1 Introduction: What Is Implementation and Why Should We Care About It?

This essay takes stock of the literature on how European Union¹ policies are implemented by the member states. The notion of policy implementation is tied to what has been called the “textbook conception of the policy process” (Nakamura 1987 : 142). This conception assumes that the policy cycle may be divided into several clearly distinguishable phases, ranging from problem definition and agenda-setting to policy formulation, policy implementation, evaluation and finally to policy termination or re-formulation. Policy implementation thus refers to “what happens after a bill becomes a law” (Bardach 1977 ) or, as one scholar has aptly put it, to the process of “translating policy into action” (Barrett 2004 : 251).

A similar, but slightly different concept is the notion of “compliance”. It has been prominent in international relations research among scholars studying the domestic fulfilment of international agreements (for an overview, see Raustiala and Slaughter 2002 ). Compliance refers to “a state of conformity or identity between an actor’s behavior and a specified rule” (Raustiala and Slaughter 2002 : 539). Thus, the compliance perspective also starts from a given norm and asks whether the addressees of the norm actually conform to it. It thus focuses less on the process than on the outcome of implementation. Moreover, compliance can occur without implementation. For example, the current practice may already conform to the status required by a norm. Conversely, implementation does not necessarily have to result in compliance but may be incomplete or contrary to the prescribed goals (Raustiala 2000 : 391–399). Irrespective of these semantic differences, most compliance and implementation research is interested in both the process of how a given norm is being put into practice and in the outcome in terms of rule conformity. In this sense, implementation and compliance are two sides of the same coin.

A third concept that is often used in the context of EU implementation research is the notion of “Europeanisation”, which points to the effects of European integration in the member states.² While both concepts, and the respective bodies of literature associated with them, have a considerable degree of overlap, it is important to keep in mind that there are also important analytical differences. The implementation of EU legislation usually entails certain policy or institutional changes at the domestic level. In this sense, implementation is one important mechanism of Europeanisation. However, the domestic effects of European integration are not confined to processes of policy transfer and the institutional adaptations associated with these. Instead, European integration may have many other, direct or indirect, intended or unintended, effects at the domestic level (Töller 2004 ). In this sense, Europeanisation research encompasses, but is broader than, EU implementation studies.

The following discussion will not focus on the wider field of Europeanisation research, as there is a separate Living Review on the “Europeanisation of national political systems”(forthcoming). Moreover, I will concentrate on implementation studies dealing with the fifteen “old” member states before Eastern enlargement, as there is a Living Review on “Europeanisation in new member and candidate states” by Sedelmeier (2006 ) , which also discusses the policy-oriented literature on how these countries implemented the acquis communautaire.

Why should social scientists care about implementation? The answer is straightforward: putting a piece of legislation or a government programme into practice does not happen automatically, nor is it a purely technical or apolitical affair. Instead, long delays and attempts at shirking seem to be a matter of everyday business in the field of implementation. In other words, if we are interested in the extent to which a particular polity is able to solve the problems with which it is confronted, we need to study not only the way it reaches decisions and the character of the resulting legal output, but also the way in which the law is executed in practice.

This is particularly true in a large and complex polity like the European Union. Due to the high number of veto players involved in policy formation, EU legislation often contains fuzzy concepts and “rhetorical compromises” in order to facilitate agreement. Moreover, EU legislation regularly leaves certain issues to the discretion of member states in order to take account of specific regional or local circumstances. In other words, crucial decisions that may decide on the success or failure of a particular policy are regularly taken at the implementation stage. What is more, it is far from self-evident that implementers will behave dutifully. The EU is marked by a highly decentralised implementation structure. It does not have its own administrative machinery to implement its legislation, but has to rely on the member states to fulfil this task. In that respect, the European multi-level system resembles the German system of co-operative federalism, in which federal legislation is carried out by the administrations of the Länder, much more than the US model of dual federalism, where each level has its own bureaucracy to put the respective laws into practice (Scharpf 1988 ).

The EU’s implementation structure appears to be even more precarious than the German one. Like in Germany, the lower level of governance is in charge of administrative enforcement. If we focus on the implementation of EU directives, one of the major legal instruments of the EU, however, it becomes apparent that also parts of the decision-making process are delegated to the domestic level. Directives only define goals that have to be incorporated into national law by member states within a certain period of time. It is only after transposition has been completed that the rules may be applied by societal target groups and enforced by administrations and the legal system at the domestic level (see Figure 1 ). The transposition phase thus adds another important “clearance point” (Pressman and Wildavsky 1973 ) to the decentralised implementation structure of the European Union. In other words, there is yet another point at which the realisation of the envisaged goals might fail.

Figure 1:

Stages of the Implementation Process

Directives thus share important characteristics of traditional international agreements which need to be ratified by member states before they become effective. Unlike international agreements, however, the implementation of EU law by member states is supervised by the European Commission and the European Court of Justice (ECJ). If the Commission detects violations of EU law, it can instigate legal proceedings against the respective member state. These proceedings may ultimately lead up to a judgement by the ECJ and, in the event of continuing non-compliance, to follow-up proceedings as a result of which the Court may impose serious financial fines. However, there is no European police force that could compel member states to obey the rules. As a result, the level of legal obligation and the extent of actual enforceability place EU law somewhere between traditional domestic law on the one hand and traditional international law on the other.

This makes the European Union a particularly interesting object of study for implementation researchers. After a relatively late start, the literature on how EU policies are put into practice domestically has in fact proliferated over the last two decades. Given the considerable body of publications that have been produced to date and the almost complete absence of any overview articles (but see Mastenbroek 2005 ), it seems to be about time to take stock of the state of the art in the field. This paper thus provides a structured overview of EU implementation studies. The next section provides an overview of the emergence and evolution of the field. It identifies three distinct waves of research,³ each with its own theoretical and empirical focus. Afterwards, the paper discusses the most important theoretical, empirical and methodological lessons to be drawn from existing studies and highlights promising avenues for future research. The final section concludes by summarising the main findings.