MPIfG Working Paper 02/11, October 2002

Transforming Social Policy in Europe? The EC's Parental Leave Directive and Misfit in the 15 Member States

Gerda Falkner, Miriam Hartlapp, Simone Leiber, Oliver Treib

Max Planck Institute for the Study of Societies

Abstract

This paper presents first results from a collaborative research project which analyzes the national transposition, enforcement and application of six European labor law Directives in all 15 member states. Looking at the case of the EC's parental leave Directive (1996), it draws conclusions about the domestic impact of European social policy standards.

It will interest practitioners as much as labor law experts that, in fact, adaptational pressure was created in all 15 member states. Although the overall economic impact of the Directive in terms of costs was rather modest, every single country had to change its rules. While misfit was rather small in Finland, France, Germany, Portugal, Spain and Sweden, the other member states were confronted with significant adaptation pressure.

Political theorists may be surprised that our results cast doubts on the theoretical usefulness of focusing too much on matches or mismatches between European policies and domestic structures. We show that a very low degree of misfit may sometimes even be a problem rather than a condition for success and that the existence of considerable adaptational pressure may under certain conditions even be conducive to smooth implementation. In addition, several member states not only eliminated the misfit created by the Directive, but raised their domestic standards above the European minimum requirements.

Zusammenfassung

Dieses Papier stellt erste Ergebnisse eines Projektverbundes vor, der die rechtliche Umsetzung, administrative Durchführung und praktische Anwendung von sechs arbeitsrechtlichen EG-Richtlinien in allen 15 Mitgliedstaaten untersucht. Anhand der Richtlinie zum Elternurlaub (1996) werden hier die Auswirkungen europäischer Sozialstandards auf der nationalen Ebene behandelt.

Für Praktiker und Arbeitsrechtsexperten gleichermaßen interessant dürfte unser Befund sein, dass die Richtlinie tatsächlich in allen Mitgliedstaaten Anpassungen nötig machte. Obwohl die von der Richtlinie verursachten Kosten insgesamt eher gering waren, musste doch jedes Land seine Regulierungen ändern. Während die erforderlichen Reformen in Deutschland, Finnland, Frankreich, Portugal, Spanien und Schweden von eher begrenzter Tragweite waren, sahen sich die übrigen Mitgliedstaaten mit beträchtlichem Anpassungsbedarf konfrontiert.

Theoretisch Interessierte werden feststellen, dass unsere Ergebnisse Zweifel daran erwecken, ob die im Rahmen der Prognose des Anpassungserfolges bislang gängige Orientierung an der Größe des verursachten Anpassungsbedarfes wirklich sinnvoll ist. Wir zeigen, dass sehr kleiner "misfit" manchmal die erfolgreiche Anpassung sogar behindern kann, und dass hoher Anpassungsbedarf unter bestimmten Bedingungen größeren Erfolg verspricht. Darüber hinaus haben mehrere Mitgliedstaaten im Zuge der Implementation nicht nur die notwendigen Anpassungen vorgenommen, sondern ihre nationalen Standards über das von der Richtlinie geforderte Maß hinaus verändert.

Contents

1Introduction [1]

For a long time, European Community studies was dominated by a bottom-up perspective. The debate between neo-functionalists and intergovernmentalists in essence revolved around the question of whether and to what extent nation states were willing to transfer crucial decision-making competencies to the European level. When scholarly attention turned away from "grand bargains" and macro-level developments to an analysis of everyday decision-making, the underlying analytical approach did not change fundamentally. When looking at the interactions between supranational, national, subnational and societal actors in European policy-making, the focus still lay on the relative influence of these actors in bringing about European policy solutions.

It was only recently that scholars developed an interest in the effects of Europeanization on domestic systems of governance. This top-down perspective has produced a number of studies dealing with the impact of membership in the European Communities (now the European Union) on such phenomena as national parliaments, party systems, state-society relationships or territorial state structures (for example, Schmidt 1999; Falkner 2000; Börzel 2001; Mair 2001; Maurer/Wessels 2001; Raunio/Hix 2001). In this context, scholars have also developed a more narrow interest in the domestic impact of European policies, as witnessed by the national implementation of European policy measures (Directives and Regulations). Focusing mainly on environmental policy, a number of scholars have pointed to the degree of fit or misfit between European rules and existing institutional and regulatory traditions as one of the central factors determining implementation performance. While some have stressed the importance of institutional fit or misfit, i.e. the degree of compatibility or incompatibility between European policies and national administrative structures and traditions (Knill 2001; Knill/Lenschow 1999; 2001), others have directed attention to policy fit or misfit, i.e. the match or mismatch between EC measures and domestic policy instruments, standards and problem-solving approaches (Börzel 2000a; 2000b). Still others have included both institutional and policy dimensions to provide a comprehensive understanding of the match and mismatch between European demands and domestic structures or legacies (Héritier et al. 1996; Duina 1997; Duina 1999; Risse et al. 2001).

Notwithstanding these differences in detail, this strand of literature shares the view that the compatibility or incompatibility between a given European policy measure and the pre-existing national traditions in the member states tells us a lot about the likeliness of implementation success or failure of that measure. The approach ultimately rests on historical and/or sociological institutionalist assumptions about the "stickiness" of deeply entrenched national policy traditions and administrative routines, which poses great obstacles to reforms aiming to alter these arrangements (see, for example, March/Olsen 1989; DiMaggio/Powell 1991; Thelen/Steinmo 1992; Immergut 1998; Thelen 1999; Pierson 2000).

Seen from this angle, European policies face deeply rooted institutional and regulatory structures. If both fit together, that is if adaptational pressure is low, implementation should be a smooth and unproblematic process easily accomplished within the given time limits. If European policies do not match existing traditions, however, implementation should be highly contested, leading to considerable delays, and involving a high risk of total failure. The following quotation provides a concise summary of this argument: "It is assumed that implementation problems only occur if there is pressure for adaptation. If an EU policy fits the problem solving approach, policy instruments and policy standards adopted at the national level, there is no reason why the public administration should resist implementation. The EU legislation can be easily absorbed into the existing legal and administrative system. Only if the implementation of an EU policy requires considerable legal and administrative changes imposing economic and political costs on the public administration, implementation failure should be expected" (Börzel 2000a: 225).

It seems that the preoccupation with matches and mismatches between European policies and national structures has led to a certain neglect of other possible factors that might influence the implementation of European policies. In a recent edited book, Thomas Risse, Maria Green Cowles and James Caporaso (2001) have suggested a number of "mediating factors" which may lead to adaptation even in the face of high levels of incompatibility. These factors are: a decision-making structure with a small number of veto points or, alternatively, a consensus-oriented decision-making culture which may be able to avoid stalemate even in systems with multiple veto actors; the presence of supporting institutions; pressures exerted by supportive interest groups; and processes of elite learning. The relevance of most of these factors seems to be highly plausible, which is not only demonstrated by the contributions assembled in the book (Green Cowles et al. 2001) but also by other empirical studies that point to the importance of similar causal conditions (Héritier et al. 2001). But do these factors really come into play only in cases of high adaptational pressure? What about the areas of low and medium levels of misfit which so far have been almost [2] totally neglected? And what other properties of national systems might act to further or hinder timely and proper implementation?

It is the aim of this paper to improve our understanding of implementation processes in the EU context. It presents results from a collaborative research project which analyzes, from a comparative perspective, the national transposition, enforcement and application of six European labor law Directives in all 15 member states. The degree of changes required and the actual transposition of European standards in the 15 domestic systems is discussed here, taking the example of the 1996 parental leave Directive. On this basis, the paper draws conclusions about the domestic impact of European social policy standards: Is the parental leave agreement of relevance only because it is the first EU-level deal between the European peak associations of labor and capital, and not as a piece of social legislation that matters in terms of social convergence in Europe (as held by earlier commentators)? [3] Furthermore, our paper discusses the relevance of the misfit hypothesis (as specified in our operationalization [4]) and tries to outline a number of additional causal conditions which have a bearing on implementation performance.

The paper is organized as follows. First, we provide an overview of the process that led to the adoption of the parental leave Directive at the European level. Second, each of the country case studies is summarized. Finally, we discuss the insights into theoretical as well as empirical aspects derived from our study of the parental leave Directive's implementation.

2The Directive and its Background

The first Commission proposal for a Directive on parental leave and leave for family reasons dates back as early as 1983 (COM [83] 686 final). On the basis of the argument that the quite diverse national provisions were thought to hamper the harmonious development of the Common Market, an approximation on the basis of Article 100 EEC Treaty was suggested. The minimum standards suggested were: three months of parental leave for either parent (to be taken up to the third birthday of the child), and an unspecified number of days off for family reasons to be decided by the individual member state. With regard to social insurance and pay, leave for family reasons was to be treated as time off with pay. By contrast, pay or indemnity for parental leave was only an option, to be met by public funds. The Commission advocated an unequivocal non-transferability of these rights. However, because of (mainly) British opposition and the unanimity requirement, adoption was impossible. Finally, the draft was set aside for almost a decade.

In 1993, the Belgian Council Presidency "tried to give social Europe back its wings" (Belgian social minister Smet, quoted in Agence Europe November 25, 1993: 9). One of its initiatives consisted in drafting a new compromise proposal on parental leave. During the Social Council's November session, the UK reportedly tried in vain to obtain a derogation from the Directive, then restated its opposition. [5] Fruitless negotiations continued until autumn 1994. Despite consensus among eleven delegations in the last relevant Council debate on September 22, 1994, deliberation was still not possible due to a British veto (Ministerrat 1994; Hornung-Draus 1995).

This was the ideal situation for an application of the Maastricht Social Agreement, which by then had already been in force for almost a year. It excluded the UK from the social policy measures adopted by the other (then) eleven member states and allowed for the adoption of Euro-collective agreements between the major interest groups on social issues that could be implemented by the EC Council Directives (for details on the Social Agreement, see Falkner 1998). Hence consultation of labor and management on the issue of "reconciliation of professional and family life" was instigated by the Commission on February 22, 1995. The Commission's consultation document outlined the importance of an initiative in the light of men and women's changing employment roles. In addition to national legal provisions, the Commission noted advances made on a voluntary basis by the social partners in practically all of the member states and therefore voiced the belief that the social partners should explore the widest possible range of issues relating to reconciliation. Seventeen interest groups reportedly responded within the six-week deadline of this first consultation procedure. According to Commission sources (COM[96] 26 final, explanatory memorandum), the answers revealed support for the promotion of equal opportunities as well as a consensus in favor of Community action on the subject, at least in the form of a recommendation. The general opinion was said to have been that the social partners should play an active role in drawing up the fundamental principles and then in implementing these through collective bargaining.

In fact, however, negotiation mandates were already being drawn up by the three major cross-sectoral federations (UNICE, [6]CEEP [7]and ETUC [8] ) while the first round of consultations was still taking place. This indicates that they were keen to show that the Euro-corporatist procedures of the Maastricht Treaty could actually be put into practice. On June 21, 1995, the Commission launched the second round of consultations. Already some two weeks after the start of the consultation, on July 5, 1995, CEEP, ETUC, and UNICE expressed their "wish to prove that they are capable of reaching binding agreements in the framework of negotiations" (Agence Europe July 13, 1995: 15; cf. European Industrial Relations Review - EIRR 259: 30).

The collective negotiations were successfully concluded after only five (out of a possible nine) months, on November 6, 1995. A draft framework agreement was submitted to the respective decision-making bodies of the three institutions (Agence Europe November 8, 1995: 15 and November 11, 1995: 12). With a view to the implementation, the ETUC, UNICE and CEEP requested the Commission to submit their framework agreement to the Council for a decision that would make the requirements binding in the member states of the Union with the exception of the UK. Soon after the formal signature of the agreement on December 14, 1995, the Commission accordingly proposed a draft Directive to the Council (on January 31, 1996; cf. Agence Europe February 1, 1996: 7). Reportedly, the draft was a matter of controversy in the Social Affairs Council (cf. Agence Europe March 29, 1996: 8). For some delegations, the content of the framework agreement left too much space for interpretation, making proper application in the member states a difficult task. Others thought that the social partners had neglected powers of the EC institutions by introducing a non-regression clause and a time limit for implementation. Nevertheless, a political consensus was reached on March 29, [9] and the Directive was formally adopted without debate on June 3, 1996. [10] The Directive had to be incorporated into national law by June 3, 1998, with the possibility of a maximum additional period of one year "if this is necessary to take account of special difficulties or implementation by a collective agreement" (Article 2 of the Directive). [11]

The parental leave agreement endows both male and female workers with an individual right [12] to parental leave on the grounds of the birth or adoption of a child. This is to enable them to take care of that child for at least three months. At the end of the leave period, the workers have the right to return to the same job, or, if this is not possible, to an "equivalent or similar job consistent with their employment contract or employment relationship" (Clause 2.5). Acquired rights have to be maintained as they stand until the end of the parental leave and to apply again thereafter. The general considerations of the text reveal that parental leave is seen as an important means of reconciling professional and family responsibilities and promoting equal opportunities and treatment between men and women. The "EC social partners" state that men should be encouraged to assume an equal share of family responsibilities, for example by means of awareness programs. However, it is crucial to stress that the agreement sets out only a few minimum compulsory standards but leaves it to member states and national social partners to establish the access conditions and modalities for applying the right to parental leave and leave for urgent family reasons.

All the more thrilling is the question concerning the practical effects of the Directive in the 15 member states, which have such different systems of labor law and gender relations.

3Implementing the Directive: Diverse Reactions to Common European Standards

Austria

Austria already had a longstanding tradition of parental leave before it became a member of the European Union. For example, the opportunity for the father to take parental leave had been introduced in 1990. Since then, either parent could take full-time leave until the child reached the age of 2, or part-time leave until the child reached the age of 4. [13] During the period of leave, private sector employees who met certain eligibility requirements received a benefit from the unemployment insurance fund of approximately 12.35 euro per day; public servants were entitled to approximately 25 percent of the wage paid in step 2 of Service Category V (International Labour Office 1994: 290). In addition, adoptive parents had the same rights to parental leave. Despite this rather well-developed system, the implementation of the parental leave Directive has caused further amendments to the pre-existing legislation. [14]

First, the rights of adoptive parents have been extended. The previous rules have caused problems regarding the adoption of children who were older than 1 year and 9 months. For those children, the minimum requirement of the Directive of 3 months parental leave would not have been guaranteed as the leave in Austria was only possible up to the child's second birthday. Therefore, the rules now allow leave up to six months even if the child at the time of adoption is already between 18 months and 7 years old.

The second important issue has been the introduction of parental leave as an individual right. Before the Directive was implemented, the father of the child could only take leave of absence if the mother was entitled to the leave, but did not take it. This meant that he could only care for the child if the mother left the leave to him - there was no equal treatment of mother and father. In addition, if the mother was, say, a student or a housewife, she had no right to take parental leave and therefore the father was not entitled to take parental leave either. This had to be changed in response to the Directive. It is interesting that, during the negotiations between the European social partners, the Austrian and German trade union representatives pressed for this individual right to be included in the European regulations and finally succeeded with their demands. In this way, Austrian trade unions have used the multi-level system to push through a reform which at the national level might have met with more resistance.

Assessing the practical effect of the required adaptations, the progress made with respect to adoptive parents has proved useful, but as to the individual right it must be said that the take-up rates for fathers have not been affected very much in practice (Interview A2: 1488-1499). It is still very unlikely that the father will take parental leave, and even more so if the mother is at home anyway. The changes taken as a whole cannot be considered very costly, either. Politically, however, the introduction of the individual right may be judged as a remarkable qualitative step towards equality between men and women.

Austria was late in transposing the Directive. It made use of the option to extend the implementation period by one year, but nonetheless the amendments did not come into force until January 1, 2000. The question arises as to why it was necessary to extend the transposition period and why Austria still did not manage to transpose in time despite the fact that the degree of misfit was rather low. This may be explained with the help of the voluntary adaptations and additional reforms carried out in the same reform process. Trade unions especially used the transposition of the Directive as an occasion to press for more far-reaching reforms to further the advantages for their members. Together with the adaptation to the binding standards of the Directive, they succeeded, for instance, in introducing more flexible notice periods for taking leave, which was favorable to employees. In addition, the possibility was introduced to postpone parts of the leave until the child is 7 years old. The interview partners clearly attributed the incentive for this flexibilization of leave options to the recommendations of the Directive (Interview A7: 379-389). Therefore, the implementation process - although at first sight the misfit was not very great - has caused a broader discussion between the government and the social partners and so has required time.

To sum up, the adaptations required by the Directive were important in a sociopolitical respect, but rather small in their practical effects. Nevertheless, transposition was delayed, which was mainly due to discussions between the government and the social partners about voluntary reforms connected to the transposition process. [15]

Denmark, Finland and Sweden

In the Scandinavian counties the main adaptations required by the Directive concerned so-called leave for urgent family reasons (Clause 3 of the framework agreement on parental leave). In Denmark and Sweden no further adaptations were required at all. In Finland, only one additional slight change was necessary with regard to an employee's return to work. In the pre-existing legislation in Finland, the person on leave had the right to return to the same or to an equivalent job. The Directive, however, prescribes that the person has to come back to the same job and only if this is not possible may he or she be forced to accept an equivalent one. This had to be adapted as well (Interview FIN3: 361-489). As far as parental leave is concerned, well-developed systems which went beyond the Directive's requirements were already in place in all three countries. There had even been pre-existing rules as regards leave for urgent family reasons (e.g. in case of child sickness), but these were not as comprehensive as the Directive prescribed. Therefore, minor adaptations were necessary in all three countries.

Finland and Sweden transposed the missing rules correctly and in time. Besides these compulsory adaptations there were no further changes associated with the transposition processes. Even the recommended rules of the Directive already pre-existed to a certain extent. [16] In addition, it seemed that, against the background of already very well-established national systems, the national governments did not see the need for any reforms which went beyond the Directive's compulsory requirements. In sum, these cases do fit well with the common picture that small adaptations can be dealt with by the member states without major transposition problems.

The Danish case, though at first sight very similar, depicts a remarkable particularity as regards the transposition instrument. In Denmark, the parental leave Directive was transposed via the collective bargaining system, because leave for urgent family reasons belongs to the areas regulated in Denmark via autonomous social partner interaction without government interference. The social partners in Denmark are eager to preserve their regulatory autonomy with regard to working conditions. Therefore the rules governing urgent family leave were not transposed via legislation, but by way of sectoral collective agreements. In addition, the main social partner organizations concluded a national agreement which intended to ensure that employees not covered by sectoral collective agreements should nevertheless be granted the rights demanded by the Directive (Clauwaert/Harger 2000: 25). While the European Treaty explicitly allows member states to transpose European Directives by way of collective agreements, [17] the European Court of Justice has defined strict criteria to judge whether such implementation is in conformity with European law. In particular, collective agreements have to ensure that all workers covered by a particular European Directive are entitled to the rights conferred by the Directive. [18] There are other cases where the European Commission has raised concerns that the coverage guaranteed by collective agreements was not sufficient and where Denmark finally decided to depart from its tradition and introduced subsidiary legislation. [19] In this case, the dialogue between the Danish government and the Commission is still going on, but the odds are that it will end up with the same result - the need to introduce supplementary legislation. If the European Commission does not accept the way Denmark has implemented the Directive, this case will underline that problems are very likely to occur if European Directives interfere with national regulatory traditions, and especially with the autonomy of social partners.

Luxembourg

Before the European Directive had to be implemented, there was no parental leave in Luxembourg in a narrow sense. Since 1988, employees could take a career break and receive an allowance during that time. But these regulations did not include a right to return to the same or a similar job, which is a very important difference. If the employee applied for re-employment within a year after maternity leave, the employer was obliged to give priority to the application, provided that there was a vacancy and the employee on leave was qualified for it. Yet still a parent using this career break always had to run the risk of not being re-employed. Only civil servants were already able to take parental leave - though unpaid - for a maximum of two years after maternity leave (International Labour Office 1994: 360-361).

These regulations have been retained in Luxembourg even after the introduction of the new parental leave rules, since the new parental leave scheme was introduced in addition to the pre-existing rules. Now families can choose which regulation they prefer: either the new kind of parental leave or the former scheme of career breaks. Therefore the introduction of the parental leave can be regarded as a novelty without pre-conditions, which is to say that the misfit for Luxembourg on this issue was considerable. In addition, the pre-existing rules regarding leave for urgent family reasons had also to be extended. Previously only some collective agreements had covered this area. As a reaction to the Directive, new legislation was introduced (Feyereisen 1999). On top of that, Luxembourg did not simply introduce the minimum requirements of the Directive, but went far beyond that. The introduction of the leave was seen (and justified) as a particular employment measure with the aim of temporarily replacing the persons on leave by unemployed people. Therefore, the leave period is longer than the Directive prescribes - 6 instead of 3 months - and in addition people are encouraged to make use of the leave, because it is very well paid, with employees receiving approximately 1487 euro per month (Interview LUX11: 570-634).

Although the degree of misfit was considerable, Luxembourg managed to transpose the Directive within the extended transposition period. The new parental leave act came into force in March 1999. [20] This is even more surprising as Luxembourg very often is among the countries which transpose with significant delay. [21] This is mainly due to a combination of administrative overload and political priorities. Luxembourg is a small country with limited administrative capacities. Therefore, there are many problems encountered in coping with the demands from Brussels on time. Under these conditions, the administrative action is guided by the imperative of "what is most important is to be done first." As a consequence, European measures pertaining to areas in which pre-existing national rules already exist, and which therefore necessitate rather small adaptations, are treated with less priority than European Directives requiring the introduction of completely new rules - like the parental leave Directive (Interview LUX1: 1000-1034). It could then be argued that in the Luxembourg case transposition was in time because the degree of changes required by the Directive was rather high.

Italy

Before the European Directive was implemented, regulations in Italy had provided for so-called optional childcare leave of six months, which could be taken after compulsory maternity leave. In principle, the mother was entitled to take this leave, but the father could also make use of it in place of the mother or if he had sole parental custody. [22] Adoptive parents had the same rights as natural parents. The leave was also paid through a benefit of approximately 30 percent of the employee's previous salary. These entitlements, however, were only guaranteed for the mother or, in exceptional cases, to the father with sole custody (International Labour Office 1994: 355). Nor was the right to return to a similar job secured for fathers (Interview I10: 75-203). Hence opportunities to take parental leave existed, even for fathers, but under rather difficult circumstances. The interview partners' perception was that "the general idea of parental leave was absent from Italian legislation" (Interview I5: 142-14). Especially since it was de facto only a prolonged maternity leave, there was no equal role for the father. Therefore, the changes induced by the transposition of the Directive can be considered as an important qualitative step. The new legislation [23] secured the individual right and equal treatment of the father as regards the return to the job. In addition, the transposition process was used to voluntarily extend the optional leave period from 6 to 10 months, and the government even introduced a bonus of one month if the event of the father taking at least three months' leave in one block. The practical consequences of these measures, however, will probably not be drastic - at least in the short or medium term. Although it is too early to assess the practical effects of the new law, it is not very likely that Italian families will be quick to change their traditional model of family responsibilities. But at least the possibilities for shared family responsibilities have been facilitated.

The transposition of the parental leave Directive in Italy was remarkably delayed. The new law has only been in force since March 2000. In November 1999, an infringement case was even referred to the European Court of Justice because of non-transposition of the Directive. But it was finally dropped in June 2000 after the adoption of law No. 53/2000 (Commission of the European Communities 2001: 31). The introduction of the new law caused fierce conflicts between management and labor. In particular, there was strong opposition from the employers' side (Interview I6: 294-327; Interview I9: 114-134). Interestingly, these conflicts were not so much focused on the compulsory adaptations, though these in turn have not been small. They were rather caused by the voluntary over-implementation and the fact that the transposition was "part of a broader law, which regards in general the relation between working time and social life and especially focuses on sharing the familial responsibility between man and woman" (Interview I5: 398-474).

In sum, there was remarkable misfit, and transposition was considerably late (until an ECJ court action intervened), but the misfit alone was not the crucial point. The transposition was connected to a broader reform which at the time was a controversial issue in the Italian political process.

Germany

The overall level of protection guaranteed by the pre-existing German Parental Leave Act ("Bundeserziehungsgeldgesetz") [24] was already well above the requirements of the European parental leave Directive. Under the act, parents were entitled to three years' parental leave during which a certain amount of allowance was paid by the state. Nevertheless, the law did not conform fully to the Directive. The right to take parental leave did not apply to employees whose partners were staying at home because they were not employed. This system was therefore inconsistent with the Directive, which guaranteed an individual right to parental leave for all employees irrespective of the employment status of their partners.

The practical implications of this inconsistency between the European Directive and the German legislation appear to be limited, since male employees (who will typically have been affected by the clause because their wives tend to stay at home in order to take care of the household and the raising of children) traditionally have accounted for only two percent of all employees taking parental leave (Vascovics/Rost 1999: 42). Since a single-income family will not be able to make a living if the sole breadwinner takes parental leave and thus receives only state parental leave allowance, which amounts to a maximum of 460 euro per month, it is highly unlikely that a large number of newly entitled employees will make use of their right. This explains why both trade unionists and employers' representatives considered the degree of changes required by the Directive to be low (Interview D3: 785-793; Interview D9: 111-156).

Despite the fact that the overall impact of the changes required by the Directive was rather slight, the Directive was transposed no sooner than two and a half years after the expiry of the implementation deadline. This seemingly surprising pattern may be explained by the political unwillingness of the conservative-liberal German government, which was in office until 1998, to adapt the German legislation to the requirements of the European Directive. The exclusion of single-income couples from the right to take parental leave was a logical consequence of the male-breadwinner-oriented family model underlying the existing Parental Leave Act (Interview D1: 30-72), which had been created in 1985, under the aegis of the conservative-liberal Kohl government. According to that logic, if one of the parents, typically the mother, was staying at home anyway, there was no need for the father to take parental leave (Interview D3: 561-566). The Directive, with its emphasis on the needs of working parents and the reconciliation of work and family life, had a different starting point, and was therefore inconsistent with the traditional German approach, even though this inconsistency manifested itself only in a point of marginal practical relevance.

The conservative-liberal government was totally opposed to adapting the German parental leave legislation to the Directive, since the latter was at odds with its conservative preferences with regard to family policy (Interview D9: 143-150). The government had already argued at the time the Directive was adopted at the European level that it was not willing to change the German law in response to the Directive (Ministerrat 1996). In 1998, when the European Commission issued a letter of formal notice, asking why Germany had not yet notified it of any transposition measures (Bulletin EU 7/8-1998: section 1.8.1), German officials again answered that there was no need for a change to the German legislation (Interview D3: 538-635). It is interesting to note that, despite the insufficiency of existing German legislation, the Commission seems to have been content with this reply. At any rate, no further infringement steps have been taken against Germany with regard to the parental leave Directive.

Following the adoption of the parental leave Directive in 1996, the Social Democrats, who were part of the opposition at that time, introduced a parliamentary motion (Bundestag 1996) in which they argued that the need to transpose the parental leave Directive should be used to enact a fundamental reform of the existing legislation, thereby transforming the male-breadwinner-oriented parental leave scheme into a more gender-neutral one. The motion, however, was voted down by the government coalition (Bundestag 1998: 22308). Only after the new "Red-Green" government coalition between the Social Democrats and the Green Party had assumed power in October 1998 was transposition of the Directive accomplished by a fundamental overhaul of the Parental Leave Act, which was adopted on July 7, 2000 and came into force on January 1, 2001. [25] The reform was based on the proposals contained in the 1996 parliamentary motion. Thus, the revised legislation not only extended the right to take parental leave to single-income couples, as necessitated by the Directive, but even went far beyond the minimum requirements of the Directive, thus reflecting some of the non-binding recommendations laid down in the European measure. For example, the reform introduced a legal right to work part-time during parental leave and allowed parents to take parental leave simultaneously.

In sum, the German case shows that the implementation of measures with little economic impact may give rise to considerable delays in transposition if they call for qualitative changes to an existing regulatory system which are at odds with the peculiar political agenda of the government of the day. Partisan composition of government also seems to be crucial for the fate of soft-law provisions enshrined in European Directives. In the German case, some of these non-binding recommendations were indeed taken into account, but only because they went well with the political aspirations of the incoming left-wing government.

The Netherlands

In the Netherlands, it was not until 1990 that a statutory parental leave scheme was introduced at all. Moreover, the 1990 act [26] provided for comparatively modest rights. Employees were entitled to thirteen weeks (or three months) unpaid parental leave, which had to be taken on a part-time basis. Thus, employees could not take full-time leave to take care of a child, but were only entitled to reduce their weekly working time to 20 hours. The length of the leave was determined by the previous working time. It amounted to 13 times the hours of a normal working week. Thus, employees whose normal working time per week had been 40 hours were entitled to reduce their hours to 20 hours for 26 weeks. While it would not have been contrary to the Directive to provide for part-time leave only, the obvious downside of the pre-existing Dutch parental leave scheme was that employees whose weekly working time was below 20 hours were not entitled to parental leave at all (Interview NL4: 60-76). Such an exclusion of part-time workers was not in line with the Directive, because Clause 1 of the framework agreement stipulated: "This agreement applies to all workers, men and women, who have an employment contract or employment relationship."

The need to extend coverage of the parental leave scheme to part-time workers would seem to be an important step, especially in the Netherlands, which has the highest proportion of part-time workers in Europe. In 1995, about 35 percent of all Dutch employees were working on a part-time basis (Visser 1999: 286). [27] Thus, the degree of changes required by the Directive appears to be considerable. The actual significance of these changes, however, is greatly reduced by the fact that a national review process of the 1990 act was already under way when the Directive was passed in 1996, and the reform proposals issued by the government as a result of that review already provided for an extension of the parental leave scheme to all employees (Clauwaert/Harger 2000: 68). Hence, the implementation of the Directive did not pose any major problems for the Dutch government. The revised parental leave legislation was passed on June 25, 1997, and came into operation on July 1, 1997. [28] While the soft-law provisions in the Directive in general did not play an important role, trade unions successfully lobbied for the incorporation of the Directive's recommendation that employees should be able to take parental leave until the child's eighth birthday. Previously, the age limit had been four years, and the initial government bill had provided for six years. Under union pressure, the bill was amended during the parliamentary process in order to raise the age limit to eight years (Interview NL4: 315-326; Interview NL10: 176-179).

The implementation of the parental leave Directive in the Netherlands demonstrates that an analysis of the "fit" or "misfit" between European requirements and pre-existing national policies has to take into account the possibility that the existing regulatory system might be a moving target, with ongoing national reform processes pushing a country's initially mismatching regulations towards the standards set at the European level.

United Kingdom

In the UK, the parental leave Directive necessitated the introduction of completely new legislation. Prior to the implementation of the parental leave Directive, employees did not have any legal entitlement to parental leave. In addition, this lack of statutory provision was not counterbalanced to a significant degree by equivalent entitlements provided by collective agreements or company practice. Thus, the UK government estimated that only about a quarter of all eligible employees already had some comparable entitlement to leave provided by their employer (DTI 1999: 5-6). The actual costs to employers of introducing these new rights were tempered by the fact that the Directive did not require member states to provide any payment during parental leave. Hence, employers mainly had to cope with the costs resulting from temporarily replacing parental leavers.

Altogether, the government estimated that the introduction of the new parental leave scheme would burden business with additional annual costs of 42 million pounds or around 69 million euro (DTI 1999: 2), which is relatively low compared to the effects in the UK of the European working time Directive or the introduction of a national minimum wage. [29] Despite the modest impact in quantitative terms, the imposition of new legal rights for employees was a considerable qualitative innovation, given the UK's liberal, intervention-free tradition in employment regulation, which was reinforced by the deregulation policy of the Thatcher and Major governments (Edwards et al. 1999: 4-15).

In light of the significant adaptational pressure, the way the Directive was implemented in the UK conforms only partly to the conventional misfit-centered view of implementation. Unlike its conservative predecessors who had vetoed the draft Directive in the Council of Ministers on several occasions during the European decision-making process (see above) and who had chosen to opt out of the European Treaty's social chapter, the Labour government which had assumed power in May 1997 was firmly committed to implementing the Directive. Blair's more positive stance towards European social regulations was made clear when he ended the British opt-out and agreed to implement the Directives adopted without the UK. As a result of the government's supportive attitude, the Directive could be incorporated into UK law within the implementation deadline. [30] This timely transposition, however, was greatly facilitated by the extremely small number of veto points which, under the British "Westminster model", would be able to stop a government from realizing its will (Tsebelis 1995; Lijphart 1999: 9-21). What is more, the virtual lack of any veto points outside of government itself is complemented by a very far-reaching reliance on delegated legislation, at least in the implementation of European Directives in the form of Regulations (Drewry 1995: 457). Although parliamentary influence is marginal in both cases, the process of enacting Regulations is much faster than the one Acts of Parliament have to go through (Interview GB7: 356-371, see also Butt Philipp/Baron 1988: 649-650).

The way in which the Directive was implemented, and the conflicts surrounding the process, are more amenable to a misfit-oriented view. As a result of strong employer pressure, but also because of its own policy of avoiding placing unnecessary burdens on business, the government chose a very minimalist route to implementation. Thus, almost all available exemption and derogation options were used, and most soft-law provisions were disregarded. What is more, the government introduced a "cut-off date" which excluded parents whose children were born before the date on which the implementation legislation came into effect (Regulation 13 of the Maternity and Parental Leave Regulations etc. 1999).

This cut-off date, which was in response to demands from employers' organizations, especially those representing small and medium-sized enterprises (Interview GB2: 184-195, 531-552), was against the terms of the Directive, and Britain's national trade union confederation, the TUC, therefore brought a case against the government to the High Court in London. This case received considerable media attention because the TUC was legally represented in court by Cherie Blair, the Prime Minister's wife, who was herself about to have a baby at the time of the Court hearing (Interview GB6: 270-277). The case subsequently was referred to the European Court of Justice for a preliminary ruling. However, the British judges had already argued that the case was likely to prevail in the European Court of Justice (Hall 2000). A few days before the case would have been heard before the ECJ, the government agreed to remedy the matter and the TUC withdrew its challenge (TUC 2001). In the meantime, the legislation has been amended so as to repeal the cut-off date. [31]

In sum, the UK's implementation of the parental leave Directive partly confirms the conventional view that high adaptational pressure leads to implementation problems. What is overlooked by such an interpretation, however, is the favorable stance of the government toward implementing the Directive and its ability to push through the largest part of the required legislative reforms within the given deadline. In addition, the importance of pressure groups acting as watchdogs and the role of national and European court powers in enforcing European standards are highlighted by the case.

Ireland

Implementation of the parental leave Directive in Ireland in almost all respects parallels the British story. In Ireland, as in the UK, the Directive necessitated the introduction of completely new legislation. Prior to the Directive, no statutory parental leave scheme had existed. Thus, the degree of adaptation required was "100 percent", as one interviewee aptly put it (Interview IRL1: 215). Similarly to the UK, the actual costs of implementing the Directive were not excessive, but still significant, with employers mainly having to cope with replacement costs, while the total burden on business was moderated by the lack of any requirement to provide for payment during parental leave (Interview IRL6: 288-302). The introduction of a statutory right for mothers and fathers to take parental leave was nevertheless a huge step for Ireland given the voluntarist tradition of Irish employment regulation, resulting in a generally low level of labor market regulation (Prondzynski/Richards 1994: 10; Prondzynski 1999: 56).

Like its British counterpart, the Irish government was firmly committed to implementing the European Directive. This time, however, it was not party politics that mattered, because the government supported the Directive even though it was made up of a conservative-liberal coalition between Fianna Fáil and the Liberal Democrats. However, the implementation of the parental leave Directive and a number of other "family-friendly" measures had been agreed between the social partners and the government in the tripartite "Partnership 2000" agreement (Interview IRL6: 312-318). Since national social partnership, which has considerably intensified during the last 15 years, is backed by all major political parties in Ireland (O'Kelly 2000: 351), it comes as no surprise that the FF-LD government fully supported the transposition of the Directive. As a consequence, implementing the Directive was altogether a rather smooth process and could be accomplished without exceeding the deadline. All that was needed was a six months' extension of the transposition period, which, in accordance with Article 2 of the Directive, was granted by the Commission (Interview IRL6: 243-249, 616-625). [32] This is all the more remarkable since deficient administrative capacities in Ireland often lead to delays in implementation.

In substantive terms, the Irish government chose a minimalist way of transposing the Directive, making use of most of the available exemptions and ignoring almost all the soft-law provisions. The only instance of over-implementation was on the issue of leave for urgent family reasons. While the Directive only stipulates that employees must be granted a few days off work without payment to care for family members who have suddenly become ill or have had an accident (Clause 3 of the framework agreement), the Irish legislation requires employers to maintain remuneration during leave for urgent family reasons, which, however, is restricted to three days per year (Section 13 of the Parental Leave Act, 1998). In striking similarity to the UK government, the Irish government decided to introduce a cut-off date excluding parents whose children were born before June 3, 1996, i.e. the date on which the Directive was adopted at the European level. As in the UK, this was a reaction to the pressure from employers to minimize the initial costs of adaptation to the Directive. It was only after the European Commission had initiated infringement proceedings that the Irish government agreed to repeal the cut-off date and to extend the legislation to all parents (Interview IRL6: 232-258). [33]

The Irish case again demonstrates that it is perfectly possible for member states to implement Directives which confront them with considerable adaptational pressure in a surprisingly fast and altogether fairly smooth way. However, sometimes concessions seem necessary to accommodate strong business opposition. Thus, both the Irish and the UK governments initially gave in to business pressure through the introduction of a cut-off date which excluded certain parents from parental leave entitlements. They both waited for external intervention by the European Commission or the European Court of Justice to "force" them into compliance with European standards.

Belgium

In Belgium parental leave did not exist in the private sector, whereas it had been known for a long time in the public sector (Interview B6: 82-137; Interview EU1: 293-319). Only a 10-day leave for urgent reasons such as hospitalization, accident or sickness of persons living under the same roof was regulated by the convention collective de travail N° 45 (CCT, national collective agreement). [34] Therefore in legal terms misfit was almost 100 percent (for private employees), or as one interview partner put it "we had to start at zero" (Interview B1: 570-570, translation by M.H.).

The degree of practical misfit was clearly more moderate. Since the mid-1980s, collective agreements often ensured parental leave (Interview B6: 82-137) and, de facto, parental leave was taken under the so-called scheme d'interruption de carrière (IDC, sabbatical). This scheme had been introduced on January 22, 1985, at the suggestion of Mrs Miet Smet, who later became Minister for Labor and Social Affaires when the EU Directive was adopted and transposed into Belgian law. Under the general IDC, all workers in the private sector can stop working or reduce their working hours for three to twelve months while receiving part of their monthly salary and being entitled to return to their job afterwards. However, the right to take such an interruption de carrière depends on the employer's agreement - unless there is a more favorable collective agreement in the enterprise - and requires replacement by an unemployed person. [35] Thus to comply with the provisions of the EU Directive important changes were necessary.

The required adaptation has taken place in time. [36] In Belgium now, all workers have the individual right to take three months' leave to educate their natural or adopted children under four years of age. During the leave period, they are protected against dismissal and their acquired rights are maintained. Following the soft-law aspects of the Directive, the right to leave is non-transferable and the leave may be taken not only on a full-time basis, but also part-time or in a piecemeal form. It could be argued that some of the standards - such as protection against dismissal or the right to return to a similar job - are not new in a qualitative sense since they have been practiced under the general IDC leave before. In addition, it should be noted that the reform mainly meant a change of instruments, since parental leave de facto was often assured by collective agreements even before the transposition by law.

Thus, Belgium seems to be an example where correct and timely transposition took place, even though misfit was considerable. [37] This is even more remarkable since in the 1980s, when the European Commission had first proposed a Directive on parental leave, the Christian democratic-liberal (CVP-PRL) coalition government in Belgium blocked the proposal along with the UK. In 1993, however, it was the Socialist Belgian Minister Miet Smet who, during the presidency of her country, reopened discussion of the long-dormant parental leave Directive and became a promoter of the issue: "We took the proposal out of the fridge, we wanted to open the debate again and we put a lot of work into it" (Interview B6: 38-47, translation by M.H., see also above). But even if support for the introduction of parental leave in Belgium is obvious, which is remarkable in light of the significant misfit that needed to be overcome, the story is more complicated. The transposition of the parental leave Directive led to severe quarrels between national actors over the question of who should be the one to implement the Directive.

After the adoption of the European Directive, Minister Smet envisaged transposition by adding a special sub-scheme devoted to parental leave to the IDC. At the same time, the social partners in the National Labor Council had started negotiating on the transposition of the EU Directive and adopted CCT N° 64 on April 29, 1997. However, they could not assure payment or continuity of social security benefits since this was outside their remit. That the social partners still went ahead with the transposition of the Directive by CCT even against the will of the government has to be seen in the context of the fact that the Directive was the first to stem from European-level social partner negotiations. It is important to note that the Belgian government supported the European social partner negotiations, because of the form of the negotiations. Social partner agreements with erga omnes application, giving social partners a quasi-legislative function, are well known in Belgium. Therefore, Belgians were proud to export the national model to the European level (Interview B5: 265-280). Minister Smet was for personal and competence reasons attached to a transposition via a sub-scheme of the IDC. [38] "The minister did not want to introduce a right to take parental leave, but to integrate it into the existing sabbatical scheme, albeit with allowance, and the social partners wanted to create a new right, but they could not provide for allowance without the agreement of the minister" (Interview B9: 296-325, translation by M.H.).

The result is a confusing coexistence of two versions of parental leave. Even though they are similar, some relevant differences exist: the CCT guarantees a right for parental leave even in SMEs, but financing is not assured. Parental leave under the IDC scheme is financed, but can only be taken if the person has not taken parental leave under the CCT 64. Owing to these two parallel schemes, the legal situation remains unclear, and debates about the prevalence of one of the schemes continue. It should be noted that transposition via one of the schemes would have been sufficient. The quarrels are not due to the Directive, but can nevertheless only be explained in its context.

France

In France almost no misfit existed at the time of adoption of the European Directive in Brussels. The general principle of parental leave financed by a state allowance (APE, allocation parental d'éducation) had for a long time been established in the Code du Travail (L 122-28). Some national standards even provided for a significantly higher protection level. For example, parental leave could be extended to a total of 36 months. This was partly due to the fact that two years before the Directive was adopted at the European level a new national law on the family was adopted. [39] Taking into account the fact that at this point the discussion in Brussels had already been under way for some years, it is not surprising that some of the newly introduced changes covered standards which the Directive would otherwise have required to be adopted. Most importantly, the right to parental leave was extended to SMEs with less than 100 workers. Previously, parental leave in these enterprises had depended upon the decision of the employer after consultation with the comité d'entreprise (works council). Parental leave also became an individual right in 1995.

This picture fits in well with the broad support found amongst national actors, ranging from the militant trade union federation Force Ouvrière (Interview F3: 206-208) and public sector unions (Interview F9: 146-178) to employers' representatives (Interview F2: 125-233) and ministry officials (Interview F7: 145-289). The underlying tenor in all interviews was that even employer representatives seem convinced that European social policy is good because "the European Directives do not disturb the French game, they do not pose any problems because we already have our Labor Code" (Interview F2: 125-233, translation by M.H.).

As to leave for urgent family reasons, a statutory entitlement for all employees to time off with pay for family reasons such as marriage, birth or adoption, or death existed at the time the Directive was adopted, but this did not include sickness (EIRR 263: 24). In addition, since 1995 exceptional leave has been allowed in cases of accident or for the care of sick children under 16 - but this scheme does not include care for older children or other family members. Even if it is difficult to argue that this change was caused exclusively by the ongoing European debate over the parental leave Directive, interference seems likely. Hence, this reform may well have been enacted in anticipation of the coming European measure. In any case, misfit remained with regard to leave for reasons like caring for older children or other family members, or in cases of sickness. Initially, however, these shortcomings were not eliminated by a reform of the existing schemes - inertia prevailed. Change did not take place until later when it was in the interest of the Jospin government to further improve the situation for employees who wanted to take leave for urgent family reasons. The remaining misfit with the European Directive will also have played its role. At the end of 2000, new leave for seriously ill children was established. [40] This leave is paid and thus the new law guarantees a more favorable regulation than prior national standards, and even exceeds the requirements of the Directive. But there is still no provision for parents to take leave in order to care for a sick partner or relative, and thus some of the requirements of the Directive still have not been met. So far the Commission has not taken any steps to correct this, maybe because the formulation of the standard in the EU Directive is very loose and diverging interpretations are possible according to the national point of view.

The non-binding provisions did not play any role in the French transposition of the Directive (Interview F7: 313-331). In the first few years, no transposition took place at all. This inertia seems to be due to the fact that there was only very limited misfit in France. At the time the Directive was adopted, adaptational pressure was only small, partly because changes had been anticipated in an earlier national reform. Under such circumstances it would certainly have been inconvenient for government to risk triggering a new debate and become involved in possible conflicts just to change some marginal details of the law and possibly include some of the soft-law provisions.

Spain

In Spain two different types of parental leave existed before the Directive was adopted in Brussels: 12 months' parental leave to take care of children with the possibility of extending this (excedencia por cuidado de hijo) and an entitlement to working time reduction for guardianship (guarda legal) - which is de facto an equivalent to part-time parental leave. Protection against dismissal was provided, but only in constitutional form as a general clause against any discriminatory dismissal (Interview E4: 632-840). The Spanish regulation on leave for urgent family reasons included cases of birth, serious illness and death of relatives up to second degree of consanguinity and kinship, without any age restriction. Even if Spanish law showed a broad understanding of family, accidents were not considered a reason for taking urgent leave (Castro Argüelles 1998: 913). Last but not least, it should be mentioned that, in cases where both parents worked, Article 46.3 of the Estatuto de los Trabajadores (ET) granted parental leave to the father and to the mother, but simultaneous leave was not possible. However, the principle of parental leave as an individual right was not contradicted by this regulation.

There was consent amongst interview partners from the ministry, employers' organizations and trade unions that the degree of misfit at the time of adoption was small (Interview E4: 618-623; Interview E2: 831-835; Interview E1: 1087-1099). Thus existing national rules were considered largely sufficient to comply with the European Directive (Interview E4: 632-840). National actors were reluctant to take action over the remaining details - especially since misfit seemed contingent on interpretation. This explains the initial inertia.

In the context of the transposition of the EC pregnant workers Directive, [41] the European Commission had made it clear that the constitutional granting of protection against dismissal was not sufficient to comply with the requirements of European standards. Therefore, Spain saw itself forced to revise the legislation on maternity leave so as to introduce explicit protection against dismissal (Commission of the European Communities 1999: 14). The conservative government used this opportunity to clarify simultaneously the existing uncertainties over protection against dismissal during parental leave and the list of grounds on which leave for urgent family reasons could be taken. The individual right clause was also revised and now definitely ensures the standard of the EU Directive (Interview E4: 632-840). [42] Once the stimulus for a wider reform had been given, transposition of the parental leave Directive was quickly incorporated into a broader reform law, and adaptation finally took place after a delay of a year and a half.

Non-binding recommendations seem in Spain to be generally considered "extras" for which trade union representatives have to fight in collective agreements (Interview E1: 1225-1233, 1201-1205). Non-compulsory standards of the EU parental leave Directive did not enter into statutory law. Despite intensive lobbying by trade unions to extend the law (Law 39/99) to include the possibility of taking parental leave until the child is 8 years old, the age limit remained 6 years (Interview E1: 1225-1233; Interview E8: 288-291).

Greece

An entitlement to three months unpaid parental leave existed in Greece since 1984 (Law 1483/1984), [43] but this law limited employees' leave for childcare to enterprises with more than 100 workers. In 1993, a national collective agreement had lowered the threshold number of employees to 50 and extended the duration to 3.5 months. Thus, in June 1996, when the Directive was adopted at EU level, the right to parental leave was still restricted to workers in enterprises with more than 50 employees (Koukiadis 1997: 387; Clauwaert/Harger 2000:45). Since the first introduction of parental leave, the right to take leave also applied to cases of adoption, and protection against dismissal was guaranteed. The right to return to a similar job existed (Interview GR2: 232-364) and acquired rights were maintained (Koukiadis 1997: 388). However, the right to take parental leave was only granted to workers whose spouses worked outside the family household, similar to the rules in Germany and Austria (International Labour Office 1994: 333; Koukiadis 1997: 387-388). Such an exclusion of certain employees from the parental leave entitlement was not in line with the Directive.

The scope of parental leave rights was enlarged by Law 2639/98 [44] and the company-size threshold was also abandoned. Now all workers in both public and private-sector enterprises are entitled to parental leave. Thus implementation was secured in time even if this broader labor law is not considered by national experts to be a transposition of the EU Directive. The Commission has initiated the first step of an infringement proceeding for non-notification, but apparently no further action has been taken.

It seems that misfit remains since parental leave is still not an individual right and provision has been made for the employer to refuse parental leave to his or her employees if more than 8 percent of the workforce ask for leave in the same year. Applications are dealt with on a "first come, first served" basis. This is not related to the size of the enterprise and it is not specified how long the leave may be postponed (Interview GR14). In combination with the child's age limit up to which the leave can be taken, this can ultimately lead to a de facto refusal by the employer (International Labour Office 1994:333; Koukiadis 1997: 387; Clauwaert/Harger 2000:45). So far, however, no practical difficulties have been reported.

In sum, misfit was considerable in legal terms and this has great relevance in practice, since there are many SME enterprises in Greece which were not covered before. With regard to the principles, most standards had already been secured for the smaller field of application. Since the government did not make any adjustments to the substantive standards guaranteed by law, it is no surprise that the recommendations of the Directive have not found their way into Greek legislation. Pressure from national actors to improve the conditions under which leave can be taken would have been needed to initiate a reform process. Such pressure is absent since societal interest in these conditions remains low as long as take-up numbers are low - and this is mainly due to a lack of adequate financial support (Interview GR2: 232-364).

Portugal

Prior to the implementation of the Directive, parental leave existed as an employee's right to take 6 months "special leave" for children under three years. Prolongation of this unpaid leave was possible up to 24 months (Interview P1: 1727-1745, Wollmann 1999: 89). Leave on the grounds of adoption was also possible if the adopted child was under three years old. But this special leave did not guarantee all the rights required by the European Directive. Even if both mothers and fathers could take the leave, interview partners claimed that the right was not individual (Interview P8: 770-815, see also Pinto et al. 1997: 578). In addition, it seems that no explicit protection against dismissal existed for parental leave, nor was the return to a similar job legally assured. [45] Regarding leave for urgent family reasons, there was no misfit since an entitlement to a maximum of 15 days' leave in order take care of any family member in cases of sickness or accident was guaranteed, and if the child was younger, the leave was even longer (Interview P1: 1788-1851).

Misfit was in overall terms medium and small in practical relevance, [46] but the changes that took place with the transposition law No. 142/99 [47] are even more considerable since they exceed the binding standards of the European Directive. National experts argued that "the Directive imposed almost no modifications to the existing law, and we only modified it in order to make the social partner agreement visible" (Interview P1: 2162-2169, translation by M.H.).

A so-called parental-leave scheme has recently been introduced. Without replacing the special leave, it simply adds the right to another type of leave. Parents can now be absent from work for three months to take care of their children. Once this parental leave has been used, parents can still take the "old" special childcare leave for another two years. At first sight, this coexistence seems to be a confusing duplication of provision instead of proper integration. On closer inspection, however, the coexistence might be explained with reference to the non-regression clause, which does not allow a lowering of prior standards when transposing European Directives. The "new" parental leave is an individual and non-transferable right but only for three months (Clauwaert/Harger 2000:75, Interview P8: 44-56). The "old" right allows for more leave, but not as an individual and non-transferable right. Portugal did not introduce the new benefits for the special leave under the old regulation, but limited them to a shorter time period required by the Directive. Nevertheless if the "old" right would have been cancelled, the standard of six months' leave would have been lowered to three months, something not allowed when transposing EU Directives.

Adaptation was delayed by a year and a half. [48] The 17th Report on the Application of Community Law reveals that there was an infringement procedure of the Commission against Portugal (1998/0441) which was suspended on December 22, 1999, shortly after the transposition legislation entered into force. National adaptation did not only successfully tackle the misfit, but also showed itself to be creative in using the Directive to address specific national problems, such as teenage pregnancy, by the introduction of grandparents' leave (Interview P8: 770-815). In addition, the age limit for adopted children was raised, and the new law fosters a role change to improve female participation on the labor market by introducing 15 days' paternity leave with pay. Thus, Portugal is a case where stalemate would have been likely according to the misfit hypothesis, while the actual Portuguese reaction shows that, even with relevant misfit, proper implementation and even over-implementation is possible. In this context, it should be noted that law No. 142/99 along with other amendments to law No. 4/84 were part of the 1996-1999 Acordo de Concertação Estratégica, a strategic, tripartite social pact. Thus consensus amongst national actors on problem definition and solutions existed.

4Conclusions: EC Standards and Misfit in the Member States

The above analysis of the EC's 1996 parental leave Directive and its implementation in the 15 member states sheds new light on the discussion of EC social standards (see 1 below) and of implementation quality determinants (see 2 below).

1. By the time of its adoption, many have argued that, the importance of the parental leave agreement lay in its existence rather than its substance. The symbolic importance of a first collective agreement at the EU level to prove that the corporatist procedures of the Maastricht Social Agreement could be operational was indeed great. At the same time, many scholars criticized the fact that the agreed minimum standards were low (for example, see Keller/Sörries 1997; 1999). Specialist journals such as the European Industrial Relations Review (EIRR) highlighted above all that the minimum three months of parental leave represented the shortest time allotted in any of the countries with a statutory right to parental leave, i.e. in Greece. One had the impression that legal changes would only be required in a very small number of countries. Concerning urgent family leave on grounds of force majeure, the parental leave agreement's provisions were initially thought to improve the status quo only in Ireland and the UK (the latter originally being outside its remit) (EIRR 263: 23).

An in-depth-evaluation, by contrast, must take into consideration the relevant background of the parental leave agreement at both the European and the national levels. First, the realistic point of comparison for the provisions of the agreement (which was incorporated in the Council Directive) is with the original Commission proposal (which itself had not recommended particularly far-reaching standards) and with the various compromise texts that had been discussed in the Social Council briefly before the Social Agreement was employed as a new legal basis (see also Dølvik 1997; Falkner 1998; Hartenberger 2001). The latter were even lower in standard than the collective agreement.

Second, and most importantly here, a study of the practical effects in the member states is needed if we are to properly assess the impact of the agreement. The empirical study summarized in this paper reveals that, in fact, adaptational pressure was created in all 15 member states. Although the overall economic impact of the Directive in terms of the costs to (private and public) employers was rather modest (in particular, compared to Directives in other policy areas [49] ), every single country had to change its rules. While policy misfit was rather small in Finland, France, Germany, Portugal, Spain and Sweden, the other member states needed significant adaptation of either the qualitative kind (turning parental leave into an individual right of every employee, male and female alike, without any general need for the employer's consent) or the quantitative kind (e.g. including important categories of the workforce, such as employees working in SMEs or private-sector employees, in the existing leave schemes). Denmark is a particular case, since the implementation of the Directive only required limited adaptation in policy terms, but with high probability will call for the introduction of generally binding legislation instead of autonomous regulation by the social partners, thereby challenging the established relationship between the state and the social partners. Even in the absence of a final assessment of the parental leave Directive's standards, [50] one should underline the fact that the number of changes actually required certainly surpasses scholarly expectations.

Furthermore, almost half of all the member states went beyond the European minimum requirements when transposing the Directive. Minor (but still significant) cases of over-implementation could be observed in France, Ireland, and the Netherlands, while the reforms enacted in Austria, Germany, Italy, Portugal, and Luxembourg were considerably more favorable than the Directive envisaged.

2. What theoretical lessons can be drawn from the way the 15 member states have implemented the parental leave Directive? On the one hand, some of our cases conform well to the view suggested by misfit-oriented scholars. For example, Finland and Sweden both had to cope with very little adaptational pressure and did not encounter any problems in fine-tuning their existing schemes so as to fully comply with the requirements of the Directive.

On the other hand, however, several cases are completely at odds with the misfit hypothesis. In Spain and France, only very little change was required. Nevertheless, transposition came considerably late. In both countries, it was the small amount of misfit itself that caused problems. Spanish officials originally argued that no changes were required at all, and only acknowledged the existence of adaptational pressure after the Commission had taken action against Spain's rules against dismissal in the context of the pregnant workers Directive. Similarly, the French administration for a long time considered its existing rules and regulations as basically sufficient and therefore saw no point in initiating a time-consuming legislative reform process only to effect some detailed fine-tuning without any real impact. The converse is true for Luxembourg. Here, the existence of considerable adaptational pressure had a positive effect on implementation performance. While Luxembourg is often plagued by considerable delays in transposing European Directives, the government in this case managed to adjust in an unusually fast and smooth way. Under conditions of permanent administrative overload, Directives requiring more important changes are apt to be treated with higher priority than measures which demand only minor changes. In addition, the Luxembourg and Irish cases stress the importance of administrative capabilities and resources for implementation performance. Although the relevance of this causal condition was already acknowledged in the first comprehensive study on implementing European Community legislation (Siedentopf/Ziller 1988a; 1988b), the insights gained by this pioneer analysis do not receive the attention they should in recent debates.

Additionally, we have found a number of more ambiguous transposition patterns, displaying affirmative as well as conflicting features. The UK and Ireland were both forced to introduce completely new legislation and thus had to cope with rather high degrees of adaptational pressure. At first sight, the incorrect transposition of the Directive, which could only be remedied after interventions by the Commission and the European Court, seems to match the predictions of the misfit-oriented view. It is nevertheless surprising that both governments were firmly committed to implementing the Directive (whose standards they would certainly not have chosen autonomously for their country in the absence of an EU Directive) and managed to transpose the central parts within the given deadline. Both cases underline the importance of political willingness on the part of governments, but they also stress the relevance of interest group pressure. Such pressure may not only be beneficial to implementation, but may also force governments to indulge in non-compliant behavior. In the two cases in question, both transpired: union pressure helped speedy transposition while employer opposition made the British and Irish governments try (in vain) to get away with an illegal cut-off date.

The role in timely implementation of specific government agendas, and in particular of the partisan composition of government, is also underlined by the German case. The conservative-liberal government refused to revise the existing legislation since the changes were at odds with the government's conservative preferences on family policy as expressed in the laws in force. Adaptation only took place after the Social Democrats had assumed power and chose to fundamentally overhaul the old legislation according to their more egalitarian views on family policy. The German case also demonstrates that the partisan composition of government is important for the fate of soft-law provisions. Some of the non-binding recommendations were indeed taken into account by the new act because they were amenable to the political aspirations of the left-wing government.

A further pattern concealed by an exclusive focus on policy misfit is the role played by issue linkage and over-implementation or gold-plating (i.e. implementation exceeding the Directive's binding standards). Thus, the Austrian case shows that implementation delays may also be due to debates about the way soft-law provisions should be reflected in the national rules. In Italy, transposition was also late because the implementation of the Directive was coupled with a comprehensive review of the legislation on maternity leave and of other issues related to greater compatibility of work and family life.

To sum up, the empirical results presented in this paper cast doubts on the theoretical usefulness of focusing too much on matches or mismatches between European policies and domestic structures. We have shown that a very low degree of misfit may sometimes even be a problem rather than a condition for success and that a considerable misfit may under certain conditions even be conducive to smooth implementation. In addition, several member states not only eliminated the misfit created by the Directive, but raised their domestic standards above the European minimum requirements. While the degree of misfit indeed may be an important condition for determining implementation success or failure, our analysis suggests that a more open theoretical approach which takes into account both institutional and actor-based factors, such as the framework of actor-centered institutionalism (Mayntz/Scharpf 1995; see also Scharpf 1997), may be a more promising starting point. Adding to the insights gained by scholars like Héritier et al. (2001) and Risse, Green Cowles and Caporaso (2001), we have demonstrated that national context variables are crucial for an understanding of implementation processes, and we have highlighted the impact of interest group pressure, administrative capabilities, party politics, issue linkage, and gold-plating.

Further research will be needed to determine the relative weight of each of these factors and to complement this picture with implementation studies on many more Directives. [51] What is certain at this stage is that our analysis reveals a much more dynamic relationship between European rules and domestic adaptation strategies than has been suggested by misfit-oriented scholars hitherto.

References

Börzel, Tanja A., 2000a: Improving Compliance through Domestic Mobilisation? New Instruments and Effectiveness of Implementation in Spain. In: Christoph Knill/Andrea Lenschow (eds.), Implementing EU Environmental Policy. New Directions and Old Problems. Manchester: Manchester University Press, 222-250.

Börzel, Tanja A., 2000b: Why There Is No "Southern Problem": On Environmental Leaders and Laggards in the European Union. In: Journal of European Public Policy 7(1), 141-162.

Börzel, Tanja A., 2001: Europeanization and Territorial Institutional Change: Toward Cooperative Regionalism? In: Maria Green Cowles/James Caporaso/Thomas Risse (eds.), Transforming Europe: Europeanization and Domestic Change. Ithaca: Cornell University Press, 137-158.

Bundestag, 1996: Antrag der Abgeordneten Hildegard Wester, Christel Hanewinckel, Hermann Bachmaier, weiterer Abgeordneter und der Fraktion der SPD: Elterngeld und Elternurlaub für Mütter und Väter. BT-Drucksache 13/6557. Bonn: Deutscher Bundestag.

Bundestag, 1998: Plenarprotokoll 13/241 vom 18.06.1998. Bonn: Deutscher Bundestag.

Butt Philipp, Alan/Christina Baron, 1988: United Kingdom. In: Heinrich Siedentopf/Jacques Ziller (eds.), Making European Policies Work: The Implementation of Community Legislation in the Member States. Vol. 2: National Reports. London: Sage, 637-713.

Castro Argüelles, M. Antonia, 1998: La incidencia de las Directivas communitarias en la reforma laboral. In: CIVITAS - Revista Espanola de Derecho del Trabajo 92, 889-917.

Clauwaert, Stefan/Sabine Harger, 2000: Analysis of the Implementation of the Parental Leave Directive in the EU Member States. Brussels: European Trade Union Institute.

Commission of the European Communities, 1999: Report from the Commission on the Implementation of Council Directive 92/85/EEC of 19 October 1992 on the Introduction of Measures to Encourage Improvements in the Health and Safety at Work of Pregnant Workers, and Workers Who Have Recently Given Birth or Are Breastfeeding. COM(99) 100 final. Brussels: Commission of the European Communities
<http://europa.eu.int/comm/employment_social/equ_opp/news/preg-en.pdf> (15.03.1999).

Commission of the European Communities, 2001: Eighteenth Annual Report on Monitoring the Application of Community Law (2000). COM(2001) 309. Brussels: Commission of the European Communities (16.07.2001).

DiMaggio, Paul J./Walter W. Powell, 1991: Introduction. In: Walter W. Powell/Paul J. DiMaggio (eds.), The New Institutionalism in Organizational Analysis. Chicago: University of Chicago Press, 1-38.

Dølvik, Jon Erik, 1997: Redrawing Boundaries of Solidarity? ETUC, Social Dialogue and the Europeanisation of Trade Unions in the 1990s. Oslo: ARENA.

Drewry, Gavin, 1995: The Case of the United Kingdom. In: Spyros A. Pappas (ed.) National Administrative Procedures for the Preparation and Implementation of Community Decisions. Maastricht: European Institute of Public Administration, 453-476.

DTI, 1998a: Draft National Minimum Wage Regulations: Public Consultation. London: Department of Trade and Industry 
<http://www.dti.gov.uk/er/nmwr/index.htm> (September 1998).

DTI, 1998b: Measures to Implement Provisions of the EC Directives on the Organisation of Working Time ("The Working Time Directive") and the Protection of Young People at Work ("The Young Workers Directive"). London: Department of Trade and Industry (April 1998).

DTI, 1999: Parental Leave Regulations: Summary of Regulatory Impact Assessment - Final Version Lodged with Parliament. London: Department of Trade and Industry (November 1999).

Duina, Francesco G., 1997: Explaining Legal Implementation in the European Union. In: International Journal of the Sociology of Law 25(2), 155-179.

Duina, Francesco G., 1999: Harmonizing Europe: Nation-States within the Common Market. Albany: State University of New York Press.

Edwards, Paul/Mark Hall/Richard Hyman/Paul Marginson/Keith Sisson/Jeremy Waddington/David Winchesteret al., 1999: Great Britain: From Partial Collectivism to Neo-Liberalism to Where? In: Anthony Ferner/Richard Hyman (eds.), Changing Industrial Relations in Europe. Oxford: Blackwell, 1-54.

Falkner, Gerda, 1998: EU Social Policy in the 1990s: Towards a corporatist policy community, European Public Policy Series. London: Routledge.

Falkner, Gerda, 2000: How Pervasive Are Euro-Politics? Effects of EU Membership on a New Member State. In: Journal of Common Market Studies 38(2), 223-250.

Feyereisen, Marc, 1999: Family Leave Introduced in Luxembourg. EIROnline Document LU9903196N. Dublin: European Foundation for the Improvement of Living and Working Conditions
<http://www.eiro.eurofound.ie/1999/03/InBrief/LU9903196N.html>.

Green Cowles, Maria/James Caporaso/Thomas Risse (eds.), 2001: Transforming Europe: Europeanization and Domestic Change. Ithaca: Cornell University Press.

Hall, Mark, 2000: Parental Leave Regulations Referred to ECJ. EIROnline Document UK0006176N. Dublin: European Foundation for the Improvement of Living and Working Conditions
<http://www.eiro.eurofound.ie/2000/06/InBrief/UK0006176N.html>.

Hardy, Stephen/Nick Adnett, 2002: The Parental Leave Directive: Towards a "Family-Friendly" Social Europe. In: European Journal of Industrial Relations 8(2), 157-172.

Hartenberger, Ute, 2001: Europäischer sozialer Dialog nach Maastricht. EU-Sozialpartner-verhandlungen auf dem Prüfstand, Nomos Universitätsschriften. Politik. Vol. 117. Baden-Baden: Nomos.

Héritier, Adrienne/Dieter Kerwer/Christoph Knill/DirkLehmkuhl/Michael Teutsch/Anne-Cécile Douillet (eds.), 2001: Differential Europe: The European Union Impact on National Policymaking. Lanham: Rowman and Littlefield.

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

Hornung-Draus, Renate, 1995: Elternurlaub: Erste Sozialpartner-Verhandlung auf der Grundlage des Sozialabkommens. In: EuroAS 3(9), 147-148.

Immergut, Ellen M., 1998: The Theoretical Core of the New Institutionalism. In: Politics and Society 26(1), 5-34.

International Labour Office, 1994: Maternity and Work, Conditions of Work Digest, 13. Geneva: International Labour Organization.

Jørgensen, Carsten, 2001: Social Partners Agree to Complete Implementation of EU Working Time Directive through Subsidiary Legislation. EIROnline Document DK0112158F. Dublin: European Foundation for the Improvement of Living and Working Conditions
<http://www.eiro.eurofound.ie/2001/12/Feature/DK0112158F.html>.

Keller, Berndt/Bernd Sörries, 1997: The new social dialogue: procedural structuring, first results and perspectives. In: Industrial Relations Journal, European Annual Review 1997, 77-98.

Keller, Berndt/Bernd Sörries, 1999: The New European Social Dialogue: Old Wine in New Bottles? In: Journal of European Social Policy 9, 111-125.

Knill, Christoph, 2001: The Europeanisation of National Administrations: Patterns of Institutional Change and Persistence. Cambridge: Cambridge University Press.

Knill, Christoph/Andrea Lenschow, 1999: Neue Konzepte - alte Probleme? Die institutionellen Grenzen effektiver Implementation. In: Politische Vierteljahresschrift 40(4), 591-617.

Knill, Christoph/Andrea Lenschow, 2001: Adjusting to EU Environmental Policy: Change and Persistence of Domestic Administrations. In: Maria Green Cowles/James Caporaso/Thomas Risse (eds.), Transforming Europe: Europeanization and Domestic Change. Ithaca: Cornell University Press, 116-136.

Koukiadis, I., 1997: Greece. In: Roger Blanpain (ed.) Legal and Contractual Limitations on Working Time in the European Union. Leuven: Peeters Press, 367-400.

Lijphart, Arend, 1999: Patterns of Democracy: Government Forms and Performance in Thirty-Six Countries. New Haven: Yale University Press.

Mair, Peter, 2001: Backbenchers Learn to Fight Back: European Integration and Parliamentary Government. In: Klaus H. Goetz/Simon Hix (eds.), Europeanised Politics? European Integration and National Political Systems. London: Frank Cass, 27-51.

March, James G./Johan P. Olsen, 1989: Rediscovering Institutions. The Organizational Basis of Politics. New York: Free Press.

Maurer, Andreas/Wolfgang Wessels (eds.), 2001: National Parliaments on Their Ways to Europe: Losers or Latecomers? Baden-Baden: Nomos.

Mayntz, Renate/Fritz W. Scharpf, 1995: Der Ansatz des akteurzentrierten Institutionalismus. In: Renate Mayntz/Fritz W. Scharpf (eds.), Gesellschaftliche Selbstregelung und politische Steuerung. Frankfurt/M.: Campus, 39-72.

Ministerrat, 1994: Vermerk des Vorsitzes für den Rat (Arbeit und Sozialfragen) (Tagung am 22. September 1994). Betr.: Geänderter Vorschlag für eine Richtlinie über Elternurlaub. Brüssel: Rat der Europäischen Gemeinschaften (13.09.1994).

Ministerrat, 1996: Entwurf eines Protokolls über die 1914. Tagung des Rates (Arbeit und Sozialfragen) am 29. März 1996 in Brüssel. Brüssel: Rat der Europäischen Union (23.04.1996).

O'Kelly, Kevin, 2000: Ireland: Shifting from Voluntarism to National Agreements. In: Jeremy Waddington/Reiner Hoffmann (eds.), Trade Unions in Europe: Facing Challenges and Searching for Solutions. Brussels: European Trade Union Institute, 339-364.

Pierson, Paul, 2000: Increasing Returns, Path Dependence, and the Study of Politics. In: American Political Science Review 94(2), 251-267.

Pinto, M./P. Furtado Martins/A. Boavida, 1997: Portugal. In: Roger Blanpain/E. Köhler/J. Rojot (eds.), Legal and Contractual Limitations to Working Time in the European Union. Leuven: Peeters Press, 555-602.

Prondzynski, Ferdinand von, 1999: Ireland: Corporatism Revived. In: Anthony Ferner/Richard Hyman (eds.), Changing Industrial Relations in Europe. Oxford: Blackwell, 55-73.

Prondzynski, Ferdinand von/Wendy Richards, 1994: Ireland, European Employment and Industrial Relations Glossary. London: Sweet and Maxwell.

Raunio, Tapio/Simon Hix, 2001: Backbenchers Learn to Fight Back: European Integration and Parliamentary Government. In: Klaus H. Goetz/Simon Hix (eds.), Europeanised Politics? European Integration and National Political Systems. London: Frank Cass, 142-168.

Risse, Thomas/Maria Green Cowles/James Caporaso, 2001: Europeanization and Domestic Change: Introduction. In: Maria Green Cowles/James Caporaso/Thomas Risse (eds.), Transforming Europe: Europeanization and Domestic Change. Ithaca: Cornell University Press, 1-20.

Scharpf, Fritz W., 1997: Games Real Actors Play. Actor Centered Institutionalism in Policy Research. Boulder: Westview.

Schmidt, Vivien A., 1999: National Patterns of Governance under Siege: The Impact of European Integration. In: Reiner Eising/Beate Kohler-Koch (eds.), The Transformation of Governance in the European Union. London: Routledge, 155-172.

Siedentopf, Heinrich/Jacques Ziller (eds.), 1988a: Making European Policies Work: The Implementation of Community Legislation in the Member States. Vol. 1: Comparative Syntheses. London: Sage.

Siedentopf, Heinrich/Jacques Ziller (eds.), 1988b: Making European Policies Work: The Implementation of Community Legislation in the Member States. Vol. 2: National Reports. London: Sage.

Thelen, Kathleen, 1999: Historical Institutionalism in Comparative Politics. In: Annual Review of Political Science 2(1), 369-404.

Thelen, Kathleen/Sven Steinmo, 1992: Historical Institutionalism in Comparative Politics. In: Sven Steinmo/Kathleen Thelen/Frank Longstreth (eds.), Structuring Politics: Historical Institutionalism in Comparative Analysis. Cambridge: Cambridge University Press, 1-32.

Tsebelis, George, 1995: Decision Making in Political Systems: Veto Players in Presidentialism, Parliamentarism, Multicameralism and Multipartism. In: British Journal of Political Science 25, 289-325.

TUC, 2001: Government in "Total Climb Down" over Parental Leave. Press Release. London: Trades Union Congress 
<http://www.tuc.org.uk/work_life/tuc-3092-f0.cfm> (30.04.2001).

Vascovics, Laszlo A./Horst Rost, 1999: Väter und Erziehungsurlaub. Schriftenreihe des Bundesministeriums für Familie, Senioren, Frauen und Jugend, Bd. 179. Stuttgart: Kohlhammer.

Visser, Jelle, 1999: The Netherlands: The Return of Responsive Corporatism. In: Anthony Ferner/Richard Hyman (eds.), Changing Industrial Relations in Europe. Oxford: Blackwell, 283-314.

Wollmann, Ines, 1999: Arbeitsrecht in Portugal. München: Rehm.

Notes

1

We would like to thank Matthias Mahlmann, Pavel Ovseiko, and Ingeborg Tömmel for their very helpful comments.

2

The major exception is Héritier et al. (2001).

3

The few publications on the effects of the parental leave Directive that exist so far either focused solely on some member states (Hartenberger 2001: 139-151) or provided no more than a rough assessment of its practical implications, based on secondary sources (Hardy/Adnett 2002). Thus, our study is the first to offer comparable findings from first-hand empirical investigations in all 15 member states.

4

In an attempt to systematize the different understandings suggested in the literature (see above), we break down the concept of fit or misfit between European requirements and domestic structures into four dimensions. First and foremost, we look at policy misfit, i.e. the extent to which existing policies have to be reformed in order to fulfill the European standards, taking into account both legal changes and their de facto relevance (which may be diminished by pre-existing collective agreements or workplace practices). Second, implementing a Directive may require administrative reforms, such as the creation of new agencies or the assignment of additional tasks or resources to existing administrative entities. Third, implementation can demand changes to the relationship between the state and interest groups, e.g. by necessitating state intervention in an area which traditionally is regulated autonomously by the social partners. Finally, the economic cost implications of these changes, both for the state and for private employers, have to be taken into account. In order to classify different degrees of adaptational pressure and compare them between the member states, we use a threefold scheme (low, medium, and high degrees of misfit).

5
At one point, a lowest-common-denominator solution seems to have emerged: the UK wished parental leave to be only granted to mothers, not to fathers. Reportedly, only the Irish delegation and the Commission were immediately against this "awful" change (as one Commission official described it in an interview), which made the Commission threaten to bring in the ECJ against this discrimination on grounds of sex.

6
The Union of Industrial and Employers' Confederations of Europe.

7
The European Centre of Enterprises with Public Participation.

8
The European Trade Union Confederation.

9
There was unanimous agreement. Adoption was, however, postponed with a view to parliamentary approval in Germany (Agence Europe March 30, 1996: 7).

10
Council Directive 96/34/EC of 3 June 1996 on the framework agreement on parental leave concluded by UNICE, CEEP and ETUC (OJ L 145, June 19, 1996, 4-9).

11
Since the conservative British government had secured an opt-out from the European Treaty's social chapter at the Maastricht summit (EIRR 216: 2), the UK was initially not covered by the Directive. Tony Blair's Labour government, which had assumed power in May 1997, signed up to the social chapter and declared its willingness to implement the Directives that had been enacted during the UK's opt-out (EIRR, 282: 2; EIRR 284: 2). As a consequence, the UK's transposition deadline was later than the one applying to the other member states. The UK had to comply with the Directive by December 15, 1999.

12
This means that a situation where the mother must refrain from her right to parental leave in order to allow the father to take it, as previously practiced in several countries, is not consistent with the agreement.

13
Maternity Protection Act, dated April 17, 1979, Bundesgesetzblatt, No. 221/1979, as amended up to version Bundesgesetzblatt, No. 833/1992; Act concerning parental leave, dated September 9, 1989, Bundesgesetzblatt, No. 651/1989, as amended up to version Bundesgesetzblatt, No. 833/1992.

14
Maternity Protection Act, dated April 17, 1979, Bundesgesetzblatt, No. 221/1979, as amended up to version Bundesgesetzblatt I, No. 153/1999; Act concerning parental leave, dated September 9, 1989, Bundesgesetzblatt, No. 651/1989, as amended up to version Bundesgesetzblatt I, No. 153/1999.

15
The correctness of the transposition measures enacted in 1999 has recently been questioned by the European Commission. The Commission claims that the new regulation concerning the individual right still includes a very slight advantage for the mother vis-à-vis the father (Interview A2b). If this turns out to be true, our argument will be strengthened, for then the transposition - despite the small misfit - will not have only been delayed, but also incorrect.

16
For example the possibility of taking at least parts of the leave until the child is 8 or the option to take leave on a part-time basis.

17
See Article 137 (ex Article 118) of the Treaty establishing the European Community.

18
See, for example, judgment of the Court of 30 January 1985. Commission of the European Communities v Kingdom of Denmark. Equal pay for men and women. Case 143/83.

19
The best-known example is the transposition of the Directive (Council Directive 93/104/EC of 23 November 1993 concerning certain aspects of the organization of working time, OJ L 307, December 13, 1993, 18-24) in Denmark (EIRR 315: 4, Jørgensen 2001).

20
Act concerning the creation of a parental leave and a leave for family reasons, dated February 12, 1999, Mémorial Series A, 1999, p. 189.

21
From the six labor law Directives our project is studying, this is the only Directive which was transposed in time by Luxembourg.

22
Act No. 1204 concerning the protection of working mothers, dated December 30, 1971, Gazzetta Ufficiale, No. 14/1972.

23
Act No. 53 concerning dispositions for the support of maternity and paternity, the right to care and education and the co-ordination of city times, dated March 8, 2000, Gazzetta Ufficiale, No. 60/2000.

24
Bundesgesetzblatt I, 1994, p. 180.

25
Bundesgesetzblatt I, 2000, p. 1645.

26
Wet van het ouderschapsverlof, Staatsblad 1990, 560.

27
These data are based on a definition of part-time work which comprises all employees working less than 35 hours a week (Visser 1999: 286). Thus, the proportion of employees with a weekly working time of less than 20 hours will be lower than 35 percent.

28
Wet van 25 juni 1997 tot wijziging van titel 7.10 (arbeidsovereenkomst) van het Burgerlijk Wetboek met betrekking tot het ouderschapsverlof, Staatsblad 1997, 266.

29
According to government estimates, the annual costs of implementing the working time Directive, which required the introduction of the 48-hour week and an entitlement to four weeks annual leave, amounted to 1.9 billion pounds or some 3.1 billion euro (DTI 1998b: Annex 3). The national minimum wage legislation that came into effect in April 1999 had estimated annual costs of 2.5 billion pounds (about 4.1 billion euro, see DTI 1998a: Annex 5).

30
The Maternity and Parental Leave etc. Regulations 1999, S.I. 1999/3312, came into force on December 15, 1999, which was exactly the date on which the implementation period ended.

31
The Maternity and Parental Leave (Amendment) Regulations 2001, S.I. No. 4010, 2001, which came into effect on January 10, 2002.

32

The Parental Leave Act, 1998 (1998, No. 30), was adopted on July 8, 1998, and came into force on December 3, 1998.

33
The text of the Reasoned Opinion issued by the European Commission against Ireland is reprinted in Clauwaert/Herger (2000: 117-118).

34
The convention collective de travail N° 45, of December 19, 1989, which established the principle of leave for urgent reasons, was ratified by arrêté royal of March 6, 1990 (Moniteur Belge of March 21, 1990). In Belgium the social partners can negotiate national collective agreements in the Conseil National du Travail (National Labour Council, CNT). These CCTs are binding and apply to all workers in the country. Many labor law regulations have their origins in this bipartite body. The resulting collective agreements are often framed subsequently by a law or arrêté royal (royal decree, AR).

35
Under the still existing IDC scheme time off can also be taken in a piecemeal way or in a combination of different forms, but cannot exceed a maximum of 60 months. If the worker is older than 50, the scheme can be used as part-time early retirement and there is no maximum limit of 60 months. Originally, only SMEs were excluded from the replacement duty, since one of the main reasons for introducing this scheme in the first place was the employment effect. Note that the necessity to replace workers taking leave under the IDC expired on December 31, 2001 for all undertakings.

36
It came into force on January 1, 1998.

37
The Commission initiated an infringement procedure against Belgium over the provision of leave for urgent family reasons. Belgium had not notified the already existing legislation, but interview partners claimed that this was just an administrative error (Interview B6: 487-519).

38
An arrêté royal from October 29, 1997 (Moniteur Belge November 7, 1997) introduced parental leave as a sub-scheme of the IDC. It was later modified on January 20, 1998, and August 10, 1998.

39
Journal Officiel N° 171, July 26, 1994, page 10739. The law came into force on January 1, 1995.

40
Law N° 2000-1257 on the financing of social security of December 23, 2000, p. 20558. Journal Officiel N° 171, December 24, 2000. The law came into force on January 1, 2001. Allowing families to deal with accidents and sickness, the leave can be taken for up to a total of 12 months for a child of any age. It is financed by the allocation de présence parental (APP) with approximately 460 euro per month if taken full-time, approximately 300 euro per month if taken part-time and approximately 610 euro per month if the parents share this leave.

41
Council Directive 92/85/EEC of 19 October 1992 on the introduction of measures to encourage improvements in the safety and health at work of pregnant workers and workers who have recently given birth or are breastfeeding (tenth individual Directive within the meaning of Article 16 (1) of Directive 89/391/EEC), OJ L 348, November 28, 1992, 1-8.

42
Law 39/99 on the reconciliation of work and family life of November 5, 1999, published in the Boletín Oficial del Estado N° 266, November 6, 1999, page 38934-42. This law came into force on November 7, 1999.

43
Efemeris tes Kuberneseos, N° 153, October 8, 1984, Page 1840.

44
Efemeris tes Kuberneseos, September 2, 1998. The law came into force in 1999.

45
Even if the ETUI report mentions Portugal as the only country where this right is still not guaranteed (Clauwaert/Harger 2000: 9 and 77), this claim seems to rest on an overtly legalistic interpretation. Although there is no explicit legal transposition, an interview partner claimed that this right is nevertheless generally assured (Interview P1: 1788-1851).

46
Note that we still need final confirmation for the Portuguese case.

47
Diário da República, IS-A N° 203, August 31, 1999, which came into force on December 1, 1999. The law is the 4th amendment to law 4/84 on the protection of motherhood and fatherhood (Diário da República, IS N° 81, April 5, 1984).

48
Previous amendments in 1995 introduced the option to take part-time leave (17/95, June 9) and a special provision for handicapped children (17/95, June 9) as well as leave to take care of chronically ill persons in 1997 (102/97, September 13).

49
Note that a comparison of the size of misfit across Directives, and notably across policy areas, is difficult. It seems that the studies on environmental and transport policy (Héritier et al. 1996; Héritier et al. 2001) uncover a size of misfit that is generally much larger than the misfit created by our labor law Directives, which demand a fine-tuning of existing national standards rather than a profound overhaul of the regulatory philosophy. We expect a much clearer picture to emerge and will offer a more thorough discussion of the misfit issue once we can compare all our six Directives.

50
This should include an analysis of the costs incurred in all member states and a comparison with other Directives in the social field and can therefore only be presented by the authors in the future, on the basis of the evaluation of all parts of our expert interviews in the 15 member states.

51
Only on this broader basis can we hope in the future to judge the extent to which national social policies are transformed through EU Directives. For an outline of our ongoing project and intermediate results, see 
<http://www.mpifg.de/fo/multilevel_en.html#Proj5.>

Copyright © 2002 Gerda Falkner et al.

No part of this publication may be reproduced or transmitted without permission in writing from the authors.
Jegliche Vervielfältigung und Verbreitung, auch auszugsweise, bedarf der Zustimmung der Autoren.

MPI für Gesellschaftsforschung, Paulstr. 3, 50676 Köln, Germany