|Comparing Europeanisation Effects: From Metaphor to Operationalisation|
|European Integration online Papers (EIoP) Vol. 7
(2003) N° 13;
|Date of Publication in : 30.12.2003|
|| Abstract |
Back to homepage |
| This paper's comments page | Send your comment! to this paper |
An impressive array of fascinating studies has been published in recent years on the relative success and failure of "Europeanisation" in a couple of policy areas.(1) This refers to the reactions in domestic systems to top-down influences from the EU-level, be they directly induced by EC law or indirectly by European policies such as the Maastricht convergence criteria (e.g. Ladrech 1994). An important sub-field of Europeanisation studies tackle the implementation of EU law in the member states. Such studies should, at least in theory, be ideal material for comparison between countries and between policies.
However, it seems that operationalising the misfit and the effects created by EU policies in such a way as to allow for a comparison not only within the (typically few) cases of any single author or any small team of authors has not yet received enough attention. Therefore, it is impossible at this stage to undertake meaningful secondary analyses of the materials collected by different authors. This paper argues that better yardsticks for inter-case comparison are indispensable if we want to go beyond interesting individual case and/or sector studies and to approach a profound process of additive insight on the process of Europeanisation.
The empirical material used in this paper stems from a collaborative research project that analyses the national transposition, enforcement, and application of six EC labour law Directives in all fifteen member states (see http://www.mpi-fg-koeln.mpg.de/socialeurope). These 90 cases of implementation of EU law allow us to analyse, in a comparative perspective, specifics of both Directives and individual member states. Approximately 180 interviews were conducted with experts from the administrations, interest groups, and labour inspections in all member states. This project is extraordinary in that it performs a full survey for a number of Directives, rather than studying only a selection of member states as common in the literature.
The purpose of this paper is not to summarise this project but to offer some of the theoretical and methodological insights gained directly or indirectly in its context. In particular, some of our operationalisations could be of interest for other research groups, too. They concern both the independent (i.e. the factors that typically are expected to determine the outcome of Europeanisation processes) and the dependent variables of Europeanisation (i.e. the outcome of the adaptation process). From our empirical findings, only some "appetizers" can be presented to invite further reading.
This article will now suggest an operationalisation of the "misfit" between European rules and pre-existing national standards that have to be replaced by the former (2.a). According to a number of implementation theorists, this can be expected to be the main individual factor determining Europeanisation success or failure. Our argument is that we need to specify much more clearly than hitherto both the different dimensions of "misfit" and the related aggregation systems. Only on this basis can the "misfit hypothesis" be a truly testable proposition, in the first place (a second criterion being the proper measurement of adaptation success, see below 3.). Following the abstract discussion of operationalisation, this paper will briefly present some empirical findings on the kinds and degrees of misfit actually generated in 90 cases (2.b). It will subsequently discuss two aspects within the operationalisation of Europeanisation effects in policy implementation, i.e. the degree of correctness of implementation (3.a) and the multiple levels of Europeanisation effects that should ideally all be taken into consideration (3.b). The conclusions underline the importance for Europeanisation studies to move on from "metaphor to operationalisation".
This section defines three analytical dimensions of adaptation requirements triggered by EU law: policy misfit, politics/polity misfit, and costs. These dimensions are discussed with empirical examples at hand. They are combined in an index of overall misfit (none, low, medium, high) than can be quantified to fit a theoretical range from 0 to 18 in the case of six Directives studied. On this basis, the empirical patterns found in 90 implementation cases are then discussed in brief. It seems promising to compare further cases from other policy areas in the light of the same approach, in the future.
Misfit with the given situation in a member state has been highlighted as the crucial explanatory factor for implementation performance in much of the recent literature on European integration. The argument is based on the assumption that one can expect a smooth implementation process if a Directive requires only small changes to the domestic arrangements. Implementation problems, by contrast, are expected if a considerable degree misfit must be rectified by a member state (see e.g. Duina 1997; Duina 1999; Duina et al. 1999; Knill et al. 1998a; Knill et al. 1999; Knill et al. 2000a; Börzel 2000a; Börzel 2000b).
This strong emphasis on misfit even on the level of theory adds to the great importance that must be attributed to this factor also from a policy analysis perspective: We can only estimate the practical effect of any EU policy if we know where the member states started their process of adaptation. In other words, establishing in a detailed manner both the status quo ante in the member states and the demands embedded in any European Directive is crucial. This is, at least in research practice, far from easy. Much EU regulation touches intricate details of national legislation that no one but a national expert can know. The great effort needed explains why qualitative implementation studies have traditionally only analysed a few cases. Although the recent literature already goes far beyond what had been offered in earlier work, we found that for our study of six Directives in all 15 EU member states (i.e., 90 implementation cases in total) we needed a much more differentiated approach. While it is easy to state that there is a misfit between any EU measure and the domestic situation in a specific member state, it is much more difficult to conceptualise this misfit in such terms as to allow a direct comparison between countries and even between different measures.
Two steps are indispensable: categorising forms of potential misfit (which we will do first here), and operationalising the degree of misfit (see later below). With regard to form, misfit can be substantive, i.e. relate to content and hence be policy-related (policy misfit), or rather apply to matters of procedure (i.e. affect politics and/or polity).
Policy misfit(2) means that the contents of e.g. an EC labour law Directive are not fulfilled in the relevant national law. This can refer to a gradual difference (e.g., two months of parental leave instead of three as a minimum) or to a matter of principle (e.g., there is no individual right to parental leave but only one attributed to the couple that has to decide which of the two parents takes it). Europeanisation can hence be of a quantitative (more or less of an existing policy) or a qualitative kind (new/replacement of national institutions or structures).
For sure, the policy misfit may in some cases appear more important on paper than in practice. We call the former aspect the legal misfit, and calculate a kind of discount in case the practical significance is comparatively smaller. For example, a new right may not have been laid down in generally binding domestic law before, but it may at the same time have been attributed to a large part of the workforce either in collective agreements or due to a specific social tradition. Furthermore, it is important to include in the concept of legal misfit an evaluation of the scope of application. In other words, we look at the coverage of any newly attributed right. The importance of such a right may, in some cases, seem very important, but may then be seriously limited by a narrow scope of application (e.g., when all atypical workers are excluded). In short, our concept of substantive misfit takes due account of both the legal misfit and its practical significance. This helps painting realistic picture of when governments should be expected to (mal-)comply with EU law, from a viewpoint of misfit-centred implementation theory.
No less difficult is attributing a size category to the misfit actually found in a specific case. We define as a large-scale legal misfit, if there are completely new legal rules, far-reaching gradual changes, and/or important qualitative innovations. Each of them will lead to high level of substantive misfit in our system under the condition that all or a significant number of workers are concerned, and that there is no important limitation on the level of practical significance. Otherwise, only a medium-level substantive misfit will result in our table. Table 1 indicates how a similar logic is applied to medium and low degrees of legal misfit. Note that the basis of evaluation in terms of large/medium/small is the significance in the context of the national labour law standards, while the comparison with other member states and other cases will take place on the basis of the degree of misfit established for each of the countries.(3)
The policy misfit is the total in this table, but it is not yet the end of our efforts to establish the misfit in a given case of adaptation requirements. Beyond substantive rules, also aspects regarding politics and/or polity can misfit EU-level rules (as outlined above).(4) In the environmental policy area, there are manifold administrative routines affected by European Directives. Sometimes even new bodies have to be set up to comply with procedural regulation stemming from the EU-level.(5) In social policy, this is much less common, and in the area of labour law studied here, this is not the case at all. Nonetheless, public-private interaction patterns are sometimes affected by European integration. A very shining example is that employee consultation patterns may be laid down in EC Directives. Our sample does not include any of these laws, but still we found misfit in the public-private field (i.e., on the politics/polity dimension). This is because even Directives that are concerned with substantive EC labour law have to be implemented in such a way as to conform to procedural European requirements. For example, the national transposition law must cover all workers included in the field of application fixed in the Directive. Since they typically cannot cover all or the workforce, collective agreements will not suffice. This disturbs the traditional pattern of labour law making in Denmark and Sweden.(6) This meant that some member states had to adapt their institutionalised national ways of policy-making in the field of labour law, due to European integration, as far as the implementation of EC Directives is concerned (for details see Simone Leiber's forthcoming dissertation 2004).
We define as a comparatively large degree of misfit on the politics/polity dimension if there is a challenge to a crucial domestic institution or procedure (e.g., the "Danish model of social partner autonomy"). A medium degree of such a kind of misfit refers to a less important but still very significant domestic institution or procedure (e.g. the freedom to derogate from working time regulation by collective agreement, in Sweden). The misfit on the politics/polity dimension is as important a part in our calculation of overall misfit as is the policy misfit (see Table 2 on our aggregation rules, below).
Finally, another crucial element of any estimation of misfit caused by EU regulation must be costs, i.e. the economic consequences (as opposed to, e.g., the citizenship dimension) of a required reform for the addressees on all levels. Costs should not be confused with any of the forms of misfit outlined above. Large-scale policy misfit can still only cost small sums of money (e.g., if a new right is attributed to a group of people where hardly any take-up will take place), and sometimes, small legal changes amount to high costs (e.g., in the field of working time standards).
Establishing the exact costs of adapting to a EU Directive for any specific country is hardly possible. First, many kinds of actors are involved. Costs may fall on different units of the state, on semi-public and on private actors or companies.(7) If there is publicised data on the costs of adaptation at all, they typically stem from interested actors. Note that even governments are interested actors, in a wider sense, since adaptation costs can be used in the debate over the pros and cons of European integration, both in general and in social policy, in particular. The most detailed data on expected adaptation costs stem from the UK Department of Trade and Industry (Treib 2004). The real costs of adaptation in practice will typically never be known. It is even doubtful if they matter at all, at least when it comes to studying implementation performance. In that context, the costs of adaptation that can realistically be expected by the relevant national and international actors seem much more important. They, too, are difficult to establish. We consider a crucial step to be the defining of cost categories and the evaluation of their potential. With our field of labour law at hand, this indeed facilitates the comparative assessment a lot. In a first step, we established the cost categories that a given Directive can potentially trigger in any member state. Secondly, we established empirically in interviews how many groups of workers and sectors are actually concerned in the 15 member states. On the basis of the our interviewees' judgements of the costs, which we compared to the costs mentioned for other countries and other Directives, we could categorise without too many problems the costs of adaptation triggered in a given member state, on a scale of low, medium, and high.
In the field of labour law, the costs created by any EU Directive are typically costs for private and public employers, sometimes for the social security system. The costs arising from the various labour law Directives studied in this book fall into six categories:
We categorise the first two types to create, at least potentially (it certainly depends on the situation in a specific country), high costs; the third and fourth will at best trigger medium-sized costs, the fifth and sixth at maximum low costs. Note that in empirical cases, the costs even for a potentially high-cost Directive may be small, and often the elimination of misfit will only demand an administrative burden with rather insignificant costs. Beyond the short-term cost potentials, we also look at possible long-term consequences of EC labour law Directives in the member states and at times, considering a longer horizon may increase the potential costs.
The aggregation of all dimensions of misfit outlined above is, for sure, all but unproblematic. Only on the basis of an explicit system of operationalising and aggregating misfit, however, can we compare many cases of Europeanisation pressure across both Directives and countries. We decided to rate high misfit on any one dimension (substantive misfit, misfit on politics/polity dimension, or costs) as a "high" on the total misfit created by a particular Directive in a particular country. This follows the logic that no dimension of misfit can eradicate the significance of adaptational pressure on another dimension (at least not when it comes to testing the importance of misfit size in terms of good or bad compliance). A high degree of misfit in terms of the national procedural patterns (or, alternatively, in terms of a specific new right granted to workers) cannot be outweighed by the fact that e.g. the costs may be low. In turn, significant levels of costs seem an important factor regardless of the abstract importance of the changes in terms of substance or politics. Consequently, our "overall misfit" consists of the highest parameter value found in the three sub-categories.(8)
In applying the system developed above in practice, we start out with the costs. We first need to establish the maximum potential costs for our six Directives outlined in Table 3 (letters refer to the cost categories listed above):
In the case of parental leave (as the only example in our sample), there is a potential for comparatively much higher long-term costs since more men might take up their right in the future. For our analysis of factors that potentially affect the transposition performance, we did not take this into account since the interviews revealed that politicians and experts either did not consider the longer-term perspective to diverge significantly from the present, or did not take this into their short-term consideration.
How much misfit did we find in actual practice, in the 15 EU member states? Table 4 lists four potential degrees of costs and the number of cases found in each.
These results suggest that our operationalisation of costs did not over-estimate the effect created by European social Directives: in 84 per cent of our cases, the costs are at best low. This seems a valid result if we take into consideration that social policy is a field where the member states have old and well-established domestic models. Each and every member state had, more specifically, an elaborate system of labour laws even before entering the EU. The Directives therefore rather refined what was already there, and added costs that seem minor if compared to, for example, the costs created when the first national laws in the field were adopted. Additionally, it is crucial to note that minimum standards in labour law are by far not the most expensive part of the EUs social policy efforts. Much higher costs are created with the social security Regulations that secure equal treatment for workers from other EU member states and the cross-national addition of social security claims (see e.g. Falkner 2003). So the scale of costs is realistic in the frame of intra-policy comparison. On the level of inter-policy analysis, too, the rather small amount of costs created by our six Directives seems adequate. As already mentioned for environmental policy, other fields show a much more revolutionalising effect of EU activity.
This does not indicate that the innovation introduced by our six Directives is negligible. Costs are but one dimension of misfit. Looking at the other elements of overall misfit as operationalised above, we find more of a departure from the domestic status quo ante in the 15 member states. Still, only a small number of the total 90 cases show large-scale misfit (10 cases), while almost exactly 50% show low levels of misfit. Again, this seems realistic if one wants to be able to use the operationalisation for intra-policy and/or inter-policy comparison.
This operationalisation furthermore allows to compare the need for adaptation arising for different member states in a much more specific way than practiced hitherto. We attribute 0, 1, 2, or 3 points for no, low, medium, or high overall misfit.
This makes evident that the highest amount of misfit for our six labour law Directives was created in the UK (for details see Oliver Treib's forthcoming dissertation 2004). This is hardly surprising, considering that this country has the most non-interventionist labour law tradition within the EU. It is well known that the UK even had an opt-out from EU social policy harmonisation under the Maastricht Treaty, which ended only when the Labour government came into office and accepted the social Directives adopted by the other member states (see e.g. Falkner 2002).
The Irish situation is somewhat similar, but there is a different economic situation explains the comparatively low protection standards preceding the EU regulation there. Against this background, it was even more impressive that the Irish government did not oppose minimum harmonisation in the social realm as harshly as the UK and accepted the Maastricht Social Agreement (Falkner 1998: 87). Denmark, too, is one of the rather reluctant countries when it comes to EU social regulation. This is, however, not due to expected high adaptation pressure but rather to a general effort to protect national autonomy. In this light, the comparatively high misfit actually created by our six Directives comes as a surprise. A closer look at our elaborate operationalisation solves the puzzle: the large-scale misfit on the politics dimension explains the rather high overall misfit. In fact, the Danish system of social partner autonomy in the regulation of working conditions has been rather strongly affected by EU social policy and had to be changed in the Direction of more state intervention (Leiber 2004; Falkner et al. 2003).
On the other end of the continuum for overall misfit are the Netherlands, France, and Spain. The latter may come as a surprise for those who assume that the Southern members are laggards in social policy, in general. This assumption is certainly not true in labour law and our operationalisation leads to a result that fits the expectation of the more initiated. The reason is that Spain, Portugal, and (to a more limited extent) Greece started, by the time of their EU entrance, from a legacy of extremely rigid regulation. It had originated in prior authoritarian regimes and was only slowly reformed in domestic reforms afterwards. This made EU labour law look rather minimalist in its re-regulative elements, and the liberalising elements were at times those that accounted for misfit (for details see Miriam Hartlapp's forthcoming dissertation 2004). In general, it should be stressed that there is no clear and consistent north-south divide in the regulative level of labour law. This is underlined by our findings that Spain shares the last place (ranks 13-15) when it comes to overall misfit, Greece ranks 8-10 and Portugal 4-5. At the same time, Sweden and Austria show a much larger misfit than protagonists of a north-south split would expect. This mirrors our findings from the 90 in-depth studies and the approximately 180 expert interviews (which gave a more general overview on the broader field of EU social policy) that even in member states with very advanced welfare and labour law systems, the ECs social Directives do at times create significant misfit.(10)
The above highlights the crucial importance of elaborate operationalisation of adaptation requirements in the study of Europeanisation. The following section will argue, however, that referring to a clearly specified and interpersonally comparable concept of misfit is just one of several prerequisites for establishing adaptation records.
In order to capture Europeanisation effects in a reliable and comparable manner, one also needs to spend time on what is usually seen as the dependent variable in Europeanisation or implementation studies, i.e. the result of adaptation processes. Our study of 90 cases very soon revealed that it is often far from self-evident if implementation is actually accomplished, or not (see section a).
Additionally, there are several levels of Europeanisation effects, and the concept of misfit as typically applied in recent works should therefore ideally be supplemented. It does not suffice to focus on the (non-)elimination of kinds and of degrees of misfit that can be established on the basis of comparing EU law and national rules. At least not when one wants to go beyond an implementation study (in the narrow sense) and to accomplish a broader study on Europeanisation effects in a particular area. Further Europeanisation effects can be observed, on the "higher" level of regime type, and at the "lower" level of individual actor behaviour (see section b below). They each matter in their own right and should therefore be taken into consideration, even in those cases where no visible effect on the level of policy is (as yet) visible.
For sure, it is useful to compare the adaptation outcome with the misfit created by any EU law. Our study revealed, however, that one needs intermediary categories in a number of cases where the categories applied to misfit are too crude to capture all aspects of the practical outcome. The complexity of this issue becomes clear if we take into consideration that a full adaptation in the sense of 100 per cent will in practice often be almost impossible to reach even in the transposition stage of a Directives life cycle (and even more so during application).(11) Often times, many different sectors of the economy and, in addition, many different worker categories are treated individually in different national laws. Consider, for example, the Working Time Directive. In many countries, there are specific laws for sectors (such as mining) and for special groups of workers (e.g. shift work laws). Taking into account that in a number of EU member states, even sub-national regional units will come into play in the field of regulation (e.g. adopting laws for Laender employees who will be exempted in federal laws on the same issue), a hugely complex matrix results in any study aiming to establish the actual adaptation to the EC Directive. Should any remaining minimal misfit on any dimension lead to a countrys scoring as a non-complier? For example, if a minimal point of one standard is not granted to a small group of public sector workers in one region, but is implemented in the rest of the country?
Such a strict view can make sense for certain perspectives (for example, for the European Commission in her enforcement activities). However, if we want to establish the policy effects of EC laws or to systematically analyse which reasons account for (non)-compliance, across many cases, it may make sense to concentrate on the relevant part of the adaptation process. In the end, it may be an irrelevant minor reason that differs from the main national transposition process, if in one of nine or 16 Laender in Austria or Germany a small-scale misfit remains. It seems fair to mention that federalism indeed is a stumbling block for perfect transposition in a number of cases. Stressing unilaterally in each case study any remaining speck of misfit would, however, distort the overall picture greatly. It would, firstly, hide even major adaptation accomplishments and, secondly, stress to an undue amount the reasons that account for the last remaining bit of non-adaptation, to the detriment of those reasons accounting for compliance with larger parts of the misfit, in earlier stages.
For all these reasons, we decided to differentiate between the date when essentially correct transposition is accomplished (and, at a later stage similarly, essentially correct application; for the moment, we shall focus on transposition here), and the date of full legal adaptation without any remaining misfit. Wherever possible we established them both.(12) Correct transposition is the full compliance with adaptation requirements in the transposition stage and notification of the relevant laws to the European Commission. Essentially correct, by contrast, refers to an essentially accomplished adaptation (this contains a qualitative consideration, how good adaptation was and whether it captured the essential) of most requirements and, at the same time, of the most central requirements(13) of any Directive (this contains two quantitative considerations: how many requirements are fulfilled and how many of those that are crucial in terms of the Directives aim). In other words, if a Directive is transposed in an essentially correct manner, only a few elements required for full adaptation may be open, and they cannot be essential in terms of the general goals of the Directive. If, as a fictitious example, a Directive on Doctors working time would be implemented in such a way as to allow doctors in one entire region in a federal state to work double overtime, this would not be essentially correct transposition. If, by contrast, a cross-sectoral Working Time Directive (like the one we studied) concerns a very large number of groups of workers, and if then one group in one out of many (or a few) regions in a federal state is not yet covered by the transposition laws, this can be essentially correct (if not much other misfit remains).(14)
When we discuss the main reasons for non-compliance we focus on the reasons for and timing of essential correctness in the transposition. For the complete impression of compliance problems, it is surely relevant to include the reasons for minor remaining deviations (and we do underline the fact that sub-national units responsibilities for covering part of the workforce accounts for unevenness in the transposition and later application of EC labour law Directives).
In terms of our cases, 21 out of 90 cases (approx. 23 per cent) were still not essentially correct by end of 2002 (note that the transposition deadlines for our six Directives fell within 30 June 1993 and 20 January 2000).(15) Full correctness was not reached in almost half of our cases, i.e. 44 of 90 cases. It is interesting to note that the Directive with the latest transposition deadline (i.e. on part-time work) does not show more cases lacking full correctness (only six) than earlier Directives. The highest number is with the Working Time Directive (twelve cases) where the deadline was 23 November 1996. The lowest number of cases lacking full correctness is with the Directive with the lowest degree of overall misfit, i.e. the Work Contract Directive (four cases). The number of cases still not essentially correct is not only smaller but also more equal across Directives (with a scale from one case to seven cases) and across countries (between 0 and 2 cases).
These findings (crudely summarised here) indicate that there is a non-compliance problem. This has with good reasons at times been contested in the literature, on the basis of available statistical data.(16) In our field of empirical analysis, one no longer needs to rely on data published by the European Commission that represent but "the tip of the iceberg" (Hartlapp 2004; Falkner et al. 2002). Our data stem from ground-level analysis and show the large amount of non-compliance that de facto exists.
Another crucial issue in operationalising the dependent variable in Europeanisation research is categorising the effects in terms of their level of impact. As outlined above, much of the recent implementation literature focused on the level of a specific policy measure. We did the same with our definition of misfit (section 2.a). However, there are levels above and below this scale. "Above" are institutionalised regime types in a particular policy area, and "below" are the strategies and unilateral actions of corporate or at times also individual actors.(17) Both should not be overlooked in implementation studies. Even effects on the lowest level may in the longer run trigger policy reform. Too strong a focus on misfit (and how much of it is eliminated during the domestic adaptation phase) may hide this from our view.
In the general debate on Europeanisation, particularly in earlier times, authors primarily discussed the highest aggregation level, i.e. on the regime type that characterised individual member states. However, only quite dramatic change will be apparent against this background (e.g., the introduction of the Euro in monetary policy or of the Common Market in competition policy, or the change from substantive regulation to process regulation in environmental policy). In our field of labour law, this would for example mean a change from a liberal to an interventionist employment system or from a corporatist to a pluralist style in public-private co-operation; in gender affairs, a change from a conservative model of family and employment to an egalitarian one would be on that level. To give a prominent example, Héritier et al.s impressive study looks at the stage of the domestic policy reform process (pre-reform, reform, post-reform) in terms of national liberalization of transport (Héritier et al. 2001). Since the reforms referred to in their study amount to a quite fundamental change of policy regime (e.g. from public to private ownership, open access or not), this would at least in many cases amount to a change in the basic model or, in other words, to change on our macro level (see Table 7 above).
Whilst this seems a powerful approach in the field of Europeanisation in the transport policy area, we need to go beyond it. Our field of social affairs, in particular labour law, does not allow for such a handy distinction of stages with a generally uniform process of change. Rather, we find countries currently undergoing liberalization processes (e.g. the southern member states which had extremely rigid labour law systems as an inheritance of their authoritarian systems) but also EU member states that recently had the tendency to re-regulate, after having had a period of vigorous liberalization during the 1980s (most importantly, the UK). Additionally, we find several parallel streams of potential policy innovation. For example, a country may regulate more intensively on issues of non-discrimination on grounds of gender or race, while at the same time liberalizing working time regulation.
What we need, in addition to paying attention to potential macro-level change, is therefore an instrument to grasp potential effects in even more detail. After all, relevant Europeanisation effects can take place within the same ideal type of welfare state regime, or of interest group involvement, while not actually shifting a country into another box. Additionally, there may be situations where an internal development made the country move towards one end of the continuum within a category box, so that without Europeanisation a move between types would have occurred. In such a case, the Europeanisation effect could be exactly that there is no move, after all (which would not be visible on the basis of a crude typology). In other words, we need to conceptualise that any stimulus coming from the EU-level meets national factors that are also potentially dynamic. This follows the reasoning of Adrienne Héritier who stresses the parallelism between national and European developments that "intersect and have a reciprocally reinforcing, counteracting, or neutralizing impact" (Héritier 2001: 2).
In order to work with a framework as parsimoniously as possible, one could refer, where it is useful to go beyond the basic type of policy regime in detecting Europeanisation effects, also to the national policy stream (Kingdon 1984). This refers to a wide notion of policy stream, including the relevant discourse.(18) It also includes parts of what Kingdon captured as the political stream (interest group campaigns) and problem stream (for a problem exists not 'as such' but must be brought to attention in a discourse). Just like longstanding institutions of both the narrow (formal systems of rules that structure the courses of actions that a set of actors may choose) and the wider sense (informal cultures), the existing national policy discourse can counteract or reinforce EU-made impulses for change. The same is true for the domestic policy stream, at large.
In addition to the relevant policy discourse, the policy stream covers a content-related dimension (i.e. recent measures adopted in the relevant area(19); compare the substantive policy misfit in Table 2) and a process-related one (i.e. recent patterns of policy-making in the field, which may somewhat differ from the national ideal type model(20); compare the politics/polity misfit in Table 2). This accounts for the fact that even longstanding national ways of doing, in the sense of regime types, may have experienced a period of gradual change (i.e. variation within the realms of the same ideal typical pattern) when they meet the top-down impulse for change, and this may account for a different response than if the same impetus came down on a similar long term rule system that has not seen any recent challenges.
To policy streams and their discourses one can attribute a probability of medium-term stability. In other words, it is not easy for individual as well as collective actors to change the course of events on this level of policy stream, but neither is it impossible. Change at that level happens more often than in case of deeply rooted national institutions (such as overall welfare state or labour law traditions), and, most importantly in our context, even in a rather protective labour law system, the dominant discourse or current policy stream may at a certain time run in the direction of more liberalization. A bit of the latter may be realized without, however, necessarily implying a change of regime. Types of labour law and social security systems, by contrast, stand for a high probability of very long-term stability. That does not exclude change a priori, but it is not realistic to expect such changes but in extraordinary cases.
Finally, what we call the micro level of Europeanisation is strategies chosen by individual or collective actors, i.e. by policy entrepreneurs,(21) which aim to influence e.g. the domestic policy resulting from a specific EU Directive. They can react quite flexibly and may, if they detect a promising benefit in terms of their self-interest in the impulse coming from the EU-level, seize the opportunity and advocate change. A change in their strategy or action may be considered an Europeanisation effect, but it will often not be easily visible, for these actors can meet the adverse interests of other actors who protect the domestic model or policy stream. Nonetheless, we argue that effects on this level should not be overlooked, for the action of individual actors, even while not (yet) changing a policy instrument to be adopted or a regime type, can be a useful indicator for potential change on these levels, at a later point in time.
From a reform probability perspective the concentration on individual actors and their strategic action may however overemphasize change. By contrast, a pure focus on regime types will tend to overly stress continuity. Policy streams, in turn, stand on middle ground. It seems, in any case, that we cannot afford to exclude from the researcher's attention all but the highest level of aggregation. Time saving that may be, but at the price of detecting change only when it amounts to small "revolutions" (in the sense of a change of regime). Smaller changes that might long before herald such events, by contrast, would be overlooked. Inadvertence towards forthcoming developments is, however, only one major shortcoming of such an exclusive approach. Another is that changes at lower levels of aggregation, which are overlooked by a pure institutionalist focus, may still be significant as such. For example, they may alert us to specific features of the European multi-level system (such as that some actors now have many more options available for influencing policy-making).
Institutions, policy streams, and actor strategies can hence be used in Europeanisation research as yardsticks against which to measure change. Combining regime types, policy streams, and strategic action, one can prevent that revolution or inertia is a matter of theoretical perspective (Knill et al. 2001b: 195). Typically, authors approaching their study from a macro-institutional perspective tend to measure the impact against a map of policy legacies and societal structures. Therefore, they tend to arrive at a more continuity-oriented conclusion than authors who focus on rational actor strategies (ibid.: 211).
In an ideal study of Europeanisation, small-scale change (on the level of chosen strategies in individual cases), medium-scale (in the sense of relevant but not revolutionary changes in policy or discourse), and large-scale changes (referring to a revolution of deeply-rooted national regimes) should all be discussed.
For sure, this will often not be feasible for research practical reasons. It is, for example, impossible to analyse in any detailed way all three levels for each and every of our 90 implementation cases.(22) Only gradually declining the levels therefore seems an excellent research-practical recommendation.(23) Since in our policy area of labour law, changes in fundamental regime type are not at stake, but changes on the level of individual actor behaviour are quite frequent, the project described here focused primarily on the meso-level. Even within this, a primary emphasis had to be laid on the level of single laws and the processes leading up to them (see our concept of misfit above 2.). However, we included questions on the dominant discourse and on the strategies of the interested actors in our extensive questionnaire. In some cases, our expert interviews hinted at relevant developments at these levels that could be useful as indicators for potential later change on a comparatively larger scale.
To give one example, our multi-level analysis of six labour law Directives revealed interesting effects on all three levels in Austria, with a view to the polity/politics dimension (see in more detail Falkner et al. 2003).
This paper focused on the operationalisation of a selection of crucial aspects in the study of Europeanisation: the misfit between European and national rules, "correct" transposition, and the levels of Europeanisation in the realm of individual policies.(24)
Why should anyone accept our complex ways of very detailed classification and comparison? Firstly, it is essential to have an explicit yardstick and explicit aggregation rules (e.g., for different kinds of misfit) if our estimation of the extent of Europeanisation demanded by any EU law should not be a matter of subjective perspective. On the basis of such an approach, a comparison between different studies on the implementation of EU law (and of Europeanisation, more generally) would be feasible, and meaningful secondary analyses could finally be done. Secondly, it is important to keep in mind that a pure focus on the misfit and its (non-)elimination hides from our view potentially noteworthy change on other levels. This is where "implementation studies" often are too narrow. We hold that looking at the recent policy stream gives a better impression. This concept (presented above) comprises not only policy content and procedure (both also part of our misfit concept), but also discourse. A full account should, however, additionally look below the threshold of adopted policies and integrate the level of individual actors whose Europeanised behaviour(25) may not immediately impact but might herald future change.
Börzel, T. A., 2000a, Improving Compliance through Domestic Mobilisation? New Instruments and Effectiveness of Implementation in Spain, in: Knill, C., Lenschow, A. (Eds): Implementing EU Environmental Policy. New Directions and Old Problems, Manchester: Manchester University Press, 222-250.
Börzel, T. A., 2000b, Why There Is No 'Southern Problem': On Environmental Leaders and Laggards in the European Union, Journal of European Public Policy 7(1), 141162.
Börzel, T. A., 2001, Non-Compliance in the European Union: Pathology or Statistical Artifact?, ECSA Seventh Biennial International Conference, May 31-June 2, 2001, Madison.
Börzel, T. A., 2002, States and Regions in the European Union. Institutional Adaptation in Germany and Spain.; in series: Themes in European Governance, edited by Føllesdal, A., Olsen, J. P., Cambridge: Cambridge University Press.
Börzel, T. A., 2003, Environmental Leaders and Laggards in Europe. Why there is (not) a 'Southern Problem', Aldershot: Ashgate.
Börzel, T. A., Risse, T., 2000, When Europe Hits Home: Europeanization and Domestic Change, European Integration online Papers (EIoP) 4(15) <http://eiop.or.at/eiop/texte/2000-015a.htm>.
Duina, F. G., 1997, Explaining Legal Implementation in the European Union, International Journal of the Sociology of Law 25(2), 155179.
Duina, F. G., 1999, Harmonizing Europe: Nation-States within the Common Market, Albany: State University of New York Press.
Duina, F. G., Blithe, F., 1999, Nation-States and Common Markets: The Institutional Conditions for Acceptance, Review of International Political Economy 6(4), 494-530.
Falkner, G., 1998, EU Social Policy in the 1990s: Towards a corporatist policy community, London: Routledge.
Falkner, G., 2002, How Intergovernmental are Intergovernmental Conferences? An Example from the Maastricht Treaty Reform, Journal of European Public Policy 9(1), 98-119.
Falkner, G., 2003, Social Policy, in: Cini, M. (Ed.): European Union Politics, Oxford: Oxford University Press, 264 - 277.
Falkner, G., Hartlapp, M., Leiber, S., Treib, O., 2002, Opposition through the Backdoor? The Case of National Non-Compliance with EU Directives. Working Paper, Political Science Series, Vol. 83 (October 2002), Vienna: Institute for Advanced Studies IHS <http://www.ihs.ac.at/publications/pol/pw_83.pdf>.
Falkner, G., Leiber, S., 2003, A Europeanization of Governance Patterns in Smaller European Democracies?, Eighth Biennial International Conference, European Union Studies Association, 2003-03-27/29, Nashville, Tennessee <http://www.mpi-fg-koeln.mpg.de/socialeurope/>.
Falkner, G., Treib, O., Hartlapp, M., Leiber, S., 2004, Complying with Europe? The Impact of EU Minimum Harmonisation and Soft Law in the Member States (forthcoming, for details please visit http://www.mpi-fg-koeln.mpg.de/socialeurope/).
Golub, J. (Ed.) 1998, New Instruments for Environmental Policy in the EU, London: Routledge.
Green Cowles, M., Caporaso, J., Risse, T. (Eds), 2001, Transforming Europe. Europeanization and Domestic Change; in series: Cornell Studies in Political Economy, ed. by Katzenstein, P. J., Ithaca/London: Cornell University Press.
Hartlapp, M., 2004, Zur Durchsetzung von EG-Recht. Eine ländervergleichende Studie über die Umsetzung und Anwendung arbeitsrechtlicher EG-Richtlinien unter besonderer Berücksichtigung von Frankreich, Belgien, Spanien, Portugal und Griechenland, Köln: Max-Planck-Institut für Gesellschaftsforschung (forthcoming in Campus Book Series; see also http://www.mpi-fg-koeln.mpg.de/socialeurope/).
Heinelt, H., Malek, T., Smith, R., Töller, A. E. (Eds), 2001, European Union Environmental Policy and New Forms of Governance. A study of the implementation of the environmental impact assessment directive and the eco-management and audit scheme regulation in three member states, Aldershot et al.: Ashgate.
Héritier, A., 2001, Differential Europe: The European Union Impact on National Policymaking, in: Héritier, A., Kerwer, D., Knill, C., Lehmkuhl, D., Teutsch, M., Douillet, A.-C. (Eds): Differential Europe: The European Union Impact on National Policymaking, Lanham et al.: Rowman & Littlefield, 1-22.
Héritier, A., Kerwer, D., Knill, C., Lehmkuhl, D., Teutsch, M., Douillet, A.-C., 2001, Differential Europe: The European Union Impact on National Policymaking; in series: Governance in Europe, edited by Marks, G., Lanham et al.: Rowman & Littlefield.
Héritier, A., Knill, C., Mingers, S., 1996, Ringing the Changes in Europe - Regulatory Competition and the Transformation of the State: Britain, France and Germany, Berlin: de Gruyter.
Héritier, A., Mingers, S., Knill, C., Becka, M., 1994, Die Veränderung von Staatlichkeit in Europa. Ein regulativer Wettbewerb: Deutschland, Großbritannien, Frankreich, Opladen: Leske + Budrich.
Holzinger, K., 2002, Transnational Common Goods: Regulatory Competition for Environmental Standards, in: Héritier, A. (Ed.): Common Goods. Reinventing European and International Governance, Lanham/Boulder/New York/Oxford: Rowman & Littlefield, 59-84.
Jachtenfuchs, M., Strübel, M. (Eds), 1992, Environmental Policy in Europe: Assessment, Challenges and Perspectives, Baden-Baden: Nomos.
Jordan, A., 2000, The politics of multilevel environmental governance: subsidiarity and environmental policy in the European Union, Environment and Planning 32, 1307-1324.
Kingdon, J. W., 1984, Agendas, Alternatives, and Public Policies, Boston/Toronto: Little, Brown and Company.
Knill, C., 2001, The Europeanisation of National Administrations. Patterns of Institutional Change and Persistence; in series: Themes in European Governance, edited by Føllesdal, A., Olsen, J. P., Cambridge: Cambridge University Press.
Knill, C., Lenschow, A., 1998a, Coping with Europe: The Impact of British and German Administrations on the Implementation of EU Environmental Policy, Journal of European Public Policy 5(4), 595614.
Knill, C., Lenschow, A., 1998b, Coping with Europe: the impact of British and German administrations on the implementation of EU environmental policy, Journal of European Public Policy 5(4), 595-614.
Knill, C., Lenschow, A., 1999, Neue Konzepte alte Probleme? Die institutionellen Grenzen effektiver Implementation, Politische Vierteljahresschrift 40(4), 591617.
Knill, C., Lenschow, A., 2000a, Do New Brooms Really Sweep Cleaner? Implementation of New Instruments in EU Environmental Policy, in: Knill, C., Lenschow, A. (Eds): Implementing EU Environmental Policy. New Directions and Old Problems, Manchester: Manchester University Press, 251-282.
Knill, C., Lenschow, A. (Eds), 2000b, Implementing EU Environmental Policy. New Directions and Old Problems, Manchester: Manchester University Press.
Knill, C., Lenschow, A., 2000c, On Deficient Implementation and Deficient Theories: The Need for an Institutional Perspective in Implementation Research, in: Knill, C., Lenschow, A. (Eds): Implementing EU Environmental Policy. New Directions and Old Problems, Manchester: Manchester University Press, 9-35.
Knill, C., Lenschow, A., 2001a, Adjusting to EU Regulatory Policy: Change and Persistence of Domestic Administrations, in: Risse, T., Cowles, M. G., Caporaso, J. (Eds): Europeanization and Domestic Change, Ithaca NY: Cornell University Press; also published as: conference paper, 1998, pp. 1-25.
Knill, C., Lenschow, A., 2001b, "Seek and Ye Shall Find!" Linking Different Perspectives on Institutional Change, Comparative Political Studies 34(2), 187-215.
Kohler-Koch, B., 2002, European Networks and Ideas: Changing National Policies?, European Integration online Papers 2(6) <http://eiop.or.at/eiop/texte/2002-006a.htm>.
Ladrech, R., 1994, Europeanisation of domestic politics and institutions: the case of France, Journal of Common Market Studies 32(1), 69-88.
Leiber, S., 2004, Nationale Sozialpartnerschaft und Soziales Europa. Eine ländervergleichende Studie zur Umsetzung arbeitsrechtlicher EU-Richtlinien unter besonderer Berücksichtigung von Dänemark, Schweden, Finnland, Österreich, Luxemburg und Italien, Köln: Max-Planck-Institut für Gesellschaftsforschung (forthcoming in Campus Book Series; see also http://www.mpi-fg-koeln.mpg.de/socialeurope).
Lenschow, A., 1999, Transformation in European Environmental Governance, in: Kohler-Koch, B., Eising, R. (Eds): The Transformation of Governance in the European Union, London: Routledge, 39-60.
Lenschow, A., 2002, New Regulatory Approaches in 'Greening' EU Policies, European Law Journal 8(1), 19-37.
Mayntz, R., Scharpf, F. W., 1995, Der Ansatz des akteurzentrierten Institutionalismus, in: Mayntz, R., Scharpf, F. W. (Eds): Gesellschaftliche Selbstregelung und politische Steuerung, Frankfurt/New York: Campus, 39-72.
Mayntz, R., Streeck, W. (Eds), 2003, Die Reformierbarkeit der Demokratie: Innovationen und Blockaden; Festschrift für Fritz W. Scharpf; in series: Schriften des MPIfG, Vol. 45, Frankfurt a.M.: Campus.
Radaelli, C. M., 2000, Whither Europeanization? Concept Stretching and Substantive Change, European Integration online Papers 4(8) <http://eiop.or.at/eiop/texte/2000-008a.htm>.
Scharpf, F. W., 2000, Interaktionsformen. Akteurzentrierter Institutionalismus in der Politikforschung, Opladen: Leske + Budrich.
Schmidt, V. A., 2003, The Boundaries of "Bounded Generalizations": Discourse as the Missing Factor in Actor-Centered Institutionalism, in: Mayntz, R., Streeck, W. (Eds): Die Reformierbarkeit der Demokratie: Innovationen und Blockaden; Festschrift für Fritz W. Scharpf, Frankfurt a.M.: Campus, 318-350.
Treib, O., 2004, Europäische Vorgaben, nationaler Anpassungsbedarf und dessen politische Verarbeitung. Eine Ländervergleichende Studie über die Umsetzung arbeitsrechtlicher EU-Richtlinien unter besonderer Berücksichtigung von Deutschland, den Niederlanden, Großbritannien und Irland, Köln: Max-Planck-Institut für Gesellschaftsforschung (forthcoming in Campus Book Series; see also http://www.mpi-fg-koeln.mpg.de/socialeurope).
(*) Most of this paper is based on joint research with Oliver Treib, Miriam Hartlapp, and Simone Leiber at the Max-Planck-Institute for the Study of Societies, and will be published in a joint book (Falkner/Treib/Hartlapp/Leiber forthcoming). For further details on the project see http://www.mpi-fg-koeln.mpg.de/socialeurope. Thanks to Oliver Treib for various comments and draft tables, and to Tina Steinbeck for research assistance. A longer previous version of this paper has been presented at the ECPR General Conference Marburg, 18-21 September 2003, Section 15 "Europeanisation: Challenges of a New Research Agenda", Panel 15.1 "Europeanisation: Concepts & Methods". Thanks for stimulating comments to the organisers and the participants and, most importantly, to the anonymous EIoP referees!
(1) Only a few path-breaking studies can be mentioned here (see, inter alia, the work by Héritier et al. 1996; Héritier et al. 2001; Duina 1999; Knill 2001; Knill et al. 2000c; Knill et al. 2000b; Green Cowles et al. 2001; Börzel 2002; Börzel 2003).
(2) This term is applied by several authors, see e.g. Börzel and Risse (2000).
(3) In our forthcoming book, each chapter explains the degrees of misfit for our cases in detail, so that our categorisation can be controlled in an inter-subjective manner.
(4) This is often called "institutional misfit" in the literature (see, e.g. Börzel et al. 2000), but since there are so many definitions of the term institution (and many of them are very broad) we prefer a more specific label.
(5) There are many studies on the EU's environmental policy and its implementation (Golub 1998; Héritier et al. 1996; Knill 2001; Knill et al. 1998b; Knill et al. 2001a; Knill et al. 2000c; Lenschow 2002; Lenschow 1999; Jordan 2000; Jachtenfuchs et al. 1992; Holzinger 2002; Heinelt et al. 2001).
(6) These countries apply a far-reaching form of corporatism in their domestic labour-law making, i.e. independent social partner action without any state involvement. They also tried to implement EC Directives in the same issue area according to their traditional pattern that cannot, however, cover the whole workforce (Falkner et al. 2003).
(7) Since the governments must be expected to protect themselves, the social insurance companies, and the enterprises from additional costs wherever feasible (or to defend such costs for either actor if it suits their political ideas), we decided not to focus on the distribution on different actors.
(8) Obviously, one could even have allowed for more potential categories of overall misfit degree. However, we doubt that the kind of data on implementation performance that can be collected in expert interviews and supplementary legal studies can ever reach a quality that allows even further differentiation without deteriorating the reliability of the results if tested in an inter-subjective manner.
(9) All data stem from a joint project with Oliver Treib, Miriam Hartlapp, and Simone Leiber and will be discussed in detail in our forthcoming book.
(10) To give a few examples, Germany had to increase the legal minimum of vacation time to four instead of three weeks and had to cut the reference period for the maximum allowed overtime, under the Working Time Directive. France had to include, under the Part-time Work Directive, workers busy between 32 and 39 hours a week. They had, due to the prior definition of part-time work, been excluded from any protection. Luxembourg had to introduce legislation granting parental leave, for the first time, under the ECs Parental Leave Directive. Under the same act, Austria and Germany had to introduce an individual right to parental leave for fathers (independent of the mothers right, and of the mothers profession).
(11) A further complication is that EC law evolves over time, e.g., via judgements of the ECJ. Since the boundaries between specification of known definitions and re-interpretation of what could be expected to be the law are blurred, one cannot but try to work on the basis of what seems to be the majority consensus on interpretation by the time of adoption. If, for example, one member state interprets one provision differently (which typically results in not having to change domestic laws) but all or most others, possibly including the Commission, rely on another understanding of the rule, we coded this as remaining misfit in the one deviating country.
(12) Note that the regional implementation laws add a large number of cases to the 90 already studied. For practical reasons, we had to rely on the information on the state of implementation in autonomous regions given by federal experts of the relevant country. We could not do a case study for each region with an own implementation law for any group of workers.
(13) If a Directive has two equally important aims, e.g. protecting pregnant workers and allowing for their employability, both must be realised in an at least essentially correct manner.
(14) Should there be any developments that go countercurrents with the goals of the Directive, this must be taken into consideration and deducted from the degree of adaptation accomplished.
(15) Special deadlines, e.g. for the UK in the case of those Directives adopted under the original opt-out, are deducted in our calculation of months of delays. So are special extensions of deadlines granted by the European Commission for intense social partner involvement in domestic decision-making on the implementation.
(16) For example, Tanja Börzel argued that "we have simply no evidence that the European Union suffers from a serious compliance deficit which is claimed by the European Commission and academics alike" (Börzel 2001).
(17) Radaelli's "domains of Europeanization and types of change" (Radaelli 2000) by contrast focus more on domestic structures overarching the domain of specific public policies. His nine elements (see his Figure 1) of potential Europeanization on the public policy level (actors, style, instruments, resources, cognitive structure of policy, paradigms, frames, narratives and policy discourse / legitimacy) are useful in analytical terms but too extensive to be applied in multi-case study. This made us look for a slimmer concept, which is still multi-leveled and differentiated. Héritier et al.'s focus is helpful but not fine grained enough for our empirical material (1994; 1996).
(18) The latter encompasses the ideas and communicative actions that can potentially drive, within institutional interactions, change in perceptions, preferences, and strategies (Schmidt 2003: 32).
(19) As long as they do not represent a fundamental reform of the regime type.
(20) See previous footnote.
(21) "Policy entrepreneurs, people who are willing to invest their resources in pushing their pet proposals or problems, are responsible not only for prompting important people to pay attention, but also for coupling solutions to problems and for coupling both problems and solutions to politics." (Kingdon 1984: 21)
(22) Inter alia, because it would be impossible to conduct a proper discourse analysis of 90 domestic adaptation processes and their context.
(23) This follows the tradition of actor-centered institutionalism (Mayntz et al. 1995; Scharpf 2000; Mayntz et al. 2003) where it is recommended to first look at the institutional level and only where necessary decline the ladder of abstraction towards the individual level.
(24) Admittedly, these aspects are but a few elements in the broader field of Europeanisation studies, and some abstract studies have offered even broader catalogues as suggestions for enquiry (Radaelli 2000) or have tackled in greater detail expectations as to when Europeanisation should be expected (see inter alia the work by Héritier et al., Risse et al. and Börzel mentioned earlier).
(25) For sure, one might even go beyond the level of action and enquire on changes in individual attitudes or knowledge, but this will at best be feasible in very small-n studies. On the difficulties of ideational diffusion in the EU, and its study see e.g. Kohler-Koch (2002).