EIoP: Text 2000-008: Full text

Whither Europeanization? Concept stretching and substantive change(*)

Claudio M. Radaelli

European Integration online Papers (EIoP) Vol. 4 (2000) N° 8; http://eiop.or.at/eiop/texte/2000-008a.htm

Date of Publication in

: 17.7.2000

Introduction

Europe matters, but how? The political systems of the European Union (EU) member states are penetrated by European policies, but what is the effect of this process? Is Europeanization making the member states more similar? Or do different domestic political structures ‘refract’ Europeanization in different directions? Has ‘Europe’ changed domestic political structures (for example, party systems and public administration) and public policy? If so, what are the mechanisms of change?

The literature on Europeanization seeks to address these issues. There is a lively debate in this field and most of the intellectual energy is concentrated on empirical research. We do not know enough about the processes of Europeanization and its effects. Thus, empirical analysis is essential. However, it has to be accompanied by a delimitation of the concept. This paper argues that research on Europeanization runs the risk of conceptual stretching. The implications of sloppy conceptual frameworks should not be overlooked. Section 1, therefore, exposes these implications and suggests ideas in the direction of conceptual precision. Section 2 ‘unpacks’ the concept of Europeanization by using a simple taxonomy. Section 3 illustrates the main mechanisms of Europeanization, before the key explanatory variables are discussed in section 4. Section 5 presents suggestions for future research. This paper has two important limitations. Firstly, it is concerned eminently with the impact of Europeanization on member states, although there is evidence of Europeanization outside the current domain of the EU. Secondly, it has a political science – public policy analysis focus, and therefore does not say much on issues falling outside these disciplines.

1. The Concept

Conceptual analysis is a fundamental step in comparative political science, as shown by the perceptive recommendations formulated by Sartori in several essays (Sartori 1970, 1984, 1991). Concepts that are not well defined lead to confusion and elusive language. Concepts that do not specify the level of analysis generate mistakes in terms of the ‘ladder of abstraction’, that is, they obfuscate the relations between genus and species. Concepts without negation are universals: they point to everything, ‘conceptions without specified termination or boundaries’ (Sartori 1970:1042).

What is the state of the current research on Europeanization in the light of conceptual analysis? The problem – I would argue – is not that different authors assign different meanings to Europeanization – this is an indicator of the vitality of the debate. Instead, the potential risks refer to (a) concept misformation, (b) conceptual stretching and (c) ‘degreeism’.

A. Definitions and their problems

Lawton (1999), for example, suggests that Europeanization is the de jure transfer of sovereignty to the EU level, and distinguishes this concept from ‘Europeification’, that is, the de facto sharing of power between national governments and the EU. Thus, Europeanization and ‘Europeification’ are identified with the emergence of EU competencies and the pooling of power. Börzel (1999:574) draws instead attention to what happens once power has been transferred to Brussels. She defines Europeanization as a ‘process by which domestic policy areas become increasingly subject to European policy-making’. This is a notion which requires further specification, if one wants to distinguish between the simple fact that there is more ‘Europe’ in domestic policies and the more profound impact of the EU in policy areas which have now become dominated by a European logic of political behavior. Additionally, the reference to an ‘increasing’ role of European policy-making may lead to the conclusion that a very large number of policies are experiencing an ‘increased’ EU dimension: how does one distinguish between Europeanized policies and the ones which are still eminently domestic?

Caporaso, Green-Cowles and Risse, in their introduction to Green Cowles et al. (2000: Page numbers not available at this stage), provide yet another option with the following definition:

‘We define Europeanization as the emergence and development at the European level of distinct structures of governance, that is, of political, legal, and social institutions associated with political problem-solving that formalize interactions among the actors, and of policy networks specializing in the creation of authoritative rules’.

To begin with, the reader is struck by the emphasis on policy networks. Arguably, the authors use the notion of networks with wide latitude, that is, networks as patterns of social interaction present everywhere except under conditions of extremely autocratic rule. Yet the EU debate witnesses different positions on the usefulness of the network approach (see Kassim 1994 for a critical position). Additionally, at least in conventional analysis, networks are considered one possible ‘mode’ of governance (as opposed to corporatism, pluralism and statism; see Kohler-Koch 1999), not an ever-present phenomenon. Thus, one would think that the relevance of policy networks is a matter of empirical (rather than definitional) analysis. More importantly still, the emphasis on the ‘creation of rules’ and ‘the European level’ suggests an extremely broad notion of Europeanization, inclusive of both EU policy and politics and their repercussions on national systems. But if Europeanization has to have a precise meaning, it has to be different and more selective than the notion of EU policy formation and European integration. Common sense indicates that Europeanization has something to do with the penetration of the European dimension in national arenas of politics and policy, a point raised by Börzel (1999).

Ladrech heads towards a more promising direction when he puts emphasis on Europeanization as process. He argues that Europeanization is an:

‘Incremental process re-orienting the direction and shape of politics to the degree that EC political and economic dynamics become part of the organizational logic of national politics and policy-making’ (Ladrech 1994:69).

By ‘organizational logic’ he means the ‘adaptive processes of organizations to a changed or changing environment’ (Ladrech 1994:71). In doing so, he underlines the role of adaptation, learning and policy change. The emphasis on organizations is broad enough as to accommodate both processes wherein networks play a role and instances of Europeanization in which there are no networks at work. The drawback is that too much emphasis on organizations may obfuscate the role of individuals and policy entrepreneurs. Further, the object of Europeanization is limited to ‘national politics and policy-making’. Perhaps one could add identities and the cognitive component of politics (Checkel 2000; Schmidt 1999). Drawing upon Ladrech’s definition, I would argue that the concept of Europeanization refers to:

Processes of (a) construction (b) diffusion and (c) institutionalization of formal and informal rules, procedures, policy paradigms, styles, ‘ways of doing things’ and shared beliefs and norms which are first defined and consolidated in the making of EU decisions and then incorporated in the logic of domestic discourse, identities, political structures and public policies.

This definition stresses the importance of change in the logic of political behavior. The latter changes through a process(2) leading to the institutionalization in the domestic political system (at the national and-or subnational levels) of discourses, cognitive maps, normative frameworks and styles coming from the EU. Another point: the definition does not mention organizations. By contrast, it accommodates both organizations and individuals. It is sufficiently broad to cover the major interests of political scientists, such as political structure, public policy, identities and the cognitive dimension of politics. It can be applied both to EU member states and to other countries.

B. Concept stretching

Turning to conceptual stretching, according to elementary logic a concept can be described by two fundamental properties, that is, extension and intension (Sartori 1970). Intension refers to the collection of properties covered by a concept. Extension represents the class of entities to which the concept applies. There is a trade-off between intension and extension. The more properties are included in the concept of Europeanization, the smaller will be the class of empirical instances. Although there are no priorities in the choice between intension and extension, a concept with high intension has high discriminatory power. Most studies of Europeanization, however, seem to privilege extension. This is probably the result of an early stage of research, when the analytic grid has to be broad enough as to accommodate a wide range of empirical observations that may have something to do with Europeanization. Thus, Europeanization is supposed to explain processes of cultural change, new identities formation, policy change, administrative innovation, and even modernization. It covers the formation of European public policy and the effects of EU decisions on national systems. It affects member states but also the wider world, as there is undoubtedly Europeanization of policy in countries applying for EU membership.

However, the more we know about Europeanization, the more exigent we should be in terms of intension, that is, the properties of the concept. Otherwise the risk of degreeism may loom large. Degreeism, as defined by Sartori (1970, 1991), occurs when differences in kind are replaced by differences of degrees. As we are not able to see the difference between a cat and a dog, we speak of different degrees of cat-dogs (Sartori 1991). The metaphor of Europeanization as a continuum and the notion of domestic political systems being ‘increasingly’ penetrated by EU policy make the distinction between the cat and the dog difficult. The point is that without boundaries it is impossible to define Europeanization. But the literature is somewhat reluctant to tell us what falls outside Europeanization. If everything is Europeanized to a certain degree, what is not Europeanized? In this case again, political scientists in the early days of research on a new topic have been hesitant to exclude possible indirect, unforeseen, and simply odd instances of Europeanization. They tend to argue that ‘a certain degree of Europeanization’ may be found almost everywhere. In terms of sociology of knowledge, one can understand this cautious approach. Political scientists do not want to preclude innovation by posing a rigid fence around a developing area of research. But this strategy has the cost of conceptual sloppiness and degreeism.

Shall one then stick to a very narrow definition of Europeanization, thus limiting the scope of analysis? Connotative precision (that is, high intension) is vital in this stage of research, but one does not need to narrow down the analysis to a few selective aspects of Europeanization. The best strategy – I argue – is to unpack the concept and to distinguish between Europeanization and other terms (thus showing what Europeanization is not). To unpack a concept – Sartori explained – is to decompose ‘mental compounds into orderly and manageable sets of component units’ (Sartori 1970:1083). In a recent paper, Morlino (1999) introduces a fundamental distinction between the Europeanization of the polity and the Europeanization of public policy. He then combines this dimension with the level at which Europeanization is observed, that is, the EU or its member states. By doing so, he rules out the hypothesis of Europeanization processes taking place outside the Union – a point that can be questioned. Additionally, one does not see the need for the second dimension, that is, the one discriminating between the EU and member states: what is the ‘Europeanization of the EU’ supposed to mean? Be that as it may, taxonomic exercises like this are indispensable if one does not wish to proceed in the (conceptual) dark.

C. Europeanization and contiguous (but different) terms

In the next section I will present my own ‘unpacking’ proposals. At the onset, however, it is indispensable to draw the line between Europeanization and other concepts, namely convergence, harmonization and integration.

To begin with, Europeanization is not convergence. The latter can be a consequence of Europeanization. Convergence is not Europeanization because there is a difference between a process and its consequences. However, Europeanization can also produce divergence. Policy studies have detected considerable variability. Some authors have found consistent convergence of media ownership policy induced by Europeanization (Harcourt 2000), but the Europeanization of transport policy has resulted in striking differences between France, Germany, Italy, the Netherlands and the UK (Héritier and Knill 2000).

Europeanization should not be confused with harmonization. In a study of French environmental policy for the agricultural sector, Petit concludes that ‘if Europeanization can increase problem-solving capacity, member state institutions continue to produce divergent solutions. Therefore, Europeanization does not necessarily accord with harmonization. And in the absence of harmonization, competitive distortions in the common market are likely to arise (Petit forthcoming: 23). Harmonization produces a level playing field, whereas Europeanization can end up in regulatory competition and even distortions of competition.

Finally, Europeanization is not political integration. Typically, theories of integration address the question ‘why do different countries join forces and build up supranational institutions’? Europeanization would not exist without European integration. But the latter concept belongs to the ontological stage of research, that is, the understanding of a process in which countries pool sovereignty, whereas the former is post-ontological, being concerned with what happens once EU institutions are in place and produce their effects (on EU ‘ontology’ and post-ontological perspectives see Caporaso 1996). This begs the question of the relationship between Europeanization and general theories of integration. As shown by Börzel (1999:576-577), theories of integration focus on the issue whether European integration strengthens the state (inter-governmentalism), weakens it, or triggers ‘multi-level governance’ dynamics. The post-ontological focus of Europeanization brings us to other, more specific, questions, such as the role of domestic institutions in the process of adaptation to Europe. The final results in terms of ‘strengthening’ or ‘hollowing out’ of the state are always conditional (for example, it depends on the configuration and response of domestic institutions) in the literature on Europeanization.

Europeanization and EU policy formation should be kept distinct at the conceptual level. But in the real world they are interconnected. European policy is not a mysterious deus ex machina situated ‘up there’. Instead, it originates from processes of bargaining, imitation, and diffusion wherein domestic governments and national interest groups play an important role, together with European institutions. Further, the outcomes of Europeanization can feedback into the process of EU policy re-formulation. National actors can draw lessons from Europeanization and seek to change or adapt EU policy. However, analytically one should distinguish between the process leading to the formation of a certain policy, and the reverberation of that policy in the national arenas. More generally, I do not see why one cannot treat Europeanization as an example of ‘second-image reversed’ process (Gourevitch 1978). If ‘second-image reversed’ research designs are legitimate tools of scientific inquiry, then Europeanization is nothing but an instance of these designs.

2. What is Europeanized and to what extent?

The aim in this section is to suggest a taxonomy that can assist political scientists in their attempt to ‘unpack’ the concept of Europeanization. The emphasis on political scientists is appropriate as Europeanization can be approached differently by other disciplines. A taxonomy is a simple device that organizes research and makes complex concepts amenable to empirical analysis. Therefore the ideas presented in this section serve the purpose of assisting the research design. To repeat: the emphasis at this stage is on the organization of research, not on explanation.

Bearing in mind the definition of Europeanization adopted in the previous section, one can approach the study of this phenomenon by raising the questions ‘what is Europeanized?’ (that is, the domains where the effects of Europeanization are supposed to materialize) and ‘to what extent?’ (that is, extension and direction of Europeanization). Figure 1 – inspired by but different from Morlino’s proposal – provides a suggested taxonomy for the empirical investigation of these two dimensions.

Let us begin with the domains of Europeanization (the first column in fig.1). Drawing upon Morlino (1999), the first important distinction is between macro-domestic structures and public policy. Public policy is not the mere output of the political system, especially in the EU, and consequently it should be distinguished from it. Public policy has dynamic effects on political structures. But analytically, and specifically in a static exercise on taxonomies, one is allowed to differentiate between policy and domestic structures. I will now turn to a review of recent research conducted on the domains indicated in fig.1, with the aim of presenting some of the emerging themes.

A. Domestic political structures

a) Institutions

Domestic political structures include institutions, public administration, inter-governmental relations and the legal structure. In terms of institutions, it has been observed that Europeanization has somewhat limited the role of parliaments and traditional territorial representation (Andersen and Burns 1996: 230). This chimes with the traditional analysis of the democratic deficit of the EU (however, confront Wincott 1998 and Majone 1996 for different perspectives). The supposedly technocratic features of the EU political system and especially the Economic and Monetary Union have instead widened the scope for executive leadership in the policy process and the role of technocrats (Featherstone 1998; Dyson and Featherstone 1999). However, these statements should be tempered by the observation that the manifestation of post-parliamentary governance represents broader political dynamics. The crisis and, perhaps, redefinition of parliaments is caused by long-term transformations of democracies, such as the increasing importance of non-majoritarian institutions and the ‘technicization’ of public policy-making (Majone 1996; Andersen and Burns 1996). Additionally, although the EU has brought about a verticalization of the relations between cabinets and assemblies, this is not a uniform pattern. Executive leadership in public finance has increased in some countries, but in other countries the relation between cabinets and parliaments was already biased towards the executive. A typical manifestation of Europeanization is that it is contingent on the type of domestic structures on which it impinges on, as will be argued in sections 3 and 4.

b) Public administration

Although some political scientists, most notably Wolfgang Wessels, have argued that a ‘fusion’ of national and EU administrations has taken place (see for example Rometsch and Wessels 1996), administrative diversity remains the norm. In addition, there are puzzles and paradoxes in search of explanation. Conventional wisdom argues that EU administration has developed around French and, in certain areas, German models. Thus, one would expect a clash of administrative traditions and cultures between the EU and the UK. Yet ‘the British machinery of government’ – to paraphrase Bulmer and Burch (1998) – has been able to adjust to Europe by maintaining British traditions.

Page and Wouters (1995) have examined both direct and indirect instances of Europeanization of public administration in member states. Direct influences are those related to the European Court of Justice’s jurisprudence, for example in terms of conditions of pay and service. But indirect influences – Page and Wouters note – may be even more powerful. For example, national bureaucrats may demand change in the structure of public administration in line with European standards. There can be a contagion of administrative structures from the EU to the national level. Brussels may provide a transfer platform both for EU administrative structures and for national best practice that are diffused and imitated via the daily exposure to the same administrative networks. Coombes (cited by Page and Wouters 1995:188) had already noted in 1968 that ‘increasing contact with other European civil servants … could produce what might be called a feedback into the British civil service from the procedures and customs of other European countries’ (Coombes 1968:61).

Although one might expect a process of Europeanization of public administration, the evidence collected so far points towards a modest effect. Page and Wouters (1995:202) conclude that ‘there is no clear EC model, distinctive and relevant to the national bureaucracies of nation states, that is likely to find its way through contagion, emulation, the demonstration effect or the natural process of adaptation to an important source of political power’. This conclusion, however, should be placed in the context of the precise – yet limited – research question raised by Page and Wouters, that is, whether the structure of domestic administrations has converged towards an EU model.

But, as argued above, Europeanization is different from convergence and harmonization. Although the EU may not be a ‘centre of gravity’ for administrative structures(3), Brussels may nonetheless provide a transfer platform through which administrative innovation travels from one country to others. For example, since 1998 there have been informal meetings of EU ministers and officials on the diffusion of administrative reform and the quality of legislation. Overall, although (a) direct contact between administrations and (b) the Organization for Economic Cooperation and Development’s Public Affairs Division (PUMA) provide the most relevant opportunities for administrative transfer, future research on policy transfer may suggest avenues for Europeanization different from the ones hypothesized by Page and Wouters. Further, the participation in EU-managed administrative networks may change the belief systems of domestic civil servants. Kohler-Koch (1996) argues that the deepest impact of EU policy is cultural. Specifically, EU regional policy, for example, has transformed the notion of what governance is all about. Propositions like this can be tested by examining the difference between the belief systems of domestic civil servants involved in the administrations of EU funds at the regional level in comparison with the beliefs of civil servants engaged in regional policy without a EU dimension.

Concluding on this point, the general impression is that there are doubts as to whether the EU is a major independent variable changing administrations across Europe. A recent survey of the effect of EU dynamics on central ministerial administration concludes that the literature is inconclusive and that the EU can be at best an intervening variable in administrative change (Goetz 2000). Goetz observes that, instead of assuming that Europeanization is a potential source of change and looking at its impact dichotomously, political scientists should account for its relative importance alongside other variables. The emphasis on absolute impacts rather than relative impacts is a sore point in the current literature on Europeanization, and we will turn to this issue in the conclusions.

c) Inter-governmental relations

In two different studies, Börzel (1999) and Goetz (1995) examine the impact of Europeanization on inter-governmental relations. Goetz finds out that the process of Europeanization has not transformed the German federal system. His conclusion rules out a major impact:

‘Our findings caution against a perspective which equates Europeanization with the systemic transformation of national systems of governance. Clearly, German inter-governmental relations have become Europeanized; but this Europeanization has been achieved without sacrificing, or even seriously calling into question, central features of national governance. Instead, those features have been affirmed, if not reinforced’ (Goetz 1995:112)

In her analysis of inter-governmental relations in Spain and Germany, Börzel (1999) argues that institutions refract the impact of Europe. Institutions determine the distribution of resources to the actors affected by Europeanization. In addition, the prevalent policy style(4) channels the response of regions towards a confrontational or cooperative strategy (vis-à-vis the central government). The German cooperative culture has produced successful adaptation, whereas Spanish regions have not gone too far with their confrontational culture.

There is a lively debate on the impact of Europeanization. To illustrate: Börzel (1999) argues that EU policy allocates more implementation costs to regions, and at the same time it increases the power of central governments. Therefore, EU policy produces a distribution of ‘say and pay’ that impacts negatively on regions. Morlino (1999), by contrast, hypothesizes that regions are empowered by Europeanization. Smyrl (1997) suggests a completely different perspective when he demonstrates that regions in the same country can be empowered or disempowered depending on the constellation of variables at the policy level, such as the structure of policy networks.

d) The legal structure

This area deserves a comprehensive treatment, but here a few comments may suffice. The transformation and Europeanization of domestic legal structures has been the object of several studies (Shaw and More 1995; Weiler 1994). Stone Sweet and Brunell (1998:76) argue, in line with a consistent body of literature, that ‘the effect of the European Court of Justice’s case law on national and supranational policy processes and outcomes in the social domain has been deep and pervasive’. The focus of these studies goes well beyond the role of the European Court of Justice in the EU policy process. Indeed, these studies offer a new theoretical perspective on the construction and dynamics of supranational governance in Europe (Stone Sweet and Brunell 1998; see also Stone Sweet 1999 on the theory of judicialization).

e) Structures of representation

Research on the European dimension of party systems and pressure groups has progressed steadily in the last decade or so, although specific analyses of Europeanization are a rather recent feature of the academic debate. Mair (1999) provides an example of rigorous analysis of the Europeanization of party systems by using the classic Sartorian dimensions of format and mechanics (Sartori 1976). His argument is that ‘there is little evidence of any direct impact on these features [that is, format and mechanics] of the party systems to be discerned. Indeed, I suggest that of the many areas of domestic politics which may have experienced an impact from Europe, it is party systems in particular that have perhaps proved to be most impervious to change’ (Mair 1999:4). There is however an indirect effect of European integration on party systems – Mair qualifies. Policy issues are still debated in the national arena, but are mostly decided at the European level. By contesting national elections on policy issues that are decided elsewhere, ‘voters are being offered a voice which is likely to have little or no effect on the practice of decision-making’ (Mair 1999:25). Europeanization thus contributes to the process of de-politicization, indifference, and popular disengagement. However, Mair does not consider the option of re-politicization at the EU level. Comparative analyses of EU policies, indeed, have revealed increased politicization (Radaelli 1999).

In contrast with Mair’s findings of limited Europeanization, the literature on pressure groups has detected a transformation of the structure and strategy of business groups and common cause groups. This phenomenon goes well beyond the formation and consolidation of European-level associations and confederations and involves the Europeanization of domestic groups (Mazey and Richardson 1993; Green Cowles in Green Cowles et al. 2000).

f) Cognitive and normative structures

Not only can Europe affect formal structure, it can also influence the values, norms and discourses prevalent in member states. In turn, cognitive transformation may change the preferences of policy-makers and therefore feedback into the process of European integration. Jachtenfuchs et al. (1999:411) observe that, although institutions reflect the interests of the actors that create and consolidate them, they are also ‘embedded in particular normative orders or structures of meaning’. Individual beliefs on the legitimacy of political order are (re)produced via structures of meaning. The analysis of cognitive and normative structure is therefore connected to the renewed interest in sociological institutionalism, preference formation and political legitimacy.

A recurrent theme in this type of analysis is diversity (as opposed to covergence). Recent projects have provided evidence of

distinct visions of what the EU polity should be (Jachtenfuchs et al. 1999),
differences in the construction of collective elites identities that give political meaning to innovations such as the Economic and Monetary Union (Risse et al. 1999),
limitations in terms of Europeanization of citizenship (Checkel 2000), and
diverse national discourses providing legitimacy to Europe (Schmidt 1999).

In terms of (d) it is useful to distinguish the general discourse on Europe (for example, the French political discourse on what is legitimate European integration, see Cole and Drake 2000) and the specific discourse on public policies (to illustrate: the discourse on legitimate tax competition, ‘social dumping’, or sustainable development). Schmidt, in her analysis of the general discourse on Europe, argues that the process of adaptation to Europe has three main dimensions: economic, institutional and discursive (Schmidt 1997). The type of discourse used to justify economic and institutional adaptation is crucial in terms of legitimacy.

France – Schmidt illustrates – has not been able to forge a discourse justifying the process of economic liberalization triggered by EU decisions. The adaptation to European policy has been justified in terms of protection from globalization and Americanization, but this makes most EU liberalizing choices (for example in the area of public utilities and culture) difficult to accept. She argues, perhaps with a hint of exaggeration, that ‘without such a legitimating discourse, the loss of socio-economic capacity in the face of European and global forces has led to public malaise and protest, stymied necessary welfare efforts, and contributed to the success of the extreme right’ (Schmidt 1999:17)(5).

According to Giuliani (1999) and Featherstone (1998), in Italy and Greece the discourse on Europe has become equivalent to modernization and normalization – in the sense of making these countries more normal, that is, more similar to the other EU partners. This explains the rush to Economic and Monetary Union of the Italians (and more recently the Greeks). To qualify for the single currency was inter alia a way of ‘loosing eccentric, extraordinary, astonishing and unique features which has certainly fascinated quite a few Italianists but probably not many Italians’ (Giuliani 1999:7).

Denmark – a country where the opposition to the single currency of ordinary people contrasts with the pro-Euro orientations of the elites – provides an example of political manipulation of the discourse(6). In the run-up to the September 2000 referendum on the Euro, the Danish social-democratic elite insisted on the link between the single currency and employment policies – the goal being one of making the Euro more ‘social-democratic’ in public percpeptions:

‘Two social democratic spokes persons, Torben Lund and Jacob Buksti, argue that they well understand that the Danes voted No to EMU in 1992. Then, the EMU was a markedly liberal project, which had low inflation rather than high employment as the prime objective. Today, the situation is different, they argue, because of the Luxembourg, Cardiff and Cologne processes’ (Marcussen 2000:20).

B. Public Policy

The case of Economic and Monetary Union has drawn attention to the problems of legitimacy. In turn, this has stimulated the analysis of the specific policy discourses that seek to provide legitimacy. This recent emphasis on discourse is a component of a broader research programme on the cognitive structure of public policy, which includes the Europeanization of policy frames (see Dudley and Richardson 1999 on steel policy and Radaelli 1997 on tax policy), policy narratives (Radaelli 2000a) and paradigms of public policy (defined by Hall 1993). But the Europeanization of public policy can take different forms. In principle, it can affect all the elements of public policy, such as actors, resources, and policy instruments (see fig.1). Additionally, Europeanization can affect the policy style, for example by making it more or less conflictual, corporatist or pluralist, or more or less regulative.

It is impossible to review the considerable amount of research on the Europeanization of public policy. The essential point, however, is simple. When contrasted with the literature on domestic structures, identities, culture and values, studies at the policy level reveal a greater impact of Europe, and in many instances convergence (Harcourt 1999 on media ownership policy; Schneider 1998 on telecommunications), direct and indirect transfer of models from Brussels (Radaelli 2000b on monetary policy, tax policy and media policy), and a profound impact of EU regulation on national competition policy and regulatory approaches (Majone 1996).

C. The analysis of change

Ideally, research on Europeanization should be organized in a matrix similar to the one portrayed in fig.1. On the one hand, one has to specify ‘what’ is Europeanized (that is, domestic structures, cognitive structures, or public policy); on the other, there is the question of ‘how much change’ has been brought about by Europeanization. Drawing upon Héritier (1998), Héritier and Knill (2000), Börzel (1999), Green Cowles et al. (2000), four possible outcomes can be discerned (see fig.1, second column): inertia, absorption, transformation and retrenchment. Taken together, they cover both the magnitude of change and its direction (retrenchment being an example of ‘negative’ Europeanization).

Inertia is a situation of lack of change(7). This may simply happen when a country finds that EU political architectures, choices, models or policy are too dissimilar to domestic practice. Inertia may take the forms of lags, delays in the transposition of directives, implementation as transformation, and sheer resistance to EU-induced change. In the long-term, however, inertia can become impossible to sustain (economically and politically). Therefore, one can submit that long periods of inertia should produce crisis and abrupt change (Olsen 1995).

Absorption indicates change as adaptation. Domestic structures and policy legacy provide a mixture of resiliency and flexibility. They can absorb certain non-fundamental changes, but maintain their ‘core’. For example, Fabbrini (1998) argues that the Italian institutions have flexed like bamboo in the wind to accommodate the pressure to qualify for the single currency, but since this goal has been secured, they have returned to the original position. Absorption – as specified by Héritier – is accommodation of policy requirements(8) without real modification of the essential structures and changes in the ‘logic’ of political behavior (Héritier 1998:21).

The ‘accommodation’ of Europe should not be confused with transformation. This is similar to what Hall labels ‘third order’ change, that is, paradigmatic change (Hall 1993). Paradigmatic change occurs when the fundamental logic of political behavior changes – for example, a change in the format and mechanics of party systems, or the adoption of a new orthodoxy in monetary policy.

However, Europeanization can also induce retrenchment. This is a very paradoxical effect, as it implies that national policy becomes less ‘European’ than it was. Héritier and Knill (2000) show that – in the Italian case at least – EU pressure to liberalize road haulage has objectively strengthened coalitions of domestic actors opposing reform. Therefore the direction of change has been one of increased intervention, rather than liberalization. How this may happen depends on the mechanisms and the factors explaining Europeanization, two issues to which we will now turn.

To sum up the results of this section: research can be organized by using two dimensions, that is, the ‘objects’ of Europeanization (domestic structures or public policy) and the dimension (and direction) of change. A typical research design should therefore treat fig.1 as a matrix and fill in the cells with empirical observations. Research conducted so far shows a higher level of Europeanization of policy, whereas structures seem to be less permeable. This brings us to the question whether policy dynamics feedback into political structures, an issue that can be dealt with by looking at the mechanisms of Europeanization and the explanatory variables. It is to these questions that we now turn.

3. The mechanisms of change

Once the concept of Europeanization and the dimensions of change have been clarified, the next step concerns the possible mechanisms of Europeanization. I will leave to the next section the identification of the key variables in the process of Europeanization. One important restriction in this and the next section is that they are concerned exclusively with public policy, and therefore do not cover other domains on Europeanization.

The literature on the mechanisms of Europeanization is still in its early days. Börzel (1999) and the contributors to Green Cowles et al. (2000) have drawn attention to the so-called ‘goodness of fit’ (in plain English, the degree of institutional compatibility) between domestic institutions and European policy. By focusing on the ‘goodness of fit’, these authors draw our attention to explanatory factors related to any mechanism of change(9). Therefore, their insights on explanation will be dealt with in the next section. Knill and Lehmkuhl (1999) have presented three mechanisms (the first based on the presence of European models, the second on the domestic opportunity structure and the third on the role of ‘minimalist’ directives in ‘framing’ integration). In a paper on policy transfer, Radaelli (2000b) draws on institutionalism in organizational analysis and presents the mechanisms of coercion, mimetism and normative pressures in EU policy diffusion. In another paper, he looks at the process of EU-induced cognitive convergence in the absence of direct compulsion from Brussels (Radaelli 1997). Finally, Kohler-Koch (1996) highlights subtle – yet crucial – mechanisms that go beyond the issue of the impact of EU policy on the ‘balance of power’. Figure 2 combines the insights of the current debate. The purpose is to organize our understanding of Europeanization, bearing in mind that at this stage it is vital to raise the right questions, instead of looking for the ultimate answer.

Consider first the role of European models (fig.2). In certain policy areas the EU prescribes the adoption of a specific model. As Knill and Lehmkuhl (1999:3) argue, ‘member states have to bring domestic arrangements into line with a ‘European model’ which is implicit in the supra-national policy decision’. Examples are provided by new regulatory policies in the areas of consumer protection, environmental policy, and health and safety at work. In this cases the EU positively prescribes the adoption of a model, and therefore one could use the term ‘positive integration’ to distinguish this mechanism from the cases in which the EU strikes down national barriers to the emergence of European markets without prescribing models. In addition to the new ‘market-shaping’ regulatory policies mentioned by Knill and Lehmkuhl, one should consider the case of the Economic and Monetary Union, in which a fully-fledged institutional model of monetary policy is being diffused to the countries of the Euro-zone.

When there are EU models, member states are under ‘adaptational pressure’. Put differently, they are under pressure to adapt to Europe. Pressure implies coercion – for example, certain directives specify a period of time at the end of which member states are compelled to introduce regulatory arrangements. But some European policies are adopted only by a limited number of member states. Flexible integration is an option available in a number of policy areas. This implies that adaptational pressure can operate with mechanisms different from coercion when countries have the option of joining EU policy or remaining outside. Mimetism illustrates this alternative channel of Europeanization. If the countries adopting EU models provide a critical mass, the remaining countries can feel the force of attraction of the EU ‘center of gravity’ and join in. This can explain the dynamics between the countries of the Euro-zone and Greece at the moment. Mimetism and coercion are mechanisms of isomorphism, that is, the tendency to become alike, well known to the new institutionalism in organizational analysis (DiMaggio and Powell 1991).

There is a qualification, however. Mimetism cannot work in areas where the incentives to opt out increase as the number of countries opting for EU policy grows. For example, if n-1 countries decide to abolish state aids or harmful tax measures, the remaining n country will enjoy the benefits of being the only tax haven or the sole area of state aid in the Union (Genschel and Plümper 1997). This introduces the difference between areas wherein the nature of strategic interaction corresponds to coordination games (to illustrate: setting standards in telecommunications) and areas where the logic of cooperation games prevails (harmful tax competition being an example). In a cooperation game such as the prisoner’s dilemma the logic of the ‘centre of gravity’ does not apply. Parenthetically, this explains why the European Commission has so far rejected the possibility of flexible integration in the area of tax policy coordination.

However, there are cases in which EU policies do not prescribe a model (see fig.2). Policies of ‘negative integration’ (or ‘market-making’ policies, as opposed to ‘market-shaping’) create integrated markets by removing barriers to trade, investment, freedom of establishment and free circulation of people. They do not say how a market should be governed in terms of institutional models, but typically emphasize the role of mutual recognition once the barriers have been removed. As Knill and Lehmkuhl (1999) argue, in these cases the ‘goodness of fit’ explanation does not perform well. The key mechanism triggered by ‘negative integration’ via mutual recognition is regulatory competition (Majone 1996; Sun and Pelkmans 1995). Knill and Lehmkuhl exemplify:

‘The abolition of trade barriers for beer, for instance, has no direct impact on how the production of beer is regulated at the national level. It only implies that beer can be freely sold in other member states, hence altering strategic opportunities and constraints for domestic producers and consumers’ (Knill and Lehmkuhl 1999:3).

The impact of this type of policy is not direct (there is no EU institutional framework or model to be diffused), but indirect, via international regulatory competition. When markets are opened, existing domestic equilibria are challenged, but there is no prescription of ‘how the new equilibria must look’ (Knill and Lehmkuhl 1999:10). The essential mechanism concerns the changing distribution of resources (and ultimately power) between domestic actors rather than the compatibility between EU and domestic and EU policy. Fig.2, accordingly, puts emphasis on the domestic opportunity structures. The overall effect of this indirect mechanism cannot be predicted a priori. It all hinges on who and how is empowered and disempowered by ‘negative integration’ at the national level. European liberalization can tip the scale in favor of pro-reform coalitions, or can end up in retrenchment.

Directly or indirectly, the mechanisms examined so far assume that EU policy has a precise direction and aims to produce specific compliance at the level of the member states. Typically, they are based on the ‘hard’ instruments of EU public policy, such as directives and decisions of the European Court of Justice. However, there are at least three other ‘soft’ ‘framing’ mechanisms of Europeanization, portrayed in the right-hand side of fig.2. To begin with, in some cases, such as railway policy, the EU proceeds by minimalist directives or non-compulsory regulations. By their nature, these instruments do not create any pressure in terms of adaptation (or ‘goodness of fit’) or international regulatory competition. Yet they can prepare the ground for major policy change. They do so by providing additional legitimacy to domestic reformers in search for justifications, by ‘inseminating’ possible solutions in the national debate, and by altering the expectations about the future. Additional legitimacy is particularly important when domestic leaders are engaged in radical reforms (Knill and Lehmkuhl 1999:12-13). The introduction of new solutions coming from Brussels can alter the perception of problems. New solutions can provide a new dimension to national policy problems and trigger learning dynamics or a different political logic. In the case of media ownership policy, the instrument of audience share was firstly aired at the EU level and then ‘inseminated’ in national political systems with important effects on national legislation in Germany and the UK (Harcourt 2000). Further, even the vaguest European policy has the potential of altering the expectations of domestic players, for example by showing that the opponents of liberalization are fighting for a ‘lost cause’ because EU policy is heading towards a totally different direction (Knill and Lehmkuhl 1999:13).

Can the EU affect national policy even in the absence of EU directives and regulations, albeit ‘minimalist’? The answer hinges on the strength of cognitive convergence. There are policy areas, such as direct taxation, where the EU has been severely constrained in terms of being able to produce a coherent corpus of directives. The European Commission and the Council have nevertheless been active in proposing high-level groups and qualified policy fora. In 1997-1999, high-level tax policy makers (who in the past avoided any reciprocal discussion of national corporate tax measures) were involved in a Council group dedicated to a systematic peer review of harmful tax practices. The group was based on a non-binding instrument, a code of conduct on harmful tax competition, nothing more than a gentlemen’s agreement amongst the Finance Ministers of the EU. So far no major directive has materialized, but the group has provided a mechanism of socialization and a vehicle for the transmission of a policy discourse based on harmful tax competition. One of the effects of socialization is that some domestic harmful tax measures which were in the pipeline have been shelved as incompatible with the emerging European paradigm of what ‘good’ tax policy should be. Policy fora and socialization processes are therefore potential channels of cognitive convergence and ultimately Europeanization. There is also evidence of British tax policy-makers bringing national corporate tax policy in line with the European paradigm although no direct compulsion from Brussels was operating (Radaelli 1997). This is a subtle yet powerful mechanism of Europeanization. It materializes when domestic policy-makers think in terms of standards of ‘good policy’ defined in Brussels.

Finally, Europeanization can produce effects that go beyond the balance of power. Kohler-Koch (1996; 1999) argues that the insistence on the implications of EU policy in terms of the balance of power can be misleading. The crucial effect of Europeanization – she suggests – is on the dissemination of the ‘network mode of governance’ (as opposed to other modes of governance, such as corporatism, statism, and pluralism) into the member states (Kohler-Koch 1999). EU policy can change the understanding and the practice of what legitimate governance is all about. This is an effect that cannot be detected at the level of a single episode; neither can it be investigated in a simple ‘balance of power’ research design, for example by looking at the battle between reformers and advocates of the status quo. Yet the long-term implications of new modes of governance can be the most powerful.

4. Towards explanation? The intervening variables

Mechanisms shed light on the process of Europeanization. But how does one explain the likelihood and direction of change? Two research projects have been recently concluded (Green Cowles et al. 2000; Héritier and Knill 2000), but more research is needed in this area. At the moment of writing, the best option is, arguably, to examine explanatory variables with the aid of the evidence made available by recent projects, but also with insights and suggestions that, at this stage, are still somewhat speculative. On balance, this section contains more ideas on how to proceed than specific hypotheses corroborated by a significant amount of cumulative empirical evidence.

The explanation provided by Börzel (1999) and Green Cowles et al. (2000) is based on the general idea of adaptational pressure. The latter – I would submit – works at its best when there is a European model, but, at least conceptually, the focus on adaptational pressure is independent from any of the mechanisms reviewed in the previous section. The relationship between adaptational pressure and change in domestic structures and policies is curvilinear. When adaptational pressure is low, because the content of EU policy is already present in a member state, there is no need to change domestic institutions. Simply put, there is a good ‘fit’ between national policy and the EU. Hence it is easy to absorb ‘Europe’. At the other extreme, when the distance between EU policies and national ones is very high, member states will find it very difficult to ‘digest’ and ‘metabolize’ European policy. Hence there will be inertia at the domestic level. The degree of change will be high when adaptational pressure falls between the two extremes. An arc on a Cartesian plan, with adaptational pressure on the X-axis and domestic change on the Y-axis can illustrate this relation.

Börzel and Green Cowles et al. explicitly refer to new institutional analysis in their investigation of adaptational pressure. Domestic institutions refract Europeanization by providing the dominant strategy in cases of ‘misfit’ between the EU and the member states. Börzel (1999) shows how Spanish regions reacted with a confrontational strategy to Europeanization, whereas German Länder preferred a co-operative strategy. Confront and cooperation are embedded in different ‘institutional cultures’. Institutions, in addition, determine the distribution of resources among domestic actors affected by Europeanization. The result is that the impact of Europeanization is contingent on institutional factors. A corollary is that Europeanization will produce diversity rather than convergence because domestic institutions differ widely.

The ‘goodness of fit’ argument is not without its problems, however. To begin with, what happens when domestic institutions are fragile (Morlino 1999)? In countries such as Belgium and Italy, domestic institutions have been in crisis or in transition in the last decade or so. They have not behaved like rigid posts, capable of fencing or shaping the process of Europeanization. Quite the opposite, in certain episodes Europeanization has been a crucial component of domestic institutional change, as shown by the impact of the Economic and Monetary Union. The interaction between Europeanization and domestic institutions is therefore dialectic.

Secondly, the notion of ‘goodness of fit’ needs further exploration. The metaphor of the ‘fit’ covers quite a broad range of elements. To illustrate: a country can have a bad or good ‘fit’ because of the presence-absence of Roman law, strength-weakness of bureaucratic structures, corporatist-pluralist style of decision-making, centralization-decentralization of power, and so on. The analysis of the ‘goodness of fit’ and adaptational pressures should become more precise by discriminating between the elements that can accommodate European pressure and the ones that are ultimately resilient to Europe. One way of describing the power of a phenomenon is to pin down as precisely as possible what it can break down and what can resist it(10).

To conclude on this point, the ideas of adaptational pressure and ‘goodness of fit’ provide a point of departure for general explanations, and future research will probably highlight more specification to this type of analysis, as shown amongst others by Knill and Lenschow’s chapter in Green Cowles et al. (2000).

Héritier and her associates are bringing the debate further in the direction specific sets of explanatory variables. Their first group of variables refers to the institutional capacity to produce change. This is still a very general hypothesis. Accordingly, it is useful to segment it into more specific propositions (see figure 3). To begin with, the presence of veto players à la Tsebelis (1995) constrains the institutional capability to produce change. The conventional analysis of the political systems can be used to show how political processes differ in the EU member states. This is a general, macro-focus, one centered on the characteristics of the political systems. But Héritier and her associates put emphasis on informal veto players as well. For example, in policies of privatization and liberalization of the utilities pressure groups can represent serious obstacles. Even small haulage companies can exercise their blackmail potential on certain political systems (Héritier and Knill 2000). By doing so, Héritier and her colleagues switch from the macro-characteristics of the political process to the specific features of public policies.

The interaction between policy dynamics and the macro-political structure is perhaps one of the most interesting areas of Europeanization. But one cannot see it if the policy level and the macro-political level are lumped together. Accordingly, one should separate the policy level from the macro-comparative analysis of the political systems and treat the ‘informal’ veto players within the wider framework of actors operating in policy systems (see below). What is the advantage of this choice? Empirical research shows that Europeanization has changed political systems independently from their macro-characteristics. The Economic and Monetary Union has transformed the state in countries as different as France and Italy (Dyson and Featherstone 1999). Analytic separation does not mean independence, however. Knill and Lenschow (2000) note that policy structures can be embedded in ‘regime cores’ – a point which invites more research on the connections between policy and politics. But – to reiterate – in order to detect the dynamic effect of policy upon political structures and the role of embeddedness one has to keep the two levels analytically distinct.

After this short digression on policy and politics, let us complete the analysis of macro-variables. The scope and type of executive leadership has to be considered. The leadership – Héritier and Knill (2000) explain – can be integrated or, at the other extreme, fragmented, short-lived and conflict-ridden. When leadership is integrated and the number of veto players is low, Europeanization hardly makes a difference. Executives willing to promote policy change can do so whether European policy exists or not. At the other extreme, fragmented leadership with strong sectorial veto players makes EU-induced change improbable. As shown by road haulage policy in Italy, Europeanization can produce retrenchment. Europeanization is instead most likely to have a high impact (in terms of policy change) under conditions of intermediate institutional capacity.

At any rate, the institutional capacity to produce change is a necessary condition, but it is not sufficient. The presence or absence of change – and its direction – depends on more specific variables at the level of the policy structure. Before we examine them, however, it is useful to draw attention to timing (fig.3). The impact of EU public policy is contingent on whether a country is already involved in a process of reform or not. For example, certain policies of liberalization in the EU have caught some countries unprepared whereas others, most notably the UK, were already on their way to de-regulation and privatization. Put differently, the analysis of the effects of European public policy on national policy systems should be conducted in parallel to the investigation of endogenous processes. The adaptational pressure of EU policy is certainly higher in the case of a country which has already undertaken reforms consistent with EU trajectories.

At a more general level, the temporal dimension is relevant to the extent that decision-makers can manipulate ‘time’ by delaying decisions, sequencing the process of adaptation, and control the rate of speed of Europeanization (for example they can follow a gradualist path or proceed by leaps and bounds). This brings Goetz (2000) to borrow from Schmitter and Santiso (1998) the categories of time, timing and tempo(11).

The final set of variables – labeled ‘policy structure and advocacy coalitions’ – have not received enough attention in the projects completed so far. Consequently, the analysis becomes rather speculative from this point onwards. Yet the policy level is crucial because the most surprising effects(12) of Europeanization have taken place via policy change even when formal political structures have remained unchanged (Dyson and Featherstone 1999; Featherstone 1998; Ferrera and Gualmini 1999; Giuliani 1999).

There are four observations on ‘policy variables’ (see fig.3). The first observation concerns the difference between policies that can be governed by technocratic, elitist circles and policies that by their very nature require a wider constellation of actors. Let us consider the examples of Economic and Monetary Union and transport policy. In the case of the single currency, the process of Europeanization has been seized by small technocratic elites. Core executives, central banks and technocrats with political power have been empowered by Europeanization. They have been able to produce dramatic change in monetary policy even in countries with low potential for modernization (such as Greece and Italy). By contrast, the liberalization of transport policy has become the hostage of small pressure groups with intense preferences, and divisions within the political establishment. One key difference between the two policy areas is – I submit – the diverse degree of technocratic capture potential. Comparatively, monetary policy can be captured by a small policy elite rather easily. To govern interest rates and the domestic financial markets, one needs clear rules about the competence of Treasury and the central bank, and the manipulation of very few policy instruments (either a monetary aggregate or the structure of interest rates). Contrast this with road haulage, where different departments (within and outside the core executive), a panoply of policy instruments, and the collaboration of pressure groups with blackmail potential are simultaneously needed. The technocratic capture potential (or insulation from pressure groups and societal pressure) of transport policy is inherently low, independently from the macro-constellation of domestic veto players in the political process.

The second observation (not totally independent from the first) concerns the balance between policy formulation-adoption and policy implementation. The management of monetary policy is all about policy formulation. Implementation concerns the reaction of the financial markets, but political and administrative structures are not directly involved. So much so that to announce a monetary policy is often equivalent to produce results via the expectations of market players. Tax policy is completely different. To draft a rule on tax avoidance may require the same length of time required by a decision on interest rates (although the insulation from pressure groups is quite different – but this pertains to the first observation) but to implement tax avoidance schemes is a laborious task involving tax inspectors, banks, other financial institutions, and occasionally operations coordinated with the police. Concluding on this point, my hypothesis is that the more the adoption-implementation balance veers towards implementation, the more problematic the process of Europeanization is.

The third observation is about policy discourse, that is, the discourse that provides a rationale and justifies change at the policy level. Policy discourse intersects with but is different from the broader discourse on Europe mentioned above in that it operates at the specific policy level (for example, pension reforms, or privatization, and so on). Institutional capacity and timing provide the potential for change, but policy change has to be considered legitimate.

The empirical analysis of discourse presents its own problems. One the one hand, there is the risk of reification of discourse. On the other, an empirical determination of discourse based on the press and the official speeches of key politicians remains too superficial and elusive. A possible way out of this dilemma is to draw upon the insights of discourse analysis provided by international relations (Milliken 1999). Another is to address the specific forms taken by policy discourse. Policy discourse, indeed, can take different shapes. Discourse can be cast in the form of policy narratives. In a policy narrative, a plot connects events by dint of causal mechanisms. By contrasting the inertial scenario with the favorite scenario, narratives provide a sense of necessity and suggest that certain courses of action are urgent and legitimate. Policy narratives are amenable to empirical analysis in a variety of forms (Roe 1994).

The emphasis on legitimacy and discourse brings us to the crucial role played by the belief systems. Europeanization processes are filtered and refracted by systems of policy beliefs. As shown by Sabatier and Jenkins-Smith (1993), the belief system can be articulated in three levels, that is, ‘deep core’ beliefs (normative and axiomatic), ‘policy core’ beliefs (that is, empirical and normative beliefs concerning the fundamental policy strategies), and ‘secondary’ beliefs. An important issue is to what extent and under which conditions Europeanization can change ‘policy core’ beliefs and facilitate learning and non-incremental change.

Taking the three observations together, they put emphasis on the constellation of actors, the type of policy and the role of belief systems. Turning to the final observation, it is therefore surprising that there has not been systematic analysis of Europeanization in terms of frameworks well grounded in policy theory. To mention only one example, one could think of the advocacy coalition framework(13). The latter is not the antidote to all the research puzzles raised by Europeanization. But at least it provides a framework for the systematic analysis of policy change over a decade or so. It assigns great importance to the belief systems and to the balance between endogenous learning dynamics and exogenous shocks. As such, it is suited for an analysis of the interplay between ‘exogenous’ EU policy and domestic ‘endogenous’ factors. Further, it provides an integrated framework for the analysis of constellations of actors, by grouping them into a number of coalitions with different belief systems. In conclusion, one option open to future research is to make use of alternative frameworks and research designs, and systematic approaches to policy change can provide useful insights.

5. Tentative conclusions

At this preliminary stage of research, it is impossible to draw neat conclusions on the results achieved by the projects on Europeanization informed by political science. However, an important achievement is that fundamental questions on the impact of the EU on member states have been raised. The risk of conceptual stretching looms large, but improvement is on its way. Taxonomies that ‘unpack’ concepts and make them amenable to empirical research can reduce the risk of concept misformation. Research can be organized by considering a matrix that includes the political entities affected by Europeanization and the extent (and direction) of change.

Research designs are still too rigid. They are limited to the analysis of ‘European effects’ in certain areas of change, but they do not control for rival alternative hypotheses. As Goetz observes, ‘alternative explanations for changes observed are rarely considered systematically in the context of Europeanization research’ (Goetz 2000:19). Future research should control for the relative impact of Europeanization (for example, in the area of administrative change Europeanization can be an intervening variable in processes of modernization and reform) and for rival alternative hypotheses. It may be difficult to unravel different causes, but this is the best way to be sure that changes observed at the national level are originated by EU dynamics, and not by other forces.

In terms of substantive change, the jury is still out on critical issues such as whether Europeanization is producing a differential Europe or substantive policy convergence. As for the amount of change induced by Europeanization, the answer depends on the level of analysis: macro-analyses of political structures detect low levels of Europeanization, whereas studies at the policy level signal a more consistent impact. The policy level seems the most exciting. Not only has public policy become Europeanized, but also the Europeanization of policy has triggered processes of transformation of the state. This dynamic effect of policy upon politics has taken place independently from the macro-characteristics of political systems (be they dirigiste, corporatist of polarized). In some cases, the Europeanization of public policy has changed the state even in the shadow of formally immutable institutions, as Giuliani (1999) argues in his analysis of change in the context of stalemate on institutional reforms. One suggestion is therefore to intensify research at the policy level, beyond the picture of political systems provided by conventional comparative politics, and make the most of comparative public policy analysis.

How can policy research be ‘intensified’ and hopefully improved? Europeanization is a process. It is a process where the cognitive dimension of political life matters. Hence the current emphasis on mechanisms and variables should not preclude the dimension of evolution, learning and the social construction of politics. Evolution and learning require frameworks of analysis sensitive to policy change over the medium-long term. One option that researchers may consider is to use the advocacy coalition framework(14). This framework is not panacea – as shown by the symposium hosted by the Journal of European Public Policy (March 2000) – but is suitable for studies of policy change centered on belief systems, legitimacy and the conflict between reformers and advocates of the status quo.

Research on Europeanization could also benefit from the considerable amount of knowledge generated by studies of the international sources of domestic policy change, that is, the so-called second-image reversed perspective on international relations (Gourevitch 1978). The volume edited by Green Cowles et al. (2000) goes in this direction, and in so doing it contributes to cumulative research (as opposed to ad hoc research). Although the European dimension presents significant peculiarities, one does not see the point of starting theoretical research on Europeanization from tabula rasa, and perhaps re-inventing propositions already well known on the role of domestic institutions as filters of international forces.

Overall, Europeanization has opened a new avenue for debate between comparative politics, comparative public policy and more traditional EU studies. If it does not confine itself to sui generis speculations and ad-hoc theorization, research on Europeanization has considerable potential for our understanding of the evolution of state structures and public policy.

References

Andersen, S.S. and T. Burns (1996) 'The European Union and the erosion of parliamentary democracy: A study of post-parliamentary governance' in Andersen, S.S. and K.A. Eliassen (Eds) The European Union: How Democratic Is It?, Sage: London, 227-251.

Börzel, T. (1999) 'Towards convergence in Europe? Institutional adaptation to Europeanization in Germany and Spain', Journal of Common Market Studies, 39(4), December: 573-96.

Bulmer, S. and M. Burch (1998) 'Organizing for Europe: Whitehall, the British state and European Union', Public Administration, 76(4): 601-28.

Caporaso, J. (1996) 'The European Union and Forms of State: Westphalian, Regulatory or Post-Modern?, Journal of Common Market Studies, 34 (1) March: 29-52.

Checkel, J.T. (2000) The Europeanization of citizenship?, in Green Cowles et al. Transforming Europe: Europeanization and Domestic Change.

Cole, A. and H. Drake (2000) 'The Europeanization of the French polity: continuity, change and adaptation', Journal of European Public Policy, 7(1): 26-43.

Coombes, D. (1968) Towards a European Civil Service, London: Chatham House.

DiMaggio, P.J. and W.W. Powell (1991) 'The iron cage revisited: institutional isomorphism and collective rationality in organizational fields' in Powell, W.W. and J. DiMaggio (Eds) The New Institutionalism in Organizational Analysis, University of Chicago Press: Chicago and London, 63-82.

Dudley, G. and J.J. Richardson (1999) 'Competing advocacy coalitions and the process of frame reflection: A longitudinal analysis of EU steel policy', Journal of European Public Policy, 6(2), June: 225-48.

Dyson, K. and K. Featherstone (1996) 'Italy and EMU as `Vincolo Esterno': Empowering the Technocrats, Transforming the State', South European Society and Politics, 1(2), Autumn: 272-99.

Dyson, K. and K. Featherstone (1999) The Road to Maastricht, Oxford University Press: Oxford.

Fabbrini, S. (1998) 'Due anni di governo Prodi: un primo bilancio istituzionale', Il Mulino, XLVII(378): 657-72.

Featherstone, K. (1998) ‘Europeanization and the centre periphery: The case of Greece in the 1990s’, South European Society and Politics, 3(1) Summer: 23-39.

Ferrera, M. and E. Gualmini (1999) Rescue from without? Italian social policies 1970-1999 and the challenges of internationalization, EUI working papers, EUF no.99-13, Florence, European University Institute.

Genschel, P. and T. Plümper (1997) 'Regulatory competition and international cooperation', Journal of European Public Policy, 4(4): 626-42.

Giuliani, M. (1996) 'Italy' in Romestch, D. and W. Wessels (Eds) The EU and Member States. Towards Institutional Fusion?, Manchester University Press: Manchester, 105-133.

Giuliani, M. (1999) Europeanization and Italy, Paper presented to the 6th biennial conference of the European Community Studies Association, Pittsburgh, 2 – 5 June 1999.

Goetz, K.H. (1995) ‘National governance and European integration: Inter-governmental relations in Germany’, Journal of Common Market Studies, 33(1) March: 91-116.

Goetz, K. (2000) Europeanizing the national executive? Western and eastern style, Paper prepared fot the UACES 30th annual conference, Budapest, 6-8 April 2000.

Gourevitch, P. (1978) 'The second image reversed: the international sources of domestic politics', International Organization, 32(4): 881-911.

Green Cowles, M. (2000) ‘The TABD and Domestic Business-Government Relations: Challenge and Opportunity’ in Green Cowles et al. Transforming Europe: Europeanization and Domestic Change.

Green-Cowles, M., Caporaso, J. and T. Risse (2000) (Eds) Transforming Europe: Europeanization and Domestic Change, Ithaca and London: Cornell University Press.

Hall, P. (1993) 'Policy Paradigms, Social Learning and the State. The Case of Economic Policy-Making in Britain', Comparative Politics, 25(3) April: 275-96.

Harcourt, A.J. (1999) Engineering Europeanisation: the role of the European institutions in shaping national media regulation, Paper presented to the Max Planck Institute seminar series, Bonn, November 1999.

Harcourt, A.J. (2000) European institutions and the media industry. European regulatory politics between pressure and pluralism. Ph.D. Dissertation, Department of Government, University of Manchester.

Héritier, A. (1998) Differential Europe: Administrative responses to Community policy, Paper presented to the conference on Europeanization and Domestic Change, EUI, Florence, 19-20 June 1998.

Héritier, A. and C. Knill (2000) Differential responses to European policies: a comparison, Max Planck Projektgruppe Recht der Gemeinschaftsgüter Preprint 2000-3, Bonn.

Jachtenfuchs, M., T. Diez, and S. Sung (1999) 'Which Europe? Conflicting models of a legitimate European political order', European Journal of Political Research, 4(4): 409-45.

Kassim, H. (1994) Policy Networks, Networks and EU Policy Making. A Sceptical View, West European Politics, 17 (4) October: 17-27.

Kohler-Koch, B. (1996) 'Catching up with change. The transformation of governance in the European Union', Journal of European Public Policy, 3(3), September: 359-80.

Kohler-Koch, B. (1999) 'The evolution and transformation of European governance' in Kohler-Koch, B. and R. Eising (Eds) The transformation of Governance in the European Union, Routledge: London, 14-35.

Knill, C. and D. Lehmkuhl (1999) How European matters: different mechanisms of Europeanization, European Integration online Papers (EIoP) Vol. 3 (1999) N· 7. http://eiop.or.at/eiop/texte/1999-007a.htm.

Knill, C. and A. Lenschow (2000) ‘Adjusting to EU environmental regulatory policy: Change and persistence of domestic administrations’, in Green Cowles et al. (Eds) Transforming Europe: Europeanization and Domestic Change.

Ladrech, R. (1994) 'Europeanization of domestic politics and institutions: The case of France', Journal of Common Market Studies, 32(1), March: 69-88.

Lawton, T. (1999) ‘Governing the skies: Conditions for the Europeanisation of airline policy', Journal of Public Policy, 19(1): 91-112.

Mair, P. (1999) The Europeanization of domestic politics: The limited case of national party systems, Paper given to the conference on Multi-Party Systems: Europeanization and the Reshaping of National Political Representation, European University Institute, Florence, 16-18 December 1999.

Majone, G.D. (1996) Regulating Europe, London: Routledge.

Mazey, S. and J.J. Richardson (1993) (Eds.) Lobbying in the European Community, Oxford University Press: Oxford.

Marcussen, M. (2000) Denmark in the EMU semi-sovereignty game’ Paper preapred for the 30^th anniversary conference of UACES, Budapest, 6-8 April 2000.

Morlino, L. (1999) Europeanization and representation in two Europes. Local institutions and national parties, Paper given to the conference on Multi-Party Systems: Europeanization and the Reshaping of National Political Representation, European University Institute, Florence, 16-18 December 1999.

Olsen, J.P. (1995) Europeanization and nation-state dynamics, Oslo: Arena working paper.

Page, E.C. and L. Wouters (1995) 'The europeanization of national bureaucracies?' in Pierre, J. (Ed.) Bureaucracy in the Modern State. An Introduction to Comparative Public Administration, Edward Elgar: Aldershot, 185-204.

Petit, E. (forthcoming) ‘Europeanization and domestic politics: Europe and the development of a French environmental policy for the agricultural sector’, Journal of European Public Policy.

Radaelli, C.M. (1997) 'How does Europeanization produce policy change? Corporate tax policy in Italy and the UK', Comparative Political Studies, 30(5), October: 553-75.

Radaelli, C.M. (1998) 'Networks of expertise and policy change in Italy', South European Society and Politics, 3(2), Autumn: 1-22.

Radaelli, C.M. (1999) Technocracy in the European Union, Harlow (Essex): Longman.

Radaelli, C.M. (2000a) 'Logiques de pouvoirs et récits dans les politiques de l'Union Européenne', Revue Française de Science Politique, 50(2), April: 255-75.

Radaelli, C.M. (2000b) Policy transfer in the European Union, Governance13(1) January: 25-43.

Roe, E. (1994) Narrative Policy Analysis, Duke University Press: Durham.

Sabatier, P. and H.C. Jenkins-Smith (1993) Policy Change and Learning. An Advocacy Coalition Approach, Westview Press: Boulder.

Sabatier, P.A. (1998) 'The advocacy coalition framework: Revisions and relevance for Europe', Journal of European Public Policy, 5(1), March: 98-130.

Sabatier, P. (1999) (ed) Theories of the Policy Process, Boulder, Westview Press.

Sartori, G. (1970) 'Concept misformation in comparative politics', American Political Science Review, 64(4), December: 1033-53.

Sartori, G. (1976) Parties and Party Systems, Cambridge: Cambridge University Press.

Sartori, G. (1991) 'Comparing and miscomparing', Journal of Theoretical Politics, 3(3): 243-57.

Schmidt,V.A. (1997) 'Discourse and (dis)integration in Europe: The cases of France, Germany, and Great Britian', Daedalus, 126(3), Summer: 167-97.

Schmidt, V. (1999) Discourse and the legitimation of economic and social policy change in Europe, Paper prepared for delivery for Roundtable 4, "Les effets d'information. Mobilisations, préférences, agendas." Association Française de Science Politique, 6è Congrès, Rennes, France 28-31 Sept. l999.

Schmitter, P. and J. Santiso (1998) ‘Three temporal dimensions to the consolidation of democracy’, International Political Science Review, 19(1): 69-92.

Schneider, V. (1998) Europeanization and the re-dimensionalization of the public sector: Telecommunications in Germany, France and Italy, Paper presented to the conference on Europeanization and Domestic Political Change, European University Institute, Florence, 19-20 June 1998.

Shaw, J. and G. More (1995) (Eds) New Legal Dynamics of the European Union, Oxford: Clarendon Press.

Smyrl, M. (1997) 'Does EC regional policy empower the regions?', Governance, 10(3), July: 287-309.

Stone Sweet, A. (1999) 'Judicialization and the construction of governance', Comparative Political Studies, 32(2), April: 147-84.

Stone Sweet, A. and T.L. Brunell (1998) 'Constructing a supranational constitution: Dispute resolution and governance in the European Community', American Political Science Review, 92(1), March: 63-81.

Sun, J.-M. and J. Pelkmans (1995) 'Regulatory competition in the single market', Journal of Common Market Studies, 33(1), March: 67-89.

Weiler, J.H.H. (1994) 'A quiet revolution: The European Court of Justice and its interlocutors', Comparative Political Studies, 26(4), January: 510-534.

Wincott, D. (1998) 'Does the EU pervert democracy? Questions of democracy in new constitutionalist thought on the future of Europe', European Law Journal, 4(4), December: 411-28.

Endnotes

(*) This paper arises out of preliminary research on Europeanization conducted for my project on Italy in the Euro-zone: institutions, discourse and policy regimes funded by the British Academy. The first draft of this paper was written during a visiting fellowship at Nuffield College, University of Oxford (Hilary term, 2000). I gratefully acknowledge the financial assistance of the British Academy and the hospitality of Nuffield College. The draft was delivered to the annual conference of the Political Studies Association (London, 10-13 April 2000) and to the International Workshop on Europeanization organized by the Research Unit on Europeanization, Bradford University (5 and 6 May 2000). I wish to thank the participants to these events and two anonymous referees for their comments; the usual disclaimer applies.

(2) The process can be more or less incremental. Indeed, one the most interesting areas of research on Europeanization concerns the time, timing and tempo of the process. See below, section 4.

(3) The resignation of the Santer Commission in 1999 exposed how far the EU is from providing a model to its member states!

(4) Börzel uses the term ‘institutional culture’ but she does not provide a definition. Arguably, the notion of policy style is better suited to encompass the cultural dimension Börzel has in mind.

(5) Cole and Drake (2000) examine Jospin’s reaction to this deficiency of discourse. They argue that the French discourse on Europe has been recast and rationalized in pragmatic ways, by accepting an acceleration of European integration in the context of subsidiarity. This is producing a sea-change in the French ‘intellectual mindset’ (Cole and Drake 2000:39) in that it allows France to import models and values of modernization from Europe, in contrast to ‘the traditional French stance of exporting French (i.e., ‘universal’) values to the rest of the world via Europe’ (Cole and Drake 2000:39).

(6) Schmidt adds that there are two dimensions of discourse. Coordinative discourse is used by elites to create policy, whereas communicative discourse relates to the communication to the mass public. This distinction has potential, as shown by the Danish case.

(7) Hall (1993) provides a different typology, suitable for the analysis of policy change, based on paradigmatic change, change of instruments an change in the levels of instruments.

(8) For example, economic incentives instead of command and control regulation, the introduction of a new policy instrument, the formal separation between infrastructures and provision of service in railways, or the modification of the legal status of public employees’ contracts.

(9) I am grateful to an anonymous referee for this observation.

(10) I am grateful to Ed Page for this comment.

(11) Time refers to when a decision is made, timing to the sequencing of decisions and tempo to the rate of speed.

(12) Giuliani (1999) employs the expression ‘bottom-up’ Europeanization to identify the un-expected impact of policy upon politics.

(13) For recent developments and an illustration of the framework see Sabatier (1998;1999).

(14) See Dyson and Featherstone (1999) and Radaelli (1998) for applications of this framework to Europeanization.