(Work in progress; comments are welcome) 1. Asya Zhelyazkova and Cansarp Kaya

Transcription

1 Notified and Substantive Compliance in an Enlarged Europe. Evidence from Two Policy Areas (Work in progress; comments are welcome) 1 Asya Zhelyazkova and Cansarp Kaya Abstract This paper presents first results from an ongoing research project on compliance with EU policies in 27 member states. The project addresses current a debate in the literature: whereas quantitative studies show that the new EU entrants from Central and Eastern Europe (CEE) are the forerunners in the legal incorporation of EU directives into national law (Knill and Tosun, 2009; Sedelmeier, 2008), detailed case studies indicate the existence of a gap between legal and practical compliance with the EU rules (Falkner et al., 2008). The project aims to reconcile these divergent findings by comparing the performance of different EU member states regarding different forms of compliance: timely notification, correct legal and practical implementation. In this paper we address the following questions: Do we observe systematic variation in compliance (a) between different EU member states, (b) between different forms of compliance such as timely notification, correct transposition and practical implementation? What factors account for variation in compliance across different forms of compliance? The presented findings are based on a dataset on member states notified and substantive compliance (both legal and practical) with 20 directives from two policy areas: Internal Market and Services (10 directives) and Justice and Home affairs (10 directives). The preliminary results show that patterns of timely notification are not systematically related to the correctness and the application of domestic laws relative to the EU requirements. 1 Address for correspondence: asya.zhelyazkova@eup.gess.ethz.ch 1

2 Introduction The success of the European integration project does not only depend on the adoption of common European policy and the Union s expansion of its membership. The effectiveness of European integration is also conditional on the extent to which the EU policies are incorporated and applied by the EU member states. Therefore, it is surprising that ten years after the major enlargement of the Union to include eight new member states from Central and Eastern Europe (CEE) (followed by Bulgaria and Romania in 2007 and Croatia in 2013), there is little knowledge about member states implementation performance in an enlarged Europe. Furthermore, whereas research on implementation prior enlargement is abundant, quantitative studies use rather imperfect proxies to describe and explain existing patterns of compliance with the EU rules. For, example, many studies use infringement proceedings initiated by the EU Commission against a member state as a formal indicator for non- compliance (Mbaye, 2001; Perkins and Neumayer, 2007). Infringement proceedings, however, refer to cases that are both detected by the Commission and on which the Commission decides to take an action. There may be many other cases of non- compliance that do not show up in data on infringement proceedings (Hartlapp and Falkner, 2009). Other studies use delays in the legal incorporation (also referred to as transposition) of EU directives as an indicator for non- compliance (Mastenbroek, 2003; Kaeding, 2008; König and Luetgert, 2009). There is no guarantee, however, that the national laws reported transpose the directives adequately. Finally, quantitative studies exclusively focus on the legal side of implementation outcomes and do not provide information about the extent to which the EU rules are applied in practice by the relevant national authorities. The present paper is part of a research project that seeks to improve our understanding about member states implementation of EU policies by addressing the following questions: Do we observe systematic variation in compliance (a) between different EU member states, (b) between different forms of compliance such as timely notification, correct transposition and practical implementation? What factors account for variation across different forms of compliance? 2

3 In this paper, we provide some preliminary findings about existing patterns in the legal and practical implementation in 27 member states regarding 20 EU directives from two policy areas: Internal Market and Services (10 directives) and Justice and Home Affairs (10 directives). Based on information provided by various expert evaluation reports and conformity studies, we make a distinction between legal and practical implementation. In addition, we compare our dataset on substantive implementation outcomes (both legal and practical compliance problems) with notified compliance to the Commission (delayed notification of implementation measures to the Commission). 2 The preliminary findings show that patterns of timely notification are not systematically related to the correctness and the application of domestic laws relative to the EU requirements. Furthermore, correct transposition does not necessarily mean that the EU rules are properly applied by the relevant national institutions, as we still observe variation in application problems, even if directives are properly incorporated in domestic law. This paper proceeds as follows. We start by discussing existing insights about compliance in both Western and CEE member states before and after their accession to the EU. Afterwards, we proceed with a description of the types of implementation issues and our operationalization of different aspects of compliance with EU law, followed by both descriptive and explanatory analysis of substantive compliance problems in 27 member states. We conclude with a discussion of the findings and the next steps in developing the project. Research on policy implementation before and after EU enlargement The topic of policy implementation across various national and policy contexts has been a major part of the research agenda of Europeanization scholars (Börzel and Risse, 2007; Cowles et al, 2001; Héritier, et al., 2001; Schimmelfennig and Sedelmeier, 2005; for overviews see Börzel and Risse, 2003; Sedelmeier, 2011). There are two sub- fields within the Europeanization literature that are 2 The data-set is part of a project funded by the Swiss National Science Foundation (SSNF) that aims to gather comprehensive data on the actual implementation of EU rules in 27 member states regarding directives from four policy areas: Internal Market and Services, Environment, Social Policy(employment and anti-discrimination) and Justice and Home Affairs. 3

4 relevant for the present research project: studies of EU enlargement and studies of member states compliance with EU laws. First, the influence of EU conditionality on CEE countries progress in implementing the acquis has dominated studies of enlargement (Hughes et al., 2004, Grabbe, 2006, Schimmelfennig and Sedelmeier, 2005; Sedelmeier, 2011). This research suggests that conditionality upsets the domestic equilibrium by introducing additional incentives for compliance with EU rules. The finding that the external incentive of membership was the key mechanism that led to the adoption of EU rules by the candidates makes the question of post- accession compliance even more salient (Dimitrova, 2010; Schimmelfennig and Sedelmeier, 2004: ; Sedelmeier, 2008). More precisely, after membership is granted, changes in the incentive structure are expected to affect member states behavior and lead to deterioration of the implementation performance of CEE countries. Instead, however, we observe consistently high compliance records of CEE countries after accession. More surprisingly, based on data on timely notification and infringement cases, the CEE member states seem to be outperforming their Western counterparts (Knill and Tosun, 2009; Sedelmeier, 2008). In contrast, studies of member states compliance with EU rules generally focus on countries implementation performance in the absence of conditionality (for overviews of the literature, see Mastenbroek, 2005; Toshkov, 2010; Treib, 2008). A rich body of literature has been dedicated to the study of transposition of EU directives by the old 15 member states and infringement cases initiated by the Commission (e.g., Börzel, 2001; Jensen, 2007; Kaeding, 2008; König and Luetgert, 2009; Mastenbroek, 2003; Mbaye, 2001; Thomson et al., 2007; Zhelyazkova and Torenvlied, 2009). Unlike studies of conditionality, the EU compliance literature puts most emphasis on domestic factors in explaining member states behavior. Whereas this literature focuses on a multitude of factors (Toshkov (2010) uncovers no less than 263 relationships emphasized by scholars!), most studies analyzing country- level differences make a distinction between preference- based and capacity- related explanations for implementation problems. Capacity- related factors combine both country- level capacity limitations (e.g., government and bureaucratic effectiveness) (Mbaye, 2001; Toshkov, 2009) and institutional constraints on the national implementation process within member states, such 4

5 as veto players (e.g., administrative coordination among domestic actors) (Dimitrova and Toshkov, 2009; Steunenberg, 2006) Preference- based explanations often refer to societal and government support for the EU or government policy preferences regarding particular EU directives (Thomson et al, 2007; Thomson, 2010). Some studies also associate member states general willingness to comply with EU legislation relates to the economic benefits that countries gain as a result of their membership in the European Union (Perkins and Neumayer, 2007). As regards to differences between the new CEE and the old Western member states, the most prominent explanatory factors that influence Western member states compliance with EU rules are related to low administrative capacities and high adjustment costs. In contrast, comparative analyses of CEE countries transposition of the acquis show that the preferences of governments and their general ideological orientations could explain variation in compliance before accession (Toshkov, 2008). Recent studies comparing the implementation efficiency in both CEE and Western member states support these findings (Yordanova and Zhelyazkova, 2011). Whereas capacity- related factors such as government effectiveness are more relevant for the old 15 member states, preference- based explanations such as support for the EU seem to explain variation in compliance in the new CEE countries. Despite growing research on member states implementation with EU rules, there are still some important challenges that EU compliance scholars currently face. One major challenge, for instance, lies in the debate between case study approaches that focus on the practical implementation of EU policies and large- N quantitative studies that rely on data from member states notified measures to the Commission. In particular, detailed qualitative studies point out to the existence of a gap between formal and practical compliance with the EU rules (Falkner et al, 2005; Falkner et al., 2008) that cannot be captured by the notification records and infringement cases (Hartlapp and Falkner, 2009). These findings suggest the need for a more systematic comparison between different types of compliance problems across EU member states and policy areas. 5

6 A comparative approach for studying legal and practical compliance in the EU: challenges and research strategy In the context of this project, we make a distinction between three forms of compliance: timely notification to the Commission, correct legal (i.e. transposition) and practical implementation. A common problem of standard compliance indicators is that they only capture notified compliance and do not distinguish between notified and substantive implementation performance on the one hand and legal and practical compliance, on the other. Thus, current research on compliance has primarily focused on the timing or the speed at which the EU policies are legally incorporated by the EU member states. The focus on timing could be easily explained by the fact that the data are readily available in the Commission databases. A more substantive explanation rests on the assumption that member state s national authorities eventually incorporate the EU directives in their national legal systems in order to avoid punitive actions taken by the Commission and the European Court of Justice (ECJ). While this assumption is reasonable, the timing of transposition tells us little about how the EU rules are incorporated in national law and implemented in practice. To address this problem, some scholars rely on data from infringement procedures as a more appropriate indicator for substantive violations of EU law. Infringement cases are opened by the EU Commission in response to violations of the EU rules and generally consist of three consecutive stages: 1) letter of formal notice (where the Commission inquires national authorities about possible violations); 2) reasoned opinion (the Commission formally establishes that a violation has occurred and requires that national authorities remedy the situation) and 3) referral to the ECJ. Whereas infringement proceedings should in theory cover problems related to both legal and practical implementation, in reality they refer to delayed notification, at least in the majority of cases. For example, in our dataset on the implementation of 20 directives by 27 member states, only one infringement case was opened against incorrect practical implementation, while the rest refers to transposition delays. There could be many reasons why application problems (or even issues related to incorrect transposition) do not show up in infringement data. The most obvious reason is that the Commission does not have the capacity to actively monitor the 6

7 implementation process in each and every member states. Even if the Commission officials obtain information about practical implementation, information about violations is unlikely to show up in the records, as the Commission will try to resolve the issue before escalating the conflict to more formal stages. Finally, the Commission may have incentives to pursue compliance with some policies and not others. For example, cases of non- transposition of major provisions of a directive may receive more attention than cases of only partial non- compliance or ambiguous transposition. Nevertheless, the latter could still cause uncertainty about the impact of the directive in the respective member state (Zhelyazkova, 2013). In short, standard compliance indicators cannot be sufficient to reveal patterns regarding member states legal and practical conformity with EU rules. In this research project, legal compliance is broadly defined as the extent to which a member states recently adopted and pre- existing legal system is in conformity with a EU directive. 3 Thus, correct transposition means that national laws are compatible with the goals of the EU directives. In addition, assessment goes beyond simply comparing the legal texts of notified implementation measures and EU rules. For example, some text of the EU directive may be missing from the national law transposing the directive because it is not relevant for that particular country. In other cases, the respective text may be covered by existing national rules and practice, even if these haven t been communicated to the EU Commission. In other words, assessment of legal conformity requires extensive knowledge about the topic covered by the directive as well as the legal system of a member state. Analysing practical implementation in a comparative framework is even more daunting. Unlike legal compliance, the nature of a practical problem varies across policy issues and domestic settings. The issue is further complicated because the actors responsible for practical implementation are no longer directly associated with the national government. Instead, directives target different societal actors within the member states (e.g., bar and restaurant holders regarding smoking 3 Member states often adopt different laws regarding an EU rule. For matters of feasibility, the following project does not focus on the extent to which each particular national legal measure conforms to an EU rule, but relies on information about a member state s legal compliance with this rule as a whole. 7

8 ban). Because we are aware of these issues, in this research project, practical implementation refers to the actions taken by national authorities to ensure the proper application of EU rules by the relevant actors (Versluis, 2007). Important aspects of practical implementation thus include 1) providing both monitory and institutional resources for the application of a law and 2) enforcement measures such as effective and dissuasive sanctions against non- compliance and efficient court proceedings. As a result, practical implementation is defined and measured at the member state level. Such operationalization ensures that the concept of practical implementation is comparable across countries and policy areas and keeps the coding of compliance manageable within the scope of the project. Nevertheless, we are aware that by constraining the analysis to governments and other central authorities, we are unable to capture most application problems that occur within member states. However, at the very least, we will be able to distinguish between domestic laws issued just as a façade to satisfy the EU Commission and cases where national authorities actually take the effort to enforce the EU requirements. It is important to note that legal and practical compliance are related to one another. Incorrect transposition also has consequences for how the EU rules are applied in practice. However, the opposite does not hold, as the correct transposition of a directive does not necessarily lead to effective implementation in practice. For example, national legislation could literally mirror the provisions of a directive, while national ministers or other relevant actors within the government may do little to make these provisions effective in practice. Therefore, the operationalization of practical implementation remains distinct from transposition. To provide an example, expert evaluation reports describe a practical problem in the Lithuanian implementation of the directive on reception conditions for asylum seekers. Even though national legislation regulates material reception conditions as required by the directive, national experts evaluated the material resources allocated by the government as not sufficient to ensure an adequate standard of living of the applicants. 8

9 Research design Selection of directives and sources In this paper, we present some first results regarding the relationship between different types compliance across 27 member states (excluding Croatia). The dataset is based on 20 EU directives from two policy areas: Internal Market and Services (10 directives) and Justice and Home Affairs (10 directives). Justice and Home Affairs directives regulate the free movement of persons, asylum and immigration in the Union. Some of these directives define the standards for reception of asylum applicants (2003/9/EC), of refugees (2004/83/EC), of a mass of displaced persons (2001/55/EC), and of victims of human trafficking (2004/81/EC). Another issue regulated by the directives is Illegal immigration. Directive 2002/90/EC defines illegal immigration, whereas Directive 2001/51/EC lists the financial penalties imposed on persons engaged in this act. Mutual recognition to expulsion decisions is regulated by Directive 2001/40/EC and assistance in removal by air by Directive 2003/110/EC. Finally, the sample also includes directives concerning the general rights of non- EU citizens, such as long- term residence (2003/109/EC) and family reunification (2003/86/EC). The Internal Market directives mainly address two issues: businesses in the internal market and single market for services. Three of ten directives are related to harmonization of businesses. They aim to regulate the takeover bids (2004/25/EC), facilitate cross- border mergers (2005/56/EC) and protect the intellectual property (2004/48/EC). The rest of the directives aim at the removal of the barriers to services. Directive 2006/123/EC is the general framework for services, whereas others focus on specific sectors. Directive 2008/6/EC aims to create an internal market in postal services, whereas the other five are essential for the integration in financial services. These directives address the transparency of the securities market (2004/109/EC), financial risk management systems (2006/48/EC and 2009/49/EC), payments services (2007/64/EC), and electronic money institutions (2009/110/EC). The focus on directives from the two policy areas is appropriate given the objectives of the study. First, they ensure variation in compliance with EU rules. 9

10 For example, Internal Market directives contain more technical rules compared to other policy areas. According to the last Commission report on the application of EU law, Internal Market has the fourth highest number of opened cases among 21 policy areas. In contrast, fewer infringement cases are opened in the area of Justice and Home Affairs, as these contain more general rules and provide more discretion to national policy makers. Second, both policy areas address related topics, which have nevertheless developed differently regarding their level of centralization at the EU level. For example, whereas market liberalization in terms freedom of movement for goods and services has also pushed for increased freedom of movement for persons, integration forces have been much slower and fragmented in the latter case. It remains unclear whether the same patterns hold for the implementation of market- related EU policies and directives that touch upon national identities and values. Table 1 presents the sources of different types of information that were used in this paper. Information about the legal and practical implementation was obtained from different expert evaluation reports and conformity studies. All studies and reports share important characteristics. Table 1: Data sources Practical implementation Legal implementation Acts and Actors JUSTICE AND HOME AFFAIR DIRECTIVES 10 directives Odysseus Reports (2008) INTERNAL MARKET DIRECTIVES 2004/25/EC N/A Marccus Partners (2012) IBA (2008) 2004/48/EC N/A Tipik (2012) 2004/109/EC CESR (2008), Mazars (2009) CESR (2008) 2005/56/EC N/A Lexidale (2013) 2006/48/EC N/A DLA Piper UK LLP (2009) 2006/49/EC N/A DLA Piper UK LLP (2009) 2006/123/EC Siemens Report (2008) Eurochambers (2011), Commission Report (2012) Milieu (2011), Eurochambers (2011), Stelkens et al. (2012), Commission Report (2012) 2007/64/EC N/A Tipik (2011) 2008/6/EC Copenhagen (2010), EGRP Copenhagen (2010), WIK (2013) (2011), WIK (2013) 2009/110/EC N/A Tipik (2013) Milleu (2011) 10

11 First, they were prepared by an external agency on the request and financial support of the Commission. Thus, the main goal of these reports was to inform the Commission about the implementation process in the member states. 4 Second, all studies and reports were based on information provided by national experts in the particular topic covered by the directive, as well as in the national legal system. Finally, the reports do not evaluate the directives as a whole, but separate articles and sub- articles of the directives. In some cases, the reports even provide separate assessment of individual sentences within a sub- article, if these could have different implications for a member state s implementation performance. Information about member states compliance with Justice and Home Affairs directives were obtained from the Conformity checking of the transposition by Member States of 10 directives in the sector of Asylum and Immigration study, conducted by the Odysseus Academic Network for Legal Studies on Immigration and Asylum in Europe. Table 2a and 2b provide an example for the information in the reports and the structure of the conformity tables illustrating the findings of the Odysseus study. Table 2a: Excerpt from the Synthesis report on the Qualification Directive Article 4(1), second clause, Q. 8 from the national report NO TRANSPOSITION AT ALL Bulgaria, Lithuania LEGAL PROBLEM Czech Republic, France, Romania, Spain PRACTICAL PROBLEM Luxembourg, Slovakia Table 2b: Example from the conformity table in the Czech Republic regarding the Qualification directive Provision of the Reference to State of Evaluation National directive Number of transposition provision of Question or transposition Sub-question Article 2(e) Q.6 Law In order Part I, Art I s 27 Article 4(1) Q.8A & Q8.B Pre-existing law Problem(1) Art 49a ASA Second Clause Article 7(1) Q.14 Law Problem(2) Art 2.8 ASA Article 20(3) Q.30.C Not Not(3) (1): Transposition of this provision is incomplete (2): Art. 7 (1) QD was transposed only with regards to the refugee status. Therefore, the ASA does not contain a concept of actors of protection from serious harm (interpreted textually) (3) This provision is not transposed into national legislation and is not regulated by existing laws. 4 However, as the Commission explicitly pointed this out to us, the contracts with the external agencies do not necessarilly mean that the EU Commission shares the opinion of the authors of the reports or that it is obliged to follow their recommendations. 11

12 Regarding the Internal Market directives, no single report exists on all directives. Instead, each directive was evaluated in separate studies (except directives 2006/48/EC and 2006/49/EC, which were analysed in one report by the same experts). In some cases, directives 2006/123/EC and 2008/6/EC (the services and the postal services directives), several reports were analysed, as they covered different parts of these directives. We additionally contacted the authors to request further information if the existing information in the reports was incomplete or if there were inconsistencies. 5 Because we used different sources of information regarding the implementation of the Internal Market directives in the 27 member states, the expert reports also differ in structure. Half of the reports have similar structure where the evaluations of each directive s provision were presented in a table. The other half of the directives are coded based on different types of reports: member states questionnaires (directive 2004/109/EC transparency requirements), a general study with country specific reports focusing on provisions (2005/56/EC) or developments in that sector (2008/8/EC), a general study without country specific reports (2004/25/EC), and a book on implementation of law (2006/123/EC). To increase the quality of the data, the information was coded by three different coders and the final results were compared. Generally, coding decisions were identical and in the few cases of disagreement, such cases were excluded from the analysis. Finally, some data limitations should be also addressed and discussed. For example, whereas all the reports provide extensive information about the legal aspect of compliance, they differ regarding their focus on practical implementation. Thus, whereas we have information about the practical implementation of all 10 Justice and Home Affairs directives, the same type of information is only available for three of the 10 Internal Market directives (the services and postal services directive as well as the directive on transparency requirements). 5 In the process of coding compliance problems, we identified inconsistencies regarding only one of the directives in the study: directive 2005/56/EC on cross- border mergers and acquisitions. However, the authors of the report helped us resolve these issues. 12

13 Another important limitation of the project is that both legal and practical compliance are measured at a single point in time, while in reality implementation performance varies over time. This is especially problematic for practical implementation, where it takes years after the adoption of a law before application problems become apparent. Likewise, problems of incorrect transposition could be eliminated by the adoption of new domestic laws or vice versa. While we are unable to take into account all changes in national legislation and practice, the time frame of the project allows us to study compliance several years after the prescribed deadlines, at a period when member states were expected to have transposed and applied the EU directives. Thus, all conformity studies used in this study were prepared several years after the transposition deadlines and cover compliance in 27 countries in the period between 2007 and This time period allows us to include in the analysis both the old 15 member states, as well as the new member states from Central and Eastern Europe. Measuring different aspects of compliance As already mentioned, the dependent variables in this study reflect three different forms of compliance: timely notification, correct legal transposition, and practical implementation. Delays in notification are generally measured on the basis of the Eurlex database, which provides the number and the implementation dates of national implementing measures. However there are cases where member states report very early measures that are not entirely relevant for the directive. For that reason, we use information on infringement proceedings that were initiated by the Commission as a result of delayed notification. In this study we present findings on reasoned opinions, which is the stage where the Commission has established that a country hasn t notified a relevant transposition measure. Information about legal implementation is obtained from expert evaluation reports. As already noted, these reports analysed compliance at the provision level. In this study, we only consider provisions that require member states to take a certain action. For example, we excluded cases where a provision was not applicable to a member state or it was optional to implement. The number of 13

14 relevant provisions varies between directives and between member states implementing the same directive. On the one hand, for eight member states, the number of non- optional and relevant provisions was only three in the directive 2004/25/EC on take- over bids (the lowest number in our dataset). In the other extreme, Luxembourg had to transpose 775 provisions from directive 2006/48/EC. At the provision level, legal implementation of each provision is coded into four categories. Full and correct implementation is coded as 3. Partial implementation is observed when a certain part of a provision is not implemented and is missing and it is coded as 2. When national experts identified a problem with the transposition of a provision, this case is considered as incorrect implementation and coded as 1. If a provision is not implemented at all, it is coded as 0. To aggregate this information to directive level, we recoded the variable as 1 if member states experienced any problems in legal implementation (including non- transposed provisions) and 0 otherwise. At the directive level, Legal non- conformity is measured as the ratio of incorrectly transposed relevant provisions in a directive from the total number of evaluated provisions. Because the reports provide separate information about practical implementation problems, this variable is also coded sedately from legal implementation. In order to understand the measure better, we provide an example from the Services Directive (2006/123/EC). For example, Article 8 (1) of the Services Directive states all procedures and formalities relating to access to a service activity and to the exercise thereof may be easily completed, at a distance and by electronic means, through the relevant point of single contact. According to the Commission report (2008) on the implementation of the directive, 12 member states transposed this rule without any legal problems. However, some years after the transposition deadline, Eurochambres (2011) surveyed member states about the implementation of the directive and asked if their points of single contact provide the means to fully complete formalities, both electronically and from a distance. The results were different than the legal implementation situation. Half of the countries that implemented the provision correctly in legal terms had points of single contact that did not provide a complete set of formalities yet. Problems associated with practical 14

15 implementation as coded in the same way as legal non- conformity. Fur the purposes of this study, the variable reflects the ratio of provisions associated practical implementation problems within a directive. Descriptive analysis: comparison of different forms of compliance In this paper, we provide some first descriptive results on how timely notification, legal conformity and practical implementation relate to one another across 27 member states and two policy areas. Figure 1 illustrates the distribution of the substantive compliance indicators across member states. Figure 1a combines legal and practical non- conforming provisions and presents country- level average scores separately for the policy areas. Instead, in Figure 1b, a distinction is based between practical and legal compliance problems, while the country scores are based on all relevant directives. The computations on practical implementation problems are based on 13 directives instead of 20 because of missing information on this variable for the Internal Market directives. Figure 1a shows that member states experienced on average more problems implementing the ten Justice and Home Affairs directives than the directives related to internal market and services. This observation is not surprising, given that the former address more sensitive issues related to immigration of third country nationals and integration of refugees. In addition, we do not observe a distinct pattern between new CEE and the old Western member states that would suggest the existence of an east- west divide in EU compliance. Whereas Bulgaria and Poland on average incur a higher ratio non- conforming issues, Slovenia, Slovakia and Lithuania excelled relative to other member states regarding the internal market directives. Nevertheless, the average country differences seem to be negligible regarding member states implementation of the internal market directives. In addition, most member states respond differently to the requirements from the two policy areas (Bulgaria is an exception, where the country scored high on both internal market and justice and home affairs directives). For example, whereas Estonia and Spain experienced relatively few compliance problems with the ten internal market directives, the national authorities in these two countries lead the way in the 15

16 ratio of problematically implemented provisions in the area of justice and home affairs. Turning more specifically to compliance problems related to practical implementation, we observe limited variation in the average scores across the 27 countries (see Figure 1b). Instead, there seems to be more heterogeneity in the extent to which countries implement different directives (as seen by the wide confidence intervals for some of the member states). The heterogeneity in implementation outcomes is especially pronounced in Ireland. However, this is likely due to the smaller number of observations that are relevant for Ireland. Figure 6 1a: Average ratio of conformity problems (legal and practical issues) across two policy areas (bootstrapped 95% CIs around the average score, 1000 draws) EU Member States bg pl lu nl it fi cz lv el cy dk ro pt fr de mt se hu uk at be est es ir sk si lt est es bg at hu ro lv nl lt mt it lu be cz el sk fr fi pl si de cy se pt Ratio of non-conforming provisions (internal market) Ratio of non-conforming provisions (justice and home affairs) 6 Denmark, Ireland and the UK were excluded from the JHA sample of directives, as only few of the 10 directives were applicable to the three member states. 16

17 Figure 1b: Average ratio of practical and legal problems across 20 directives (bootstrapped 95% CIs around the average score, 1000 draws) EU Member States bg est es at ro lv nl lu lt it hu pl cz mt fr be el fi dk cy de sk si se uk pt ir ir hu it dk pl mt est pt si el lv cz lt sk bg fi de fr cy es ro nl at uk be se lu Ratio of legally non-conforming provisions Ratio of practically non-conforming provisions Figure 1b does not illustrate the relation between the different forms of compliance. Instead, in Figure 2 (a- b) member states average scores on the three dependent variables are plotted against each other to check whether there are any discernable patterns. Figure 2a compares the average country scores for legal and practical implementation, but it also makes a distinction between directives belonging to one of the two policy areas. Interestingly, the two policy areas differ in their relationship between legal and practical compliance. In the area of internal market and services, all member states exhibit (on average) similar and rather low share of legal compliance problems, whereas their share of application problems varies considerably across the 27 countries. We observe the opposite relation for the Justice and Home Affairs directives. Whereas the ratio of legal problems varies considerably across the EU member states, the average score related to practical implementation problems is rather low for all member states. 17

18 Figure 2a: Legal and Practical conformity across two policy areas Legal and Practical Implementation Legal non-conformity(ratio of incorrect provisions) at bg ro lt lv hu nl be lu mt it cz fr lu sk nl si pl el de fi cy se es uk se pt est es dk at ro de bg cy fr it fi est be sk ir Application problems(ratio of practical problems) cz mt lt pl lv el si hu pt Internal Market and Services Justice and Home Affairs The differences in legal compliance corroborate general expectations that justice and home affairs directives are more difficult to implement as they address nationally sensitive issues. In contrast, the internal market directives often consist of rather technical requirements. Furthermore, the Commission puts much high emphasis on the transposition of internal market directives. The observed differences in practical implementation are more difficult to explain. One possible explanation relates to the way the variables for legal and practical implementation are operationalized in this study. If provisions are generally transposed incorrectly in national law, related practical problems will also ensue and we are unlikely to observe many more practical problems separate from those related to legal transposition. Still, the finding that member states considerably vary in the practical implementation of the correctly transposed in general internal market directives suggests that even technical policies are likely cause implementation problems. The emphasis of the Commission on compliance with internal market polices seem to ensure their legal conformity, but the Commission impact does not extend to member states practical implementation. Figure 2b provides some visual comparison for member states average scores on different compliance indicators computed based on all directives in the 18

19 current dataset. To capture problems with delayed implementation, we use the ratio of infringement cases (reasoned opinions) initiated by the Commission as a result of transposition delays. In addition, the figure also distinguishes between Western and the new member states from Central and Eastern Europe (CEE) as preliminary evidence regarding the alleged west- east divide regarding the application of EU law. Figure 2b: Comparison between legal, practical and delayed implementation Legal non-conformity(ratio of incorrect provisions) Ratio of reasoned opinions(delayed notification) Legal and Practical Implementation lu at se be nl ro es bg est cy fr uk de fi el mt it sk lt cz lv pl si dk pt hu Application problems(ratio of practical problems) lu se cy nl el be es fr de mt it pt uk lt fi ro sk cz lvsi siest pl at bg dk hu Application problems(ratio of practical problems) ir Notified and Practical implementation ir Ratio of reasoned opinions(delayed notification) Ratio of reasoned opinions(delayed notification) Notified and Legal Implementation ir pt uk se si cy el be de frmt sk fi dk it hu pl lt cz lu nl es lv ro at est bg Legal non-conformity(ratio of incorrect provisions) Notified and Substantive Implementation ir pt uk se si de be fr cy dk el lu mt it pl lt fi sk cz nl hu lv ro at es est Substantive non-compliance(legal and practical combined) Western member state CEE member state bg Based on Figure 2b, we do not observe any clear relationship between legal and practical implementation problems, on the one hand, and substantive 19

20 implementation problems (legal and practical) and delayed notification, on the other. This finding is also corroborated by simple correlation tests (e.g., Pearson rho), which show no significant correlations between the different aspects of compliance. Furthermore, based on our measures for legal and practical implementation, there is no evidence for a distinct west- east divide. In congruence with existing research, however, the CEE member states receive fewer infringement cases due to delays in transposition than their Western counterparts. Explanatory analysis Whereas the descriptive analysis does not illustrate any distinctive country- level patterns in different forms of compliance, it still shows variation across member states and directives in their national implementation outcomes. Even though, we did not formulate specific hypotheses regarding the relevant factors that could influence different types of compliance problems, it is still important to analyze whether the same factors explain variation in the different dependent variables. To this end, we conducted preliminary statistical tests on the likelihood of member states share of compliance problems. The statistical analysis includes both capacity- related and preference- based factors employed in existing research on implementation. For each directive, we defined the parties that were in office when that directive was transposed into national law based on the database of Döring and Manow (2010). In accordance with existing research, we assume that a member state s willingness to comply with a EU directive is reflected in its level of support for EU integration. This variable was coded based on data from the Party Policy in Modern Democracies project (Benoit and Laver, 2006) and the Chapel Hill data set (Marks et al., 2006). In addition, changes in government composition before implementation could disrupt the implementation of directives, as new political elites may not share the views of the preceding government that participated in the adoption of the directive at the EU level. As regards capacity- related variables, we include data on a member state s bureaucratic efficiency. The measure was directly obtained from the World Bank Governance indicators database (Kaufmann et al, 2005). In addition, the 20

21 existence of veto players may negatively affect the transposition of EU law. Based on information provided by the expert evaluation reports, we identified the relevant domestic actors responsible for transposition and application of each directive. The variable reflects the number of different ministries responsible for implementing a directive. The statistical analysis additionally compares the performance of Western countries without communist legacies (coded as 1) relative to the new CEE member states regarding different types of compliance problems. We include another categorical variable for policy sector to check whether there are systematic differences between Internal Market (coded as 1) and Justice and Home Affairs directives. Finally, we control for directive characteristics such as policy complexity (measured by the number of recitals in a directive) as well as the number of provisions in a directive that are relevant for a member state. Table 3 presents the results from the statistical models on the different forms of compliance. Whereas the Model 1 applies standard logistic regression on the likelihood of reasoned opinions against delays, the rest of the models analyze member states share of substantive problems (legal and practical combined) in a directive (Model 2), the ratio of incorrectly transposed provisions (Model 3) and the share of practical problems (Model 4). In congruence with the descriptive analysis, we observe significant differences between the new CEE member states and Western countries only with respect to delayed notification. In addition, the level of bureaucratic efficiency in a country is also relevant for compliance problems associated with delayed transposition, but not for more substantive compliance indicators. In contrast, the analysis shows that the number of ministries responsible for implementing a directive has a positive effect only on the share of practical implementation issues in a directive. Regarding the effects of the preference- based factors, surprisingly, government support for the EU has significant positive effect on the share of incorrectly transposed provisions (dependent variable: legal non- conformity), whereas change in government decreases the share of issues associated with practical implementation. However, these contradictory results are driven by a couple of outliers and the results are not robust across different model specifications. 21

22 Table 3: Analysis of different aspects of compliance in 27 member states regarding 20 directives Infringements RO (Delays) Substantive non-compliance Legal non-conformity Practical non-compliance Variables Coef. (S.E) Coef. (S.E) Coef. (S.E) Coef. (S.E) Western member state 2.563*** (0.527) (0.228) (0.236) (0.225) Govt. change ** (0.316) (0.230) (0.230) (0.242) Govt. effectiveness ** (0.388) (0.187) (0.188) (0.229) N ministries ** implementing a directive (0.119) (0.020) (0.019) (0.039) Govt. EU support * 0.103** (0.022) (0.141) (0.057) (0.046) Complexity (N recitals) *** 0.014*** 0.027** (0.008) (0.003) (0.003) (0.009) N relevant provisions ** *** *** ** (0.001) (0.001) (0.001) (0.010) Policy Area *** *** (Internal Market=1) (0.423) (0.246) (0.261) (0.548) Constant * ** *** ** (1.259) (0.622) (0.520) (0.849) Pseudo loglikelihood N Note: Robust Standard errors (observations clustered in member states) Regarding the policy- level characteristics, the number of recitals increases the share of substantive problems (practical and legal combined and separate), but has not effect on infringement cases due to delays. The only variable that has the same effect on all compliance indicators is the number of relevant provisions for a member state. In particular, higher number of relevant provisions is negatively associated with compliance problems. This result is not too surprising given that the longest directives in our sample consist of many technical provisions and 22

23 annexes that could facilitate transposition by providing more specific implementation guidelines to the member states. Finally, as shown by the descriptive analysis, Internal Market directives incur fewer issues related to legal non- conformity than Justice and Home Affairs directives. In contrast, differences in practical implementation between the two policy areas are not statistically significant. This is likely due to the low number of existing reports on the practical implementation of internal market directives (only three out of ten directives). Conclusion The aim of this paper was to provide some first insights about member states legal and practical implementation across two policy areas (Internal Market and Justice and Home Affairs) in an enlarged European Union. Whereas existing research does not distinguish between different types of implementation problems, we tried to discuss and describe possible ways to disentangle (both conceptually and empirically) delayed and incorrect legal implementation as well as practical implementation problems. Understanding these differences is important, given that recent findings suggest the existence of a gap between what scholars refer to as law in the books and law in action (Versluis, 2007). Furthermore, discrepancies between formal and practical compliance have been specifically attributed to the new member states from Central and Eastern Europe. In particular, recent studies suggest that while the CEE member states excel in their efficiency of transposing the EU laws, these laws fail to be applied in practice by the relevant domestic actors. In this paper, we introduce our research approach to studying different types of compliance problems in a cross- country comparative framework. We also provide some preliminary evidence on how different compliance indicators vary across member states and the two policy areas, and test the argument that the new CEE countries are different from their Western counterparts regarding different forms of compliance with EU legislation. Finally, we compare the link between standard capacity- related and preference- based factors on compliance problems associated with delayed notification and incorrect legal and practical implementation. 23