource:

May, P. J., & Jochim, A. E. (2013). Policy regime perspectives: Policies, politics and governing. Policy Studies Journal, 41(3), 426–452.Find this resource:

Maynard-Moody, S., & Musheno, M. (2000). State agent or citizen agent: Two narratives of discretion. Journal of Public Administration Research and Theory, 2(2), 329–358.Find this resource:

Nilsen, P., Stähl, C., Roback, K., & Cairney, P. (2013). Never the twain shall meet? A comparison of implementation science and policy implementation research. Implementation Science, 8(63), 1–12.Find this resource:

Oosterwaal, A., & Torenvlied, R. (2012). Policy divergence in implementation: How conflict among decisive legislators reinforces the effect of agency preferences. Journal of Public Administration Research and Theory, 22(2), 171–195.Find this resource:

Ostrom, E. (2007). Institutional rational choice: An assessment of the institutional analysis and development framework. In P. A. Sabatier (Ed.), Theories of the policy process (2d ed., pp. 3–17). Boulder, CO: Westview Press.Find this resource:

O’Toole, L. J., Jr. (1986). Policy recommendations for multi-actor implementation: An assessment of the field. Journal of Public Policy, 6(2), 181–210.Find this resource:

Pressman, J., & Wildavsky, A. B. (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. Berkeley: University of California Press.Find this resource:

Rothstein, J. (1998). Just institutions matter: The political logic of the universal welfare state. Cambridge, U.K.: Cambridge University Press, 1998.Find this resource:

Sabatier, P. A. (1986). Top-down and bottom-up approaches to implementation research: A critical analysis and suggested synthesis. Journal of Public Policy, 6(1), 21–48.Find this resource:

Sabatier, P. A., & Weible, C. M. (2007). The advocacy coalition framework: Innovations and clarifications. In P. A. Sabatier (Ed.), Theories of the policy process (2d ed., pp. 189–220). Boulder, CO: Westview Press.Find this resource:

Sætren, H. (2014). Implementing the third generation research paradigm in policy implementation research: An empirical assessment. Public Policy and Administration, 29(2), 84–105.Find this resource:

Simon, H. A. (1945). Administrative behavior. New York: Free Press.Find this resource:

Stone, D. (2012). Policy paradox: The art of political decision-making (3d ed.). New York: W. W. Norton.Find this resource:

Trumble, W., & Brown, L. (Eds.). (2002). Shorter Oxford English dictionary (Vol. 1, A–M, 5th ed.). Oxford: Oxford University Press.Find this resource:

Tummers, L. G. (2013). Policy alienation and the power of professionals: Confronting new policies. Cheltenham, U.K.: Edward Elgar.Find this resource:

Van Meter, D., & Van Horn, C. E. (1975). The policy implementation process: A conceptual framework. Administration and Society, 6(4), 445–488.Find this resource:

Wildavsky, A. B. (1979). Speaking truth to power: The art and craft of policy analysis. Boston: Little, Brown.Find this resource:

Wilson, J. Q. (2000). Bureaucracy: What government agencies do and why they do it. New York: Basic Books.Find this resource:

Wilson, W. (1887). The study of administration. Political Science Quarterly, 2, 197–222.Find this resource:

Winter, S. C. (2012). Implementation. In B. G. Peters & J. Pierre (Eds.), Handbook of public administration (pp. 255–263). London: SAGE.Find this resource:

Zuurmond, A. (2012). Understanding infocracies: Trends in the transformation of organizations. In I. Th. M. Snellen, M. Thaens, & W. B. H. J. van de Donk (Eds.), Public administration in the information age: Revisited (pp. 164–181). Amsterdam: IOS Press.Find this resource:


(1.) As implied by the character of an encyclopedia, much of what is presented in this article concerns insights gained over the years and in social interaction with others. A part of those insights has been developed and written about in a longstanding collaboration with Michael Hill. He is thanked for his comments on an earlier version of this account.

(2.) Analytical frameworks have been developed that could serve as alternatives to the stages heuristic. They are equally aimed at structuring various aspects in the analysis of government-in-action but do so from a less normative point of reference. Laurence Lynn Jr and his collaborators developed the logic of governance framework. Elinor Ostrom elaborated, in association with others, the institutional analysis and development framework. Inspired by these two analytical constructs the present author with Michael Hill introduced the multiple governance framework, or trias gubernandi; see Hill & Hupe, 2014.

(3.) Although the reactions to new policies empirically may vary; see Lars G. Tummers, Policy alienation and the power of professionals: Confronting new policies (Cheltenham, U.K.: Edward Elgar, 2013).

(4.) For a differentiation of types of agencies see James Q. Wilson, Bureaucracy: What government agencies do and why they do it (New York: Basic Books, 2000).

(5.) Evelyn Z. Brodkin speaks of policy politics; see her “Implementation as Policy Politics,” in Implementation and the policy process: Opening up the black box, eds. Dennis J. Palumbo and Donald J. Calista (Westport, CT: Greenwood Press, 1990), 107–131.