email icon Email this citation


New Concepts — Old Problems? The Institutional Constraints for the Effctive Implementation of EU Environmental Policy

Christoph Knill and Andrea Lenschow

International Studies Association
41st Annual Convention
Los Angeles, CA
March 14-18, 2000



The theoretical discussion concerning the problem solving capacity of the European Union (EU) has focused in the past on problems of policy formulation and decision making. Here, the structural constraints built into the EU multi-level governance system have received primary attention. 1 Even though the ability of the EU to develop political programmes and particular policies is certainly a central criterion for assessing its problem solving capacity, it must not be overlooked that the effective implementation of these programmes and policies is essential in order to achieve their objectives. Nevertheless, only relatively recently policy implementation has become an issue in the literature on the performance of European governance (Collins/Earnshaw 1992; Jordan 1999; Knill/Lenschow 1998).

The evolution of the academic discussion has been similar in the field of International Relations (IR) and specifically the sub-field of regime analysis. Initially the primary questions of regime analysts have targeted the conditions for regime building and regime maintenance (cf. Kohler-Koch 1989; Hasenclever/Mayer/Rittberger 1997) — this is analogue to EU studies first focusing on the process of European integration per se. Studies of the evolution of particular regimes emphasised the puzzle of agreeing/deciding on international norms in a presumed anarchical international context. Hence, similar to the EU literature, the capacity of international regimes was measured in terms of decision making. Only recently, the scholarly attention in the field of IR has shifted to the question of norm compliance and the implementation of international agreements on the ground (cf. Haas/Keohane/Levy 1993; Jacobsen/Brown Weiss 1995; Sand 1992; Victor 1995; Zürn 1995).

Both for the effectiveness of international regimes and of EU policies it is true that processes during policy formulation and decision making have an impact on the subsequent implementation phase. The constellation of interests will affect the nature and design of the agreement to be implemented, raising more or less resistance on the part of the implementing actors on the ground. Notwithstanding this linkage, we will divide analytically the two phases and focus on the latter. We ask which factors influence the implementation of — in our case — EU environmental legislation on the national level.

Despite the noted parallels in IR and Public Policy in terms of their empirical interests, there also exist important differences. Most notable is the difference in problem perception — also indicated in the name of this workshop "Why do social actors comply?". This is the question raised by IR scholars who's vision is influenced by the expectations stated by the realist school — or more generally, rational choice approaches. Realists are generally surprised by cooperative behaviour in international relations, hence also with the creation of regimes and subsequently the compliance with international norm. Studies of norm compliance in international relations are constructed in opposition to the still prevailing realist expectation. By contrast, public policy analysts are puzzled by non-compliance. Typically coming from the perspective of national politics, the default expectation is the obedience with government and law and it needs to be explained "why do social actor not comply?" Here it is the image of the effective 'Leviathan' that is challenged. In scrutinising the implementation performance of the EU multi-level system — which is 'more' than a regime but 'less' than a state (cf. W. Wallace, 1983), these two perspectives meet. In this paper, we will depart from a Public Policy perspective, however link our observations to the IR literature where ever appropriate.


The 'Winds of Change' in EU Environmental Policy: Reactions to an Increasing Implementation Gap

It is no secret that the implementation of many EU policies has been deficient and that environmental policy stands out as particularly problematic (CEC 1996, EP 1996; Collins/Earnshaw 1992). The Commission’s own statistics on the implementation of EU environmental legislation reveal serious deficits: In 1995 member states notified their implementation for only 91 percent of the Community’s environmental directives, some member states leaving more than 20 percent not transposed (CEC 1996). In the same year, the Commission registered a total of 265 suspected breaches of Community law, whereby the environmental field accounts for 20 percent of all infringements registered by the Commission during that year. In October 1996 over 600 environmental complaints and infringement cases were outstanding against member states, with 85 cases awaiting determination by the Court of Justice (ECJ) (ibid., 2).

Politically even more explosive is the link between policy outputs, outcomes and the legitimacy of EU policy making in general. Thus, there is a growing awareness that implementation deficits might call into question the legitimacy and credibility of the integration process (ibid., 6). Indeed, it is in response to experiences with ineffective implementation, that EU critics call to drive back the EU regulatory output. In this context it is often difficult to distinguish the analytic from the ideological — undoubtedly EU sceptics have exploited the 'news' of implementation deficits to push a deeper, systemic anti-EU agenda (cf. EP 1993; Golub 1993). Evidently, policy implementation is more than a technical issue and of high political significance.

As a consequence of substantial implementation problems and the related political debate, we observe the emergence of new institutions and new policies which differ significantly from traditional approaches in EU environmental policy. Traditional approaches rely on technocratic and interventionist forms of top-down policy making where uniform and detailed requirements apply to all national administrations involved in the practical implementation and enforcement, as well as to private actors, such as industrial operators, who are addressed by the European policy in question. At the heart of the 'new approach', which is reflected in the 1993 fifth Environmental Action Programme of the EU (CEC 1993), is a reorientation in the mode of governance towards network-style and bottom-up forms of policy formulation and implementation (Lenschow forthcoming). Increasing importance is placed on procedural regulation, self-regulation, public participation and voluntary agreements as well as horizontally rather than hierarchically structured processes of policy development and enforcement.

The 'new approach', in many instances, is praised as a panacea to overcome problems of ineffective implementation. The advantage of new instruments vis-à-vis traditional regulatory measures is considered two-fold. First new instruments are argued to leave member states more leeway to comply with EU requirements by taking account of domestic context conditions. In contrast to the detailed and rigid forms of top-down instruments that are to be uniformly implemented regardless of the physical, economic or political context, new instruments focus on establishing basic procedures for improving environmental awareness and behaviour while setting no concrete environmental targets. Second, new instruments also target the policy context directly and aim to change context factors in order to facilitate the formal and practical implementation of environmental policy in general. By enlarging the channels for societal mobilisation through more transparent processes and participatory opportunities and in providing economic incentives for industrial self-regulation, new instruments are explicitly directed at increasing environmental awareness and responsibility of societal actors. Hence, by raising the level of acceptance with environmental objectives, it is hoped that new instruments may not only be effectively implemented themselves, but that they also provide a positive framework for the implementation of other measures. For instance, new information requirements may trigger greater public control of potentially polluting actors who may otherwise fail to comply with obligations imposed by other policies. Furthermore, these institutional and procedural innovations are equally directed at policy makers and implementers. By creating a more deliberative context for policy making they raise awareness of possible implementation problems among the policy makers; equally, participatory procedures induce a higher level of acceptance with EU policy on the part of policy implementers.

Indeed, there is a lot to be said for this shift in environmental regulation. It seems quite reasonable to expect that the implementation of policy measures which allow for compliance in the light of the specific institutional, political and socio-economic conditions at the domestic level are less likely to suffer from the resistance of subordinate administrative actors dealing with practical enforcement or of societal actors addressed by the policy in question. In a similar way, it seems to make much sense that the explicit orientation to stimulate and motivate domestic support might further improve implementation performance.

Besides these plausibility arguments, advocates of the new approach can rely on theoretical support, as the new approach responds to the insights gained in implementation research. The various schools in public policy implementation research have converged in acknowledging the need to secure the support of the implementing actors and the relevant interests in the policy planning process (cf. Cerych/Sabatier 1986; Ingram/Schneider 1990; Peters 1993). Likewise, there is some agreement that the socio-economic context needs to be considered during the design phase (Sabatier 1986). Likewise, relevant studies in the IR literature point to the need for regimes to contribute to "capacity and concern building" (Zürn 1995, 20) in order to improve norm compliance (see also Underdal 1998 on models of "domestic politics of implementation" and on "social learning and policy diffusion" and Haas 1998 with a similar outlook). 2 The Agenda 21 and the recently adopted Convention on Access to Environmental Information, Public Participation in Environmental Decision Making and Access to Justice in Environmental Matters (Aarhus, 23 -25 June 1998) convey this new spirit also in international environmental politics. 3


Questions, Puzzles, Explanations

In view of the political and scientific 'advance praise' rendered on behalf of the new approach, it is the objective of this paper to begin a critical assessment of the actual contributions of these regulatory and institutional (governance) reforms to improve environmental policy effectiveness in the EU. We ask the following questions: First, how successfully are these new measures implemented themselves? From a comparative perspective, are new policy instruments implemented better than top-down, interventionist regulations? Second, are new policy instruments indeed capable of raising awareness, acceptance and initiative on the level of implementation? More generally, to what extent are they successful tools to increase the problem solving capacity of the European multi-level system?

We will raise in our analysis some scepticism with regard to the great promise associated with new instruments. Making reference to a number of empirical studies we note that, first, 'new' policy instruments suffer from significant implementation deficits themselves, questioning their overall effectiveness. Second, when comparing old and new policies in terms of implementation effectiveness, no significant differences emerge. Instead, successes and failures vary across policies and countries, without indicating a direct causal linkage between policy type and implementation effectiveness.

The fact that the new approach has achieved only limited success so far may have two possible explanations. Either the theory remains deficient, or its practical application. We will argue in this paper that some truth lies in both. From this conclusion another question emerges: Are there solutions to address these theoretical and practical deficits with respect to the implementation of EU environmental policy? Which concrete steps, if any, can be taken to improve the implementation effectiveness of European policies, and hence the problem solving capacity of the Community?

We argue that at least some of the above problems can be addressed by adopting an institutional perspective in evaluating the promises of either top-down or bottom-up instruments. Such an institutional perspective, which so far plays only a minor role in implementation theory, takes account of institutional adaptation requirements implied by European institutional and policy changes. We suggest that the domestic institutional implications of European policies rather than the choice of the policy type are primarily responsible for implementation performance. An institutional perspective will not only provide a synthesis in the sometimes polarised debate between old and new approaches and address some of the resulting theoretical confusions, but it may also provide a more promising departure for practical policy design.

We proceed in the following way: To elaborate on our argument, we will introduce the major policy innovations associated with the new approach in EU environmental policy and briefly, note successes and failures in their implementation. The empirical basis of this assessment has been elaborated at greater length elsewhere; the focus of this paper is the attempt to explain why also new instruments suffer from weak implementation. We develop an institutionalist explanatory framework and present potential solutions to overcome both theoretical and practical deficits. First however, we provide a definition of effective implementation which will be the foundation of all subsequent discussions.


The Conception of Effective Implementation

Before moving to the empirical and theoretical discussion of effective implementation, we need to clarify what we mean by this term. This is no easy task as the literature offers a variety of definitions that differ in their conception of success. For reasons further specified below, we define effective implementation as the degree to which the formal transposition and the practical application of institutional and instrumental changes correspond to the objectives defined in European legislation. Hence, it is the compliance with these objectives rather than a normative evaluation of environmental quality improvements that concerns us here — we focus on output rather than outcome.

Implicit in an outcome oriented definition would be an assumption of a mechanistic causal relationship between policy objectives, instruments and policy outcomes; it implies a match between objectives and outcomes. If the objective of European legislation is, for instance, to achieve a certain level of drinking water quality throughout the Community, effective implementation is achieved as soon as the prescribed level is reached. This concept is analytically problematic as it obscures the actual link between policy instruments and outcomes. Whether policy objectives are achieved or not depends on contingencies of the political, economic, and social policy context beside the deliberate choice of policy instruments. How do we know whether drinking water quality has improved as a result of European legislation rather than as the consequence of other factors completely independent from European developments, such as different weather conditions, privatisation, or economic decline? This perspective ignores that the success of policy implementation in terms of policy outcome is not predictable given the scientific uncertainties and socio-economic complexities underlying a given problem constellation (Baier/March/ Sætren 1990; Lane 1995, 110).

A contrasting approach in the literature explicitly tries to address these problems by applying an even broader perspective on implementation. Here, policy objectives and instruments are no longer defined as benchmarks to be reached, instead it is expected that they may undergo modifications during the process of policy implementation. Implementers should have the flexibility and autonomy to adjust the policy in the light of particular local requirements, changes in the perception or constellation of policy problems, as well as new scientific evidence on causal relationships between means and ends. From this perspective, successful implementation is judged by the extent to which the perceived outcomes correspond with the preferences of the actors involved in the implementation process (Ingram/Schneider 1990). Effective implementation is analysed from a process-related rather than output- or outcome-based perspective. The crucial question for evaluating implementation success is to what extent did a certain policy allow for processes of learning, capacity building and support building in order to address policy problems in a decentralised way consistent with the interests of the actors involved? Implicit is the absence of a baseline for evaluating implementation results, however. First, it fails to offer a measuring rod for learning, capacity or support. When do these processes actually occur and work successfully? Secondly, and maybe even more significantly from the perspective of evaluating the impact of new policy instruments, this definition ignores the nature of the link between the (EU) policy and the local process of learning and problem solving. After all, local processes could have entirely different origins. In order to assess whether the EU policy has had some impact on these local processes, we need to observe whether its — however open — 'instructions' have been complied with.

Our definition of effective implementation avoids the conceptual problems of both schools just described. Rather than analysing the accomplishment of policy objectives in terms of outcomes, we restrict our investigation to the compliance with the legal-administrative tasks (i.e., output factors) specified in the legislation. These also perform as the baseline for evaluating the impact of new instruments and institutions. Hence, we start from the assumption that changes in outcome (be it environmental quality or local learning and capacity) depend on — more or less demanding — legal and procedural adaptation to EU policies. EU policies may require the reallocation of administrative competencies, the creation of new administrative structures or the adaptation of existing procedures and rules. We ask, whether both the formal transposition and practical application of the activities associated with the new approach in EU environmental policy took place in a way that is consistent with the concrete output requirements in these measures. Have existing legal and institutional arrangements been modified in a way that corresponds to the objectives underlying European legislation?

This in some ways narrow conception of implementation effectiveness offers several advantages for the purpose of this paper. Not only do we escape conceptual problems implicit in other definitions discussed above. Even more important in our context, the investigative path from EU policy — via compliance with output objectives — to the substantive impact of the new approach allows us to develop a tight linkage between implementation effectiveness and the problem solving capacity of the multi-layered system of the EU. Effective implementation, in terms of the member states’ general preparedness to accept and comply with European policies, is a basic factor indicating the level of the EU’s problem solving capacity. Changes in policy outcomes follow from this, even though the causal path becomes more blurred at this stage and hence more difficult to follow empirically; in cases of ineffective implementation the link between EU policy making and the achievements of the desired outcomes would in any case be interrupted mid-way.

Besides these substantive reasons for investigating (first) the level of compliance with output objectives, this definition of implementation effectiveness has an analytical advantage. Such a formal definition provides the opportunity to compare implementation results of very different policy measures. By contrast, it would be problematic to measure and hence compare the contributions of different policies (e.g. the Directive for free access to environmental information and the Large Combustion Plant Directive) to the achievement of a normative policy goals (such as air quality). In other words, our definition of implementation effectiveness allows us to compare and draw conclusions from a variety of EU policy cases.

Let us now turn to the investigation of the question whether the EU problem solving capacity may improve due to a turn towards the new regulatory approach.


The Limits of the ‘New Approach’

Since the emergence of environmental policy as an autonomous policy field in the EU in the 1970s it has followed hierarchical, interventionist patterns of policy steering. Policies established detailed, substantive standards, leaving member states little room for flexible adaptation. Environmental legislation in the fields of air and water pollution are typical for this era in EU policy making (cf. Haigh 1996). These measures define concrete emission and quality standards and expect the application of best available technologies. Hence, besides environmental objectives the legislation also regulate the means by which they should be achieved, reducing the scope for national responses.

The 'new' approach outlined above is mirrored in a number of recent regulations and directives. The Directive on the Freedom of Access to Environmental Information reflects efforts to improve the transparency of and public participation in environmental regulation. The Directive includes a passive right to information, hence all public authorities in possession of environmentally relevant information must allow the public access to these data; The request for information does not need to be justified. The Directive on Environmental Impact Assessment points in a similar direction. Besides the implicit cross-media approach in project authorisation procedures, the Directive defined far-reaching rights for public participation. The Eco-Audit and the Eco-Label Regulation, in turn, introduce economic incentives in order to encourage industrial self-regulation. The Eco-Audit Regulation establishes a framework that allows firms to implement an environmental management scheme which may be publicly registered if it becomes validated by an external auditor. The Eco.Label Scheme targets consumer behaviour by signalling for individual products their environmental production standards and properties.

Implementation deficits of EU air and water legislation have become common knowledge and have been described at some length in the literature (Collins/Earnshaw 1992; Krämer 1992; Rehbinder/Stewart 1985; Ward/Lowe/Butler 1997). Recent studies, however, note that new instruments as well suffer from weak implementation. Comparative country studies indicate that neither the choice of policy instrument nor the country under investigation render clear explanatory patterns (Knill 1998; Knill/Lenschow 1998). Similarly, other studies focusing on the level of performance of new instruments in several countries conclude that implementation problem occur at a higher rate than could be expected on the basis of theory as well as political rhetoric (Börzel 1997; Caddy/Favoino 1997; Bouma, Kimber, Wright, all forthcoming). Hence, the new approach directed towards horizontal governance, regulatory transparency, self-regulation, procedural regulation or self-regulation does not automatically imply effective implementation.

In order to explain these implementation failures, we advance three arguments explaining the weak performance of the new approach. First, the theoretical literature on implementation and the potential of of bottom-up policy instruments suffers from inconsistencies and ambiguities in its argumentation and hence, its implications for actual policy making. Second, political practice rarely reflects the 'pure' application of theoretical insights; i.e. in many instances new policy instruments are actually hybrids containing elements of both old and new approaches. Third, it is often overlooked in theory and practice that the new approach suffers from its own typical weaknesses which might lead to additional implementation problems. In particular, the factor usually viewed as the basic advantage of new instruments, namely their context orientation, remains a theoretically diffuse and practically complex category.


Theoretical Ambiguity

The fact that the emergence of new strategies to improve the implementation effectiveness of EU environmental policy are well in line with recommendations drawn from implementation theory does not automatically imply that the relevant theoretical body reflects a dominant and undisputed paradigm. To be sure, there are numerous studies emphasising the advantages of bottom-up concepts, indicating the need to consider the specific context constellation in which a certain policy is implemented (such as interests, capabilities, motivations of implementers and target groups) (Lipsky 1980; Berman 1980). The top-down approach, by contrast, focuses on policy content, design and the presence of enforcement mechanisms as the basic factors affecting implementation performance (Krämer 1992; Lübbe-Wolff 1996). Yet other analyses suggest a mix of top-down and bottom-up elements as the perfect solution (cf. Sabatier 1986: 23-25) or advocate a contingent approach to the choice of policy instruments (Ingram/Schneider 1990; Peters 1993). In short, implementation theory is characterised by no single and consistent conclusion which allows for general policy recommendations in order to improve implementation effectiveness.

The empirical richness of the implementation case studies of the 1980s, in particular, contributed that scholars increasingly shied away from seeking a universal model (Mayntz 1983; Windhoff-Héritier 1987). Instead the focus was on mid-level conclusions (e.g. on the appropriateness of certain policy instruments in light of distinctive policy problems and context characteristics), in the best case amounting to general contingency models (Hanf/Scharpf 1978; Linder/Peters 1989; Ingram/Schneider 1990; Peters 1993). But even these mid-level or contingent models, identifying different problem/context constellations in which either top-down or bottom-up measures are expected to be more successful, ended up contradicting one another. For instance, Ingram and Schneider (1990) argue that in constellations of low support for a policy, bottom-up instruments emphasising learning and support building will lead to better implementation performance than detailed top-down intervention. This view contrasts with an argument advanced by Cerych and Sabatier (1986), stating that clear and specific objectives might enhance learning by lower level agents because it produces obvious performance indicators. Another example is the recommendation to apply bottom-up concepts in constellations characterised by high uncertainty and complexity in order to allow for sufficient flexibility to react to new developments in the light of specific context conditions and the generation of ideas useful for the further evolution of a policy (Ingram/Schneider 1990). Again, there are also good arguments to justify detailed top-down regulation in such constellations. First, too much discretion for subordinate agents might imply that nothing happens at all (Lane 1995: 112; Lübbe-Wolff 1996). Second, even a top-down policy that takes little account of the problem complexity, might succeed in stimulating learning processes by trial and error. Such list of contradictory advice that can be derived from implementation research could be continued. It illustrates the general ambiguity of implementation theory. Even attempts to classify contingencies, identifying particular problem or context constellations, fail to establish convincing causal linkages between policy choice and implementation effectiveness.

The IR literature on compliance is not to the same extent shaped by a polarised debate. Those that advocate the strengthening of the institutional structure and sanctioning potential of international regime (an approach in line with the 'top-down' approach in the public policy debate) have directed their argument against realist thinkers (s.a.) rather than the advocates of more flexible and 'bottom-up' measures. Hence, both approaches seem to exist side by side rather than in declared opposition to one another. Among the latter, there is a sense that the perception of poor implementation has as much to do with unrealistic expectations (cf. Chayes/Chayes 1993) as with 'treatable' factors. Considerations of possible 'treatments' to improve implementation, in turn, have remained on a very general level. Proposal on how to actually 'design' institutions or instruments to further capacity or concern building on the ground are left to future research (or other disciplines). Peter Haas, for instance, calls for further research "to combine institutional and constructivist analysis to better understand how institutional design can enhance learning" (1998, 33) and consequently better implementation, on the ground. Through such relatively modest claim making, IR scholars have avoided the same level of contradictory conclusions as the students of public policy.

Neither the ambiguous results of public policy studies nor the so far limited contributions from the perspective of International Relations have gone unnoticed. But both insights seem to have been forgotten in the EU context, considering the sometimes euphoric reception of new approaches by many political and societal actors involved.

To be sure, this euphoric response should not only be read as the result of Commission bureaucrats and member state politicians misunderstanding implementation theory. It must be traced to a number of political factors as well. Besides processes of simple trial and error learning (abandoning deficient top-down concepts in favour of new approaches) or mimetic learning (the imitation of successful domestic concepts), political and ideological aspects have had an impact on recent institutional and policy changes at the supranational level. The empowerment of sub-central levels of governance as well as the increasing reliance on market forces and industrial self-regulation resonate well with the current subsidiarity debate and the neo-liberal call to 'slim' the regulatory state. But even though political factors may have contributed to ignoring the theoretical ambiguities with respect to the merits of the new approach, the political reality of its increasing adoption justifies some further investigation of its inherent limits or problems.


Is 'New' Really New?

Implicit in the controversial discussion of the potential of the new approach is the assumption that such an approach is clearly delineated and that new policy instruments as well as the appropriate institutional innovations are distinct from traditional, top-down instruments and structures. Empirical evidence suggests that this distinctiveness is a myth, at least on the level of actual application. Most new instruments or institutions are hybrids, combining deliberative and discretionary with hierarchical and inflexible characteristics. As a consequence, they share many problems with traditional tools, both during the decision making process as well as at the implementation stage.

To elaborate, it is often argued that new instruments are preferable as they render greater political acceptability in the decision making as well as implementation process. Its flexibility and responsiveness to local framework conditions suggests that new policy designs are less prone to invite political conflict. Consequently, decision making is expected to be relatively smooth. Equally, given the context orientation and discretionary structure, new instruments imply a lesser imposition on implementing actors than classical regulation, hence favouring acceptance and compliance also on the ground. Political acceptance is expected to increase even more if deliberative institutional structures are developed in addition. The reality however looks less rosy.

Several studies of environmental policy making show that political conflict does not stop, or even diminish, at the issue of new instruments. The work by Héritier, Knill and Mingers (1996) for instance investigates in some detail the conflicts around the adoption of the EMAS Regulation and the Access to Information as well as the Environmental Impact Assessment Directives. Also institutional innovations are far from unproblematic, considering for instance the protracted battle prior to the establishment of the European Environment Agency (Jimenez-Beltran 1996). These experiences show that measures of the new approach are rarely free from competency conflicts nor do they 'fit' every regulatory system. Quite on the contrary, because new instruments typically focus on the procedural level — e.g., establishing channels for awareness raising and learning — they may impose quite significant reform pressures for national and local administrations. Hence, it is doubtful that new instrument are really 'new' in the sense of a reduced adaptation challenge for the decision makers and implementing actors.

Take for example the Environmental Impact Assessment Directive. In administrative terms the Directive assumes horizontally integrated structures that enable a comprehensive assessment of any public or private project across environmental media, such as water or air as well as across geographical areas (regardless of political or administrative boundaries). Both administrative requisites violate against the federal and hierarchical structures in Germany which therefore fought the Directive in the decision making process and now stands out with a minimalistic approach to implementation (Héritier/Knill/Mingers 1996; Knill/Lenschow 1998). Similar problems have arisen in Italy due to its fragmented administrative structure that is characterised the duplication of competencies (Caddy/Favoino 1997). It seems doubtful that the substantive objectives of the Directive will be reached this way.

Equally apparent is the considerable adaptive challenge in the case of the Directive on the Freedom of Access to Environmental Information. Even though this Directive has been phrased quite openly, it also includes several procedural requirements that imply far-reaching changes in the administrative practice of some member states. Specifically, the fact that the Directive applies to everybody, regardless of personal interests or motives, is in contradiction to the definition of information rights in most member states and has resulted in restrictive and ineffective implementation of the Directive (Kimber forthcoming). Once again, Germany stands out as particularly resistant. German administrative law provides for access to data only in the case of personal involvement in the procedure to which the data apply (Winter 1996).

The Eco-Label Regulation represents even more openly a hybrid of old (hierarchical) and new (self-regulatory) features. This combination contributed that the Regulation has been considered a 'flop' so far. The Regulation includes detailed instructions on how to develop evaluation criteria and the certification of environment-friendly products. Even though the procedures provide for the wide inclusion of affected industries and national authorities, the final decision over those criteria and measurements that will apply EU-wide is taken 'hierarchically' by the Commission. This is in stark contrast to the initial idea of industrial self-regulation and has led to lags in the implementation of the Regulation due to diverging interests of the different national authorities and industrial actors involved in the process (Wright, forthcoming).

Political conflict and administrative resistance occurring during the implementation of new instruments follow not only from their 'lack in purity' resulting in the presence of elements that contradict national structures despite the assumed flexibility of the policy tool. Conflict and resistance may have even deeper reasons, touching on principal questions of regulatory philosophy. It must be realised that new and old regulatory approaches are rooted in a more general belief system (or causal theory; cf. Sabatier 1986) and certain assumptions with regard to human behaviour in the policy process. In brief, advocates of the new approach trust in citizens' self-initiative (particularly if the institutional framework and knowledge base facilitates initiative) while classical regulation assumes more passive but typically law-abiding citizens (especially if the threat of sanctions adds pressure to comply). In most member states such either society- or state-centred thinking has become institutionalised in the form of regulatory philosophy or administrative traditions and will influence the decision making process (Dyson 1980; Damasky 1986; van Waarden 1995). Hence, not all policy types will resonate equally well with the national administrative tradition. Even though the concrete challenges for adaptation may be lower for new instruments (which is already a questionable assumption, as indicated above), on the level of regulatory philosophy the political, and in the end also administrative, challenge for adaptation may be substantial (Knill/Lenschow 1998). Here exists no principal difference between old and new instruments, even in their 'pure' form.

Returning to the implementation of the Information Directive, resistance in this case is not merely due to the concrete adaptive challenged implied in its procedural demands; even more importantly, the Directive contradicts administrative traditions in many member states (Kimber, forthcoming). In Germany it is the Rechtsstaatsprinzip which emphasises the protection of individual rights and therefore limits access to administrative procedures to those persons directly affected (Burmeister/Winter 1990). Similarly restrictive responses to the Directive in Spain and Italy can be traced to their authoritarian state traditions (Börzel forthcoming; Caddy/Favoino 1997). Even in Great Britain the Directive collided initially with the traditional 'secrecy' in British regulatory practice. A more effective implementation was possible only after fundamental administrative reforms in the greater context of a neoliberal trend in Britain pointed toward a less state-centred, more transparent and open administrative practice (Knill 1998).

By contrast, the Eco-Audit Regulation fits very well into the British tradition of procedural governance and industrial self-regulation. Also in the German case, the Regulation corresponds with general patterns of industrial self-regulation in the larger context of corporative governance (Benz/Götz 1996). More problematic, however, has been the implementation of the Regulation in France. Here, market-oriented instruments are mediated in their practical application by the strong influence of the French administrative, whose role is traditionally dominant in the French governance system (Mény 1993; de Montricher 1996). As a consequence, French industry has signed up very hesitantly to the Eco-Audit scheme. It fears that environmental data that are revealed in the Eco-Audit verification process will subsequently be used by the administration against industry in authorisation procedures and the like (Bailey 1997).

Are new instruments and structures really new; that is, do they escape some of the problems usually attributed to traditional regulatory measures? The answer to these questions seems to be a rather unambiguous No. Most new instruments are in fact hybrid solutions, implying, usually on the procedural dimension, considerable adaptation pressure for at least some EU member states. From this follows political conflict in the policy formulation and decision making stages, deficits in the resulting compromise solutions, and the rather doubtful acceptance on the part of the policy implementers as their discretion to mould the instrument according to local needs and facilities is in fact more limited than expected. Even if new instruments and institutions are pure in their novelty with respect to the substantive requirements, we must realise that they cannot be new on the level of state and legal traditions. Here they may imply enormous challenges to at least some member states.


Is 'New' Really Better?

The previous discussion was not supposed to imply that so-called new instruments are in fact no different at all from traditional approaches. The intention was to show that the contrast is less encompassing than often thought and that both approaches share some characteristics, and hence problems. The main difference between old and new policies remains the more procedural emphasis of the latter, going hand in hand with more flexible and open-ended implications on the substantive side. This openness to context factors of the new approach may imply its own set of problems, however.

A somewhat idealistic and ad hoc notion of 'context orientation' among advocates of the new approach has resulted in a rather wide-spread unawareness of inherent problems. This idealistic notion may be a consequence of the rather vague conceptualisation of context factors in implementation theory; these factors are entirely contingent and not embedded in a greater theoretical framework that helps identify the actors, institutions, ideas, as well as the linkages or dynamics between them. What is the relevant context and what are its implications for the operation of new instruments? To what extent can the policy context be modified by the requirements of new instruments? What Peter Haas (1998, 33; see above) has identified as issues in need for further research is not sufficiently problematised in the discussion on bottom-up policy approaches and new instruments.

The first question that arises is whether the open-ended nature of new instruments, as opposed to clearly specified obligations, really provides sufficient incentives for motivating and mobilising sufficient administrative and societal support in favour of effective implementation. As part of its open and flexible design, the new approach assumes the presence of motivated and capable implementers who are willing and able to work with the policy instrument in the local context and toward the intended goal. There is evidence that questions the assumption of principally motivated actors as it is implicit in the new approach, however. For instance in the case of the already mentioned Information Directive, the rather vague definition of the authorities and the kind of data that are affected by the Directive has invited public administrations to interpret their obligations narrowly and the exemptions widely. In the end, hardly any changes to national practices were deemed necessary and the objectives of the Directive often dodged (Scherzberg 1994; Kimber forthcoming).

Equally the implementation of the EIA Directive has suffered from (to) narrow interpretations of its vague formulations. Especially the question if a favourable EIA should serve as a precondition for a project's authorisation is left unanswered in the Directive. This ambiguity has been exploited widely in the practical implementation and rendered the EIA practically meaningless (Bailey 1997; Alder 1993; Héritier/Knill/Mingers 1996). It appears that the avenues for wider public participation serve primarily to legitimise a project rather than to take account of public concerns.

Moreover, the implementation of the Eco-Audit Regulation indicates that the unclear incentive structure emanating from many new instruments is may also be due to the weak integration with already existing regulatory instruments. The Regulation does not include rules or guidelines concerning a possible easing of authorisation or monitoring procedures for those industrial actors that participate in the Eco-Audit Scheme. Also the link to other management schemes, such as the British Standards Institution or the international Standards Organisation, has long been unclear (Zito/Egan 1997). As a consequence, the incentives to implement an environmental management scheme are too weak for many firms and the participation rate relatively low. At the same time, the intended financial incentive of the Regulation — produced by stimulating the awareness of the cost factor of environmental pollution — is undermined by the negligence of the polluter-pays principle in traditional, regulatory environmental policy. Only in Germany, and partly due to explicit efforts to increase the coherence of regulatory measures, the industrial participation in the Eco-Audit Scheme is far above the European average (Bouma forthcoming; Knill/Lenschow 1998).

Secondly, and independent of vagueness in formulation or unclear incentive structures, new instruments are rooted in the general assumption that actors may be mobilised. In reality the group of effectively stimulated actors tends to be small and limited to those actors are both in the possession of necessary resources and capacities and are at least partly mobilised already. Therefore, the effectiveness of new instruments is highly constrained in countries where the environmental movement is weak and lacks resources. These countries on the other hand are the ones that may be the primary addressees of 'bottom-up' measures (Börzel forthcoming; Caddy/Favoino 1997).

It is interesting that the issue of 'capacity raising' has been much more central in the international debate to enhance the effectiveness of environmental regimes than in the EU. 4 The transfer of resources may be a more 'ordinary' political instruments in these fora — due to its seemingly limited intrusion in domestic politics on the one hand and the clearly enormous income gap between the members of international environmental regimes on the other hand. The issue of procedural innovation as implied in the new instruments just described has been raised only recently and not yet resulted in the same regulatory activity. Nevertheless, the problems raised so far will equally apply to international regimes once efforts for "concern building" (Zürn 1995, 20) will become more manifest.

In summary, aside from sharing some characteristics and problems with the classical regulatory approach, new instruments have a number of inherent weaknesses. Flexible design and open institutional and procedural structures may create confusion rather than incentives to act. Furthermore, the often indirect signals sent by new instruments presuppose favourable context factors. There is some evidence, however, that the openness and so-called context orientation of new instruments invites administrative resistance or sabotage as much as they may open new channels for learning and flexible adaptation to new challenges. Advocates of new instruments seem to overestimate the (positive) malleability of the local context. In practice, the context orientation of the new approach works only for a limited set of already favourable context constellations (e.g., low initial opposition, sufficient capacity and resources, some initial motivation, complementarity with regulatory and institutional structures, etc.). Ambiguous theory has helped to mystify particularly the latter aspect. To push the theoretical framework further in this respect, we need to develop general indicators concerning the malleability of the context. We argue in the next part that an institutional perspective contributes to distinguish more or less malleable and/or favourable framework conditions.


Addressing the Deficits: Towards an Institutional Perspective in Implementation Research

In previous parts we have presented analytically derived arguments and some empirical evidence that new institutional designs and new instruments may not contribute as expected to reduce the implementation deficits that were widely attributed to the predominant use of hierarchical, command-control structures and instruments. Our analysis has pointed to an overstated distinction between old and new as well as several particular weaknesses of new instruments in order to explain this disappointment. In this part we will show that these explanations have an institutional perspective in common.


Implementation and Institutions: The Argument

Notwithstanding the overall revival of neo-institutional approaches, implementation research has remained surprisingly unaffected by these theoretical developments. This may be due to the fact that implementation research was already an 'industry in decline', when institutions came "back in" to the work of political scientists from the mid-1980s onwards (Evans/Rüschemeyer/Skocpol 1985). Only recently, and especially in the context of research on the implementation of EU policy in the member states, scholars began to analyse the impact of institutions on implementation effectiveness in an explicit and more consistent way (Duina forthcoming; Knill 1998; Knill/Lenschow 1998; Timmermans et al. 1998).

In the past, institutions were integrated in the analysis merely from the perspective of adequate design. Analysts asked what are the optimal structural and organisational arrangements that would permit effective implementation of a certain policy (cf. Pressmann/Wildavsky 1973; Scharpf 1978). This thinking relies on the implicit assumption that national institutions would easily adapt to the suggested 'model' structure. Problems of institutional change were ignored. Many IR scholars that deal with the compliance with international regimes similarly tend to emphasise the structure of decision making and its role for subsequent effective implementation. Here we discover a trend to favour more inclusive, information and communication enhancing structures (cf. Zürn 1995, 20 for an overview; Dilling forthcoming and Favoino 1997 note similar trends in EU environmental politics). This perspective continues to assume the malleability of existing institutional structures. Analysts are interested in the impact of varying institutional designs on the skills, resources and capacities of relevant actors. They ask what is the perfect design facilitating the supply of implementing authorities with sufficient financial, legal and personal resources.

Without denying the importance of adequate institutional design, we argue that such a perspective remains incomplete as long as it ignores the problems associated with the process of adjusting the existing institutional arrangements to the defined 'ideal' arrangements. Favoino (1997) drives home this point in discussing explicit attempts to improve the context for policy implementation through institution building in the EU Commission. Favoino shows that institutional reforms will only be successful if there is some minimum congruence with existing institutional arrangements.

Problems of institutional adjustment are not restricted to measures which are explicitly directed at institutional change or reforms but equally relevant for the implementation of individual policy instruments. As elaborated above, regulatory policies often require changes of well-established administrative structures, procedures and practices at the domestic level, hence create pressures for institutional adaptation during the process of implementation. This 'by-product' of policy making is often overlooked in implementation research, either due to its problem-oriented focus on adequate policy design or isolated reception and perception of certain policies in their specific context (including the interests and capabilities of implementers and target groups). Furthermore, as outlined above it is often wrongly assumed that the institutional adaptation pressure of new instrument is negligible due to their less impositional nature.

Linking EU policy making and effective implementation through the intermediate step of institutional adaptation is not only empirically but also analytically relevant. It is one of the few generally accepted findings in the otherwise diverse neo-institutionalist literature (Hall and Taylor 1996) that institutional change, regardless if required explicitly or implicitly, rarely takes place in a smooth and unproblematic way. Existing institutions 'matter', and they do so mainly by constraining the options for future changes and adaptations. Thus, an institution’s normal life is characterised by persistence and continuity; fundamental or path-breaking changes are confined to exceptional situations or external shocks.

The strong focus in the more recent IR literature on capacity building reflects an awareness of the limits of adaptive capacities on the ground. Underdal states even more generally that "compliance will, in certain circumstances, tend to be inversely related to the amount of behavioural change required by regime rules" (1998: 12). However this literature ignores the institutional roots of resistant behaviour and more or less adaptive capacities. If more differentiated observations are made at all, it focuses on the behavioural change asked from 'the state' (implied seems a unitary actor image) or powerful domestic interests (see Underdal for a good review of the literature). A focus on 'the state' is too undifferentiated for the analysis of the implementation of regulatory policies which at the very least involve the administration on the national and often sub-national level. 5 Also the primary focus on domestic interest seems problematic in our investigation of the potential of new instruments. These instruments target the perception of interests by domestic actors through information exchange and learning procedures. Hence it is important to ask under what conditions are some interests more malleable than others rather than assuming interests as fixed. Here we suggest an institutional approach and argue as follows.

Institutional persistence increases with an institution’s depth and breadth (Krasner 1988). Institutional depth refers to the extent to which institutions structure preferences and beliefs of actors by providing cognitive frames of reference. For instance, the paternalistic state tradition with a strong emphasis on the rule of law in Germany pre-structures the administration's behaviour in the environmental field; instruments implying administrative discretion or a transparent relationship with the public are perceived as misfits in this regulatory framework and hence resisted. Institutional breadth, in turn, relates to the costs of institutional adaptation, including not only the number of additional institutional changes required as a result of comprehensive institutional linkages, but also the institutional impact on the distribution of power and resources between different actors. The deficits in implementing the EU Environmental Impact Assessment Directive are related to this aspect. The Regulation implies the establishment of broad institutional linkages in order to organise an integrated, cross-media assessment. Hence, different ministries (e.g., industry, agriculture, land planning, environment, health, etc.) and/or administrative agencies need to coordinate their procedures in order to facilitate the full implementation of the Directive. The institutional breadth of this requirement has been one of the reasons for resistance (cf. Börzel and Caddy forthcoming).

The emphasis on institutional continuity and stability in the neo-institutional literature does not imply that institutions are entirely static. Indeed, institutions constantly adapt to external pressures. However, the scope of these adaptations is constrained by the existing institutional arrangements; that is, the institution’s "roots and routes" (Olsen 1995, 6). Institutional change is expected to be incremental and path-dependent. Consequently, changes tend to occur on the level of "second order changes" (Hall 1993), "secondary aspects" (Sabatier 1993) or "changes within the core" (Knill/Lenschow 1998). They do not challenge an institution’s very identity.


Implications for Implementation Research

Applying the neo-institutional theory to problems of effective implementation, we find that an institutional perspective may help to address conceptual deficits in two ways. First, it may reduce theoretical ambiguity by focusing on aspects of implementation which existing approaches so far have treated as a 'black box'. Second, an institutional approach allows for a more systematic account of the impact and malleability of 'context conditions' on implementation effectiveness.

It follows from these neo-institutional insights that, rather than being affected by the choice of the policy approach per se, effective implementation is basically dependent on the degree of institutional fit between existing institutional arrangements and the institutional implications emerging from European policies. Implementation is likely to be ineffective if the institutional implications of EU policies contradict strongly entrenched patterns of already existing institutions. Such contradictions may exist on the level of regulatory structures, hence they could affect institutional innovations as they are presently designed to ease the European policy formulation and implementation process, such as independent agencies or organisational networks. Contradictions may equally exist on the level of regulatory practices and procedures as implied in environmental policies. Let us repeat again, new instrument are not excluded from this possible problem as many of them also assume high adaptive capacity of the institutional framework.

'Pure' new instrument, of course, should not assume certain structures on the ground but rather facilitate the adaptation of the policy tool to the respective 'context conditions'. The notion of context, however, remains unspecified in implementation theory and problematic in practice. In practice the policy's flexibility with respect to those context factors that are very stable and inflexible may be more important than flexibility and regulatory discretion with respect of those factors that would relatively easily adapt to the requirements of the policy instrument. We suggest that an institutional perspective would advance our understanding of the nature of context factors and their impact on the implementation process: One would need to proceed by first analysing the level of institutional embeddedness of the various relevant context factors and then compare the found levels of context malleability with the respective degrees of actual flexibility implied in the policy. We may find for instance that a policy is highly flexible with respect to the exact administrative procedures selected to achieve the policy's objective while assuming that the administration is capable of choosing an appropriate procedure; turning to the actual context, we may face an administrative structure that has no tradition and hence capacity of independent decision making on the part of implementing officials. In such case, the so-called context orientation of the policy instrument has actually hampered effective implementation as the institutional dimension of the relevant context has been ignored.


Implications for European Problem Solving Capacity: 'Bounded Innovation'

Applying an institutional perspective to policy implementation may also add some insights to the present theoretical and practical debate on the problem solving capacity of the EU multi-level system. In the past, this debate has primarily focused on the policy formulation phase. Decision traps were identified that hinder the adoption of optimal policy solutions and hence limit top-down problem solving (Scharpf 1985). This aspect of the theoretical discussion is not further advanced in this paper. Nevertheless revealing has been the rather clear evidence that old as well as new policies may fall victim to such decision traps. The argument that new instruments, due to their open and flexible texture, are less likely to be trapped between conflicting interests could not be confirmed.

Our main concern is the linkage between effective implementation and problem solving. We have argued that it makes sense to focus on compliance with specified policy objectives as a pre-stage and -condition to actual problem solving in terms of policy outcomes. Following our theoretical and empirical considerations, new instrument are not necessarily superior to old instrument in this respect. Both instrument types may demand a too high level of institutional adaptation in order to set in motion a process of problem solving — instead, the policy may be resisted by its implementers and/or addressees.

This conclusion may yield a general policy dilemma for the problem solving capacity of the EU multilevel system: Imposing the 'ideal' policy design from the top while ignoring the potential impact of existing institutions may significantly reduce the chance for effective problem solving due to administrative resistance to adapt. Designing policy in conformity to existing institutions, however, may reduce the possibility to generate changes and solve problems (cf. Timmerman et al. 1998). New instruments may face the related dilemma: Those measures that approach the issue of problem solving indirectly, by facilitating local learning processes, may impose a too minor (or indirect) challenge for triggering any reaction. Hence, local problem solving capacity remains static as the existence of new institutional avenues or new instruments are ignored in practice. This result may be most likely in low capacity and low mobilisation situations, for instance in Southern and Central-East European states (cf. Börzel; Caddy; Knill/Lenschow all forthcoming). Policy designs that are most likely to have environmentally positive effects may be those that are least context friendly, in terms of addressing the limits of political and administrative feasibility, or vice versa.

To overcome these problems, policy makers have to develop policies that require something, but not too much from member states. In other words, there exists a curvilinear relationship between the level of adaptation pressure implied in EU policies and the effectiveness of implementation (cf. Cerych/Sabatier 1986). Only if EU legislation fits into the 'bounded space for innovation', it — be it old or new — may contribute to problem solving in the EU.



In this paper we revisited a sub-discipline in public policy, namely implementation research, that has become highly salient for EU environmental policy in recent years, but continues to suffer from theoretical weaknesses and ambiguities. EU policies are based on some of these ambiguous theoretical considerations; we have shown that the EU's shift towards a new approach in institution building and the choice of policy instruments may have been affected by a superficial analysis of the true character of the chosen measures and a unduly optimistic perspective of the anticipated effects.

Not denying the relevance of many factors and both perspectives, we have identified in this chapter a 'black box' which most models in implementation research have in common, namely the institutional framework. We suggest that focusing on this black box will help resolving some of the ambiguities and contradictions in existing theory. Looking at institutional adaptation pressure helps us qualify both the conditions under which a certain policy design is capable of solving the problem addressed and the conditions under which bottom-up 'learning' tools will actually find some resonance in the given institutional and political context. Hence the institutional perspective advances top-down and bottom-up policy perspectives and offers a bridge that links arguments of an often polarised debate.

An institutional perspective to policy implementation may provide some additional insights to the present discussion on the problem solving capacity of the EU multi-level system. This debate has been primarily concerned with the European capacity to formulate and develop regulatory policies. The Community’s capacity to solve regulatory problems is in a similar way defined by the extent to which its policies are actually implemented. Notwithstanding the dependency of implementation effectiveness on the prior formulation phase, we argue that additional insights can be gained from a separate analysis of the implementation phase and conclude that European policies, regardless whether they are of the old or new type, only contribute to the problem solving capacity of the EU when they contain substantive but no fundamental institutional implications for existing arrangements.

Finally, we have noted in this paper interesting parallels in the Public Policy and IR literature on the problem solving capacity of the EU and international regimes. In both sub-disciplines there has been a shift in attention to the issue of compliance or implementation of EU or international norms. Neither sub-discipline conducts a systematic analysis of the institutional challenge implied in EU and international norms. In the case of international norms the requirements effecting institutional adaptation may indeed be smaller than this is the case with EU directives and regulations, however such adaptation pressure also exists as recent calls for capacity and concern raising measures indicate. So far, an institutional framework of analysis has been proposed in the IR literature only for a future research agenda. This is an agenda that IR and Public Policy have in common.



Alder, John (1993). Environmental Impact Assessment - The Inadequacy of English Law. Journal of Environmental Law 5 (2), 203-20.

Baier, Vicki E., James G. March, and Harald Sætren (1990). Implementierung und Ungewißheit. In: James G. March, ed. (ed.), Entwicklung und Organisation. Kritische und konstruktive Beiträge, Entwicklungen und Perspektiven. Wiesbaden: Gabler, 170-184.

Bailey, Patricia (1997). The Implementation of EU Environmental Policy in France. In: Christoph Knill, ed., The Impact of National Administrative Traditions on the Implementation of EU Environmental Policy . Florence: European University Institute.

Benz, Arthur und Klaus H. Goetz (1996). The German Public Sector: National Priorities and the International Reform Agenda. In: Arthur Benz und Klaus H. Goetz, eds., A New German Public Sector? Reform, Adaptation and Stability . Aldershot: Dartmouth, 1-26.

Berman, Paul (1980). Thinking About Programmed and Adaptive Implementation: Matching Strategies to Situations. In: Helen Ingram and Dean Mann, (eds.). Why Policies Succeed or Fail . London: Sage, 205-27.

Börzel, Tanja A. (forthcoming). Improving Compliance through Domestic Mobilisation? New Instruments and the Effectiveness of Implementation in Spain. In: Christoph Knill und Andrea Lenschow, eds., Implementing EU Environmental Policy: New Approaches to an Old Problem . Manchester: Manchester University Press.

Bouma, Jan Jaap (forthcoming). Environmental Management Systems and Audits as Alternative Environmental Policy Instruments. In: Christoph Knill und Andrea Lenschow, eds., Implementing EU Environmental Policy: New Approaches to an Old Problem . Manchester: Manchester University Press.

Burmeister, Joachim H. und Gerd Winter (1990). Akteneinsicht in der Bundesrepublik. In: Gerd Winter, ed., Öffentlichkeit von Umweltinformationen. Europäische und nordamerikanische Rechte und Erfahrungen . Baden-Baden: Nomos, 87-128.

Caddy, Joanne (forthcoming). Implementation of EU Environmental Policy in Future Member States. The Case of EIA. In: Christoph Knill und Andrea Lenschow, eds., Implementing EU Environmental Policy: New Approaches to an Old Problem . Manchester: Manchester University Press.

Caddy, Joanne und Marieava Favoino (1997). The Implementation of EU Environmental Policy in Italy. In: Christoph Knill, ed., The Impact of National Administrative Traditions on the Implementation of EU Environmental Policy . Florence: European University Institute.

CEC (Commission of the European Communities) (1993). Towards Sustainability. A European Community Programme of Policy and Action in Relation to the Environment and Sustainable Development. Luxembourg: Office for Official Publications of the European Communities.

CEC (Commission of the European Communities) (1996).Thirteenth Annual Report on Monitoring the Application of Community Law (1995). Luxembourg: Office for Official Publications of the EC

Cerych, Ladislav and Paul Sabatier (1986). Great Expectations and Mixed Preferences: The Implementation of European Higher Education Reforms. Stoke on Trent, UK: Trentham Books.

Chayes, Abram and Antonia Handler Chayes (1993). On Compliance. International Organisation 47(2), 175-205.

Collins, Ken and David Earnshaw (1992). The Implementation and Enforcement of European Community Environment Legislation. Environmental Politics, 1 (4), 213-249.

Damaska, Mirian R. (1986). The Faces of Justice and State Authority. A Comparative Approach to the Legal Process. New Haven: Yale University Press.

Dilling, Rasmus (forthcoming) Improving Implementation by Networking: The Role of the European Environment Agency. In: Christoph Knill, ed., The Impact of National Administrative Traditions on the Implementation of EU Environmental Policy . Florence: European University Institute.

Duina, Francesco (1999). Harmonizing Europe: Nation-States within the Common Market. New York: Suny Press.

Dyson, Kenneth (1980). The State Tradition in Western Europe. A Study of an Idea and Institution . Oxford: Martin Robertson.

European Parliament (EP) (1993). Application of the Principle of Subsidiarity to Environment and consumer Policy. Europe Environment 402.

European Parliament (EP), Committee on the Environment, Public Health and Consumer Protection (1996). Working document on implementation of Community environmental law (Rapporteur: Ken Collins), PE 219.420, Brussels.

Evans, Peter, Dietrich Rüschemeyer, Theda Skocpol, eds. (1986). Bringing the State Back In. Cambridge: Cambridge University Press.

Favoino, Marieva (1997) Commission's Internal Implementation of the Environmental Integration Principle: Towards a Framework of Analysis. In: Christoph Knill, ed., The Impact of National Administrative Traditions on the Implementation of EU Environmental Policy . Florence: European University Institute.

Golub, Jonathan (1993). Recasting EU Environmental Policy: Subsidiarity and National Sovereignty. In Ute Collier, Jon Golub, and Alexander Kreher, (eds). Subsidiarity and Shared Responsibility New Challenges for EU Environmental Policy. Nomos.

Haas, Peter M. (1998) Compliance with EU directives. Insights from international relations and comparative politics. Journal of European Public Policy 5(1), 17-37.

Haas, Peter M., Robert O. Keohane, Marc A. Levy (1993), eds.. Institutions for the Earth. Sources of Effective International Environmental Protection Cambridge/MA.

Haigh, Nigel (1996) The Manual of Environemntal Policy: the EC and Britain. London: Catermill Publishing.

Hall, Peter A. (1993). Policy Paradigms, Social Learning and the State. The Case of Economic Policymaking in Britain. Comparative Politics 25, 275-296.

Hall, Peter A. and Rosemary C. R. Taylor (1996). Political Science and the Three New Institutionalisms. MPIFG Discussion Paper 96/6. Köln: Max-Planck-Instititut für Gesellschaftsforschung.

Hanf, Kenneth and Scharpf Fritz W. (1978), eds., Interorganizational Policy Making: Limits to Coordination and Central Control . London: Sage.

Hasenclever, Andreas, Peter Mayer, Volker Rittberger (1997) Theories of international regimes. Cambridge, Cambridge University Press.

Héritier, Adrienne, Christoph Knill and Susanne Mingers (1996). Ringing the Changes in Europe. Regulatory Competition and the Transformation of the State . Berlin: de Gruyter.

Ingram, Helen and Anne Schneider (1990). Improving Implementation Through Framing Smarter Statutes. Journal of Public Policy 10 (1), 67-88.

Jacobsen, Harold K. and Edith Brown Weiss (1995) Strengthening Compliance with International Environmental Accords: Preliminary Observations from a Collaborative Project. Global Governance 1(1), 119-148.

Jimenez-Beltran, D. (1996). The European Environmental Agency. In: Alexander Kreher, ed., The New European Agencies - Conference Report . EUI Working Paper, RSC No. 99/49. Florence: EUI.

Jordan, Andrew (1999). Implementation: A problem without a political solution? Environment and Planning C: Government and Policy 17,

Kimber, Clíona (forthcoming). Implementing European Environmental Policy and the Directive on Access to Environmental Information. In: Christoph Knill und Andrea Lenschow, eds., Implementing EU Environmental Policy: New Approaches to an Old Problem . Manchester: Manchester University Press.

Knill, Christoph (1998). European Policies: The Impact of National Administrative Traditions on European Policy making. Journal of Public Policy 18 (1), 1-28.

Knill, Christoph and Andrea Lenschow (1998). Coping with Europe: The Implementation of EU Environmental Policy and Administrative Traditions in Britain and Germany. Journal of European Public Policy 5(4).

Knill, Christoph and Andrea Lenschow (forthcoming). Do New Brooms Really Sweep Cleaner? Implementation of New Instruments in EU Environmental Policy. In: Christoph Knill, ed., The Impact of National Administrative Traditions on the Implementation of EU Environmental Policy . Florence: European University Institute.

Kohler-Koch, Beate (1989). Regime in den internationalen Beziehungen. Baden-Baden, Nomos.

Krämer, Ludwig (1992). Focus on European Environmental Law . London, Sweet & Maxwell.

Krasner, Stephen D. (1988). Sovereignty: An Institutional Perspective. Comparative Political Studies 21 (1), 66-94.

Lane, Jan-Erik (1995). The Public Sector. Concepts, Models and Approaches. London: Sage.

Lenschow, Andrea (forthcoming) Transformation in European Environmental Governance. In: beate Kohler-Koch and Rainer Eising, eds. The Transformation of European Governance London:Routledge.

Linder Stephen and B. Guy Peters (1989). Instruments of Government: Perceptions and Contexts. Journal of Public Policy 9 (1), 35-58.

Lipsky M. (1980). Street-Level Bureaucracy . New York: Russell Sage.

Lübbe-Wolff, Gertrude (1996). Stand und Instrumente der Implementation des Umweltrechts in Deutschland. In: Gertrude Lübbe-Wolff, ed., Der Vollzug des europäischen Umweltrechts . Berlin: Erich Schmidt Verlag, 77-106.

Mayntz, Renate (1983). Implementation von regulativer Politik. In: Renate Mayntz, ed. Implementation politischer Programme II. Opladen: Westdeutscher Verlag, 50-74.

Mény, Yves (1993). Government and Politics in Western Europe. Britain, France, Italy, West Germany . Oxford: Oxford University Press.

Montricher, Nicole de (1996). France: In Search of Relevant Changes. In: Johan P. Olsen und B. Guy Peters, eds., Lessons from Experience. Experimental Learning and Administrative Reform in Eight Democracies . Oslo: Scandinavian University Press, 243-77.

Olsen, Johan P. (1995). Europeanization and Nation State Dynamics. Working Paper 9/95. Oslo: ARENA.

Peters, B. Guy (1993). Alternative Modelle des Policy-Prozesses: Die Sicht "von unten" und die Sicht "von oben". In: Adrienne Héritier, ed., Policy-Analyse. Kritik und Neurorientierung, PVS Sonderheft 24/1993, 289-306

Pressman, J. and A. Wildavsky (1973). Implementation . Berkely: University of California Press.

Rehbinder, E. and R. Stewart (1985). Environmental Protection Policy. Integration Through Law. Berlin: de Gruyter.

Sabatier, Paul A. (1986). Top-Down and Bottom-Up Approaches to Implementation Research. Journal of Public Policy 6, 21-48.

Sabatier, Paul A. (1993). Advocacy-Koalitionen, Policy-Wandel und Policy-Lernen: Eine Alternative zur Phasenheuristik. In: Adrienne Héritier, ed., Policy-Analyse. Kritik und Neurorientierung, PVS Sonderheft 24/1993, 116-148.

Sand, Peter H. (1992) The Effectiveness of International Environmental Agreements. Oxford, Grotius.

Scharpf, Fritz W. (1978). Interorganizational Policy Studies: Issues, Concepts, and Perspectives. In: Kenneth Hanf and Fritz W. Scharpf, eds. Interorganizational Policy making: Limits to Coordination and Central Control . London: Sage, 345-70.

Scharpf, Fritz W. (1985). Die Politikverflechtungs-Falle: Europäische Integration und deutscher Föderalismus im Vergleich. Politische Vierteljahresschrift (4), 323-356.

Scharpf, Fritz W. (1994). Mehrebenenpolitik im vollendeten Binnenmarkt. MPIFG Discussion Paper 94/4. Köln: Max-Planck-Instititut für Gesellschaftsforschung.

Scherzberg, Arno (1994). Freedom of Information - deutsch gewendet: Das neue Umweltinformationsgesetz. Deutsches Verwaltungsblatt 109, 733-745.

Timmerman, Arco et al. (1998). The Design of Policy Instruments: Perspectives and Concepts. Paper presented at the 56th Annual Meeting of the Midwest Political Science association, Chicago, 23-25 April.

Underdal, Arild (1998). Explaining Compliance and Defection: Three Models. European journal of International Relations 4(1), 5-30.

Victor, David (1995). The IIASA Study on the Implementation and Effectiveness of International Environmental Commitments. Case Studies, Context, Hypotheses, Definitions, Research Methods. Laxenburg.

Waarden, Frans van (1995). National Regulatory Styles. A Conceptual Scheme and the Institutional Foundations of Styles. In: Brigitte Unger and Frans van Waarden (eds.), Convergence or Diversity? Internationalization and Economic Policy Response. Aldershot: Avebury, 333-372.

Wallace, William (1983). Less than a Federation, More than a Regime: The Community as a Political System. In: Helen Wallace et al., eds. Policy-Making in the European Community, 2nd edn. Chichester: John Wiley and Sons, 403-436.

Ward, Neil, Phillip Lowe and Henry Butler (1997) Implementing European Water Quality Directives. In: S.Baker, M.Koussis, D.Richardson and S.Young, eds., The Politics of Sustainable Development London, Routledge.

Windhoff-Héritier, Adrienne (1987). Policy-Analyse. Eine Einführung . Frankfurt: Campus.

Winter, Gerd (1996). Freedom of Environmental Information. In: Gerd Winter, ed., European Environmental Law. A Comparative Perspective. Aldershot: Dartmouth, 81-94.

Wright, Robert (forthcoming). Implementing Voluntary Policy Instruments: The Experience of the EU Ecolabel Award Scheme. In: Christoph Knill und Andrea Lenschow, eds., Implementing EU Environmental Policy: New Approaches to an Old Problem. Manchester: Manchester University Press.

Zito, Anthony R. and Michelle Egan (1997). Environmental Management Standards, Corporate Strategies and Policy Networks. Manuscript. Florence: European University Institute.

Zürn, Michael (1996). Die Implementation internationaler Umweltregime und 'positive Integration'. MPIFG DiscusionPaper 96:3.



Note 1: Fritz W. Scharpf (1985, 1994) has noted in this context the problems of joint decision making, pointing out the particular difficulties in the area of 'positive integration', or market correcting regulatory measures. Back.

Note 2: The 'top-down' approach has been implicit in discussions of widening the sanctioning capacities of international regimes. Back.

Note 3: To our knowledge, an adaptation of existing regimes in line with these proposals has not yet occurred. Also, on the international scene an analysis of the effectiveness of the 'new approach' is still outstanding.Back.

Note 4: In the EU the Cohesion Fund has been established in 1992 to facilitate resource transfers in the field of environmental politics. Back.

Note 5: Admittedly, the regulatory impose of international regimes may be less extensive than EU regulations; nevertheless national administrations are likely to be involved in the implementation even here. Back.