<%BANNER%>

Conforming or Muddling Through

Permanent Link: http://ufdc.ufl.edu/UFE0041597/00001

Material Information

Title: Conforming or Muddling Through Explaining Variations in Compliance with European Union Environmental Policy
Physical Description: 1 online resource (264 p.)
Language: english
Creator: Livanis, Ioannis
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: bargaining, clout, compliance, corruption, domestic, ecj, emas, eu, european, federalism, implementation, industry, intergovernmentalism, international, poisson, politics, principal, regulatory, salience, veto
Political Science -- Dissertations, Academic -- UF
Genre: Political Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: European Union (EU) member states have often failed to transpose and implement EU directives into their national legal order. While this is the case in most sectors of policy, the environment sector seems to bear the brunt of non-compliance. The danger is obvious. Failure in a ?less salient? sector such as the environment could decrease members? belief in ?European solutions? for more salient sectors, and if EU policy is not fully implemented by all member states it risks becoming an empty paper with only a slight effect on environmental quality. Most importantly, non-compliance is an expensive and time consuming reality that serves to take up most of what the Commission time. Three waves of literature have failed to produce an empirically reliable understanding of the determinants of non-compliance, while there is also little theoretical cumulativeness. This study uses transposition rates of EU environmental directives for the 1998-2007 period, distinguishing between the supranational and intergovernmental characteristics of the EU construct, which leads to different sources of theory (in particular International Relations and Regulatory Federalism), and different explanations of non-compliance than previously envisioned. A time-specific fixed effects Poisson estimator is used to demonstrate that non-compliance is a function of both domestic and EU level influences. This study lends support to the ability of supranational institutions to induce compliance using management mechanisms. For instance, it finds that the European Commission?s funding of Non-Governmental Organizations increases compliance. However it also lends support to the strateginess of the decision to comply by member states. Bargaining power in the Council of Ministers increases non-compliance, and so does the existence of powerful industries and cozy regulatory efficiencies. Finally, from a domestic politics standpoint, policy salience and governmental commitment are shown to increase compliance, while corruption and need for regulation are shown to decrease compliance. These findings suggest that the full picture of non-compliance can be attained if theorists take a complementary rather than disjunctive look into the explanations of non-compliance, as implementation in the EU is an intergovernmental enterprise and it should be understood in its own terms, always keeping in mind to look both ways.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Ioannis Livanis.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Hedge, David M.
Local: Co-adviser: Kreppel, Amie D.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041597:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041597/00001

Material Information

Title: Conforming or Muddling Through Explaining Variations in Compliance with European Union Environmental Policy
Physical Description: 1 online resource (264 p.)
Language: english
Creator: Livanis, Ioannis
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: bargaining, clout, compliance, corruption, domestic, ecj, emas, eu, european, federalism, implementation, industry, intergovernmentalism, international, poisson, politics, principal, regulatory, salience, veto
Political Science -- Dissertations, Academic -- UF
Genre: Political Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: European Union (EU) member states have often failed to transpose and implement EU directives into their national legal order. While this is the case in most sectors of policy, the environment sector seems to bear the brunt of non-compliance. The danger is obvious. Failure in a ?less salient? sector such as the environment could decrease members? belief in ?European solutions? for more salient sectors, and if EU policy is not fully implemented by all member states it risks becoming an empty paper with only a slight effect on environmental quality. Most importantly, non-compliance is an expensive and time consuming reality that serves to take up most of what the Commission time. Three waves of literature have failed to produce an empirically reliable understanding of the determinants of non-compliance, while there is also little theoretical cumulativeness. This study uses transposition rates of EU environmental directives for the 1998-2007 period, distinguishing between the supranational and intergovernmental characteristics of the EU construct, which leads to different sources of theory (in particular International Relations and Regulatory Federalism), and different explanations of non-compliance than previously envisioned. A time-specific fixed effects Poisson estimator is used to demonstrate that non-compliance is a function of both domestic and EU level influences. This study lends support to the ability of supranational institutions to induce compliance using management mechanisms. For instance, it finds that the European Commission?s funding of Non-Governmental Organizations increases compliance. However it also lends support to the strateginess of the decision to comply by member states. Bargaining power in the Council of Ministers increases non-compliance, and so does the existence of powerful industries and cozy regulatory efficiencies. Finally, from a domestic politics standpoint, policy salience and governmental commitment are shown to increase compliance, while corruption and need for regulation are shown to decrease compliance. These findings suggest that the full picture of non-compliance can be attained if theorists take a complementary rather than disjunctive look into the explanations of non-compliance, as implementation in the EU is an intergovernmental enterprise and it should be understood in its own terms, always keeping in mind to look both ways.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Ioannis Livanis.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Hedge, David M.
Local: Co-adviser: Kreppel, Amie D.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041597:00001


This item has the following downloads:


Full Text

PAGE 1

CONFORMING OR MUDDLING THROUGH: EXPLAINING VARIATIONS IN COMPLIANCE WITH EUROPEAN UNI ON ENVIRONMENTAL POLICY By IOANNIS LIVANIS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010 1

PAGE 2

2010 Ioannis Livanis 2

PAGE 3

To my parents Theodosios and Konstantina; my br others Charilaos and Gri gorios; and my sister in law Maria, who all made this happen 3

PAGE 4

ACKNOWLEDGMENTS First and foremost, I would like to express my deep gratitude and sincere appreciation to my advisor, Dr. David M. Hedge, for his outstanding guidance, encouragement and advice during my graduate studies and the development of this dissertation. I would like to especially acknowledge Dr. Amie Kreppel for the endless discussions, advice and encouragement during the research process that contributed to the qui cker completion of this dissertation; she has always been a source of motiva tion and inspiration a nd without her support, and that of the Center for European Studies, this dissertation wo uld not have been possible. Sincere appreciation is also extended to the other members of my committee Dr. M. Leann Brown, Dr. Lynn Leverty and Dr. Elias Dinopoulos for their guidance, and c onstructive criticisms th at led to improvements in this dissertation. I would also like to express my immensurable gratitude to my parents, Theodosios and Konstantina Livanis; and my br other Charilaos, for their cont inuous love and moral support, despite the distance. I especially want to thank my brother Grigorios Livanis and my sister in law Maria Chatzidaki, with whom I had the privileg e and pleasure of living together for the first years of my studies, and whos e patience, unending support, and positive encouragement made this possible. Finally, I would al so like to acknowledge all my friends, especially Spiro and Stathy, for their support and humor that made this process lighter. 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS ............................................................................................................... 4 LIST OF TABLES ...........................................................................................................................8 LIST OF FIGURES .........................................................................................................................9 ABSTRACT ...................................................................................................................... .............11 CHAPTER 1 INTRODUCTION ................................................................................................................ ..13 1.1 Introduction .......................................................................................................................13 1.2 Objectives .........................................................................................................................15 1.3 Overview ...........................................................................................................................16 2 THE EU IMPLEMENTATION RESEARCH ........................................................................19 2.1 Introduction .......................................................................................................................19 2.2 An A-Political Beginning ...............................................................................................21 2.3 The Misfits .....................................................................................................................23 2.4 Politics are In ......................................................................................................... ........26 2.5 Quantitative, Qualitative, and Dependent Variables: Faults and Caveats ........................35 2.6 Conclusion ........................................................................................................................41 3 EUROPEAN UNION ENVIRONMENTAL PO LICY: INSTITUTIONS, PROCESSES, AND CURRENT CONDITIONS ...........................................................................................43 3.1 Introduction .......................................................................................................................43 3.2 A Short History of Environmental Governance: Instruments and Strategies ...................45 3.2.1 Idealistic Begi nnings, 1973-1982 ...........................................................................45 3.2.2 The Internal Market is Coming, 1982-1987 ...........................................................49 3.2.3 The Single European Act and the Fourth EAP, 1987 1992 .................................51 3.2.4 Roll-Back and Context-Oriented Governance, 1992-2002 ....................................53 3.2.5 The 6th EAP and the Thematic Strategies, 2002-2012 ..........................................56 3.3 The EU Institutions and Compliance M echanisms: Cooperati on and Deterrence ...........58 3.3.1 The Commission and Monitoring ...........................................................................59 3.3.2 The Commission and the European Court of Justice .............................................61 3.3.3 The Basic Principles of EU Law and the National Courts .....................................68 3.3.4 The European Parliament .......................................................................................71 3.3.5 The Council of Ministers ........................................................................................72 3.3.6 The European Council and European Institutions ..................................................74 3.4. The Compliance Gap .......................................................................................................74 3.4.1 The Gap in the EU ..................................................................................................75 5

PAGE 6

3.4.2 The Gap in the States and the Environment ...........................................................81 3.5 Conclusion ........................................................................................................................91 4 ACCOUNTING FOR THE COMPLIANC E GAP: INTERNATIONAL RELATIONS.......94 4.1 Introduction .......................................................................................................................94 4.2 Why do States Comply with International Commitments? ..............................................96 4.3 Realist Thought and Non-ComplianceEnforcement ....................................................101 4.4 Neoliberal Institutionalism: Coopera tion and EffectivenessManagement ...................106 4.5 Social Constructivism: a Normative Approach to Non-Compliance .............................114 4.6 Conclusion ......................................................................................................................117 5 ACCOUNTING FOR THE COMPLIANCE GAPREGULATORY FEDERALISM .......119 5.1 Introduction .....................................................................................................................119 5.2 American Federalism and the EU, Comparable? ...........................................................119 5.3 U.S. Regulatory Federalism and the EU .........................................................................127 5.3.1 Centralized FederalismPrincipals and Agents ...................................................130 5.3.2 Bottom-Up Approaches ........................................................................................139 5.4 Conclusion ......................................................................................................................150 6 DATA, METHODS, AND ANALYSIS ..............................................................................152 6.1 Introduction .....................................................................................................................152 6.2 Operationalization of Compliance ..................................................................................153 6.3 Data Selection: Infringements as Biased Indicators of Compliance ..............................158 6.4 The Dependent Variable: Transposition .........................................................................161 6.5 Operationalization of Independent Va riables: Internat ional Relations ..........................164 6.6 Operationalization of Independent Variables: Regulatory Federalism ..........................167 6.7 Econometric Models .......................................................................................................172 6.7.1 Panel Poisson Regression Model ..........................................................................175 6.7.2 Pooled or Population-Averag ed Poisson Regression Model ................................177 6.7.3 Fixed Effects Poisson Regression Model .............................................................179 6.7.4 Random Effects Poisson Regression Model .........................................................183 6.7.5 Negative Binomial Regression Models ................................................................185 6.8 Data Description .............................................................................................................188 6.9 Model Selection ..............................................................................................................195 6.10 Discussion and Conclusions .........................................................................................206 7 CONCLUSION .................................................................................................................. ...218 7.1 Introduction .....................................................................................................................218 7.2 The Unnecessary Simplificatio n of the Goodness of Fit .............................................218 7.3 Domestic Politics and the Return of Administrative Explanations ................................223 7.4 Alarms, Patrols, Sticks, and Carrot s: Enforcement Versus Management ......................233 6

PAGE 7

APPENDIX A TRANSPOSITION DATA, ISSUES AND CHANGES ......................................................240 B FAVORABLE PUBLICS IN TH E EUROBAROMETER SURVEY .................................244 C SALIENCE IN THE EU ROBAROMETER SURVEY .......................................................245 D RANDOM EFFECTS POISS ON WITH TIME DUMMIES ...............................................247 LIST OF REFERENCES .............................................................................................................249 BIOGRAPHICAL SKETCH .......................................................................................................264 7

PAGE 8

LIST OF TABLES Table page 3-1 Closure decisions 1998-2007, by stage reached ................................................................65 3-2 Total number of newly detected infri ngement cases, by year of detection and by origin ........................................................................................................................ ..........76 6-1 Variable definitions ...................................................................................................... ....190 6-2 Descriptive summary statistics of the predictors of non-compliance ..............................191 6-3 Overall, between, and within variati on of the predictors of non-compliance ..................194 6-4 Estimates of Pooled and Pa nel Poisson regression models .............................................197 6-5 Estimates of Pooled and Panel Ne gative Binomial regression models ............................205 A-1 Commission categor ies and sectors .................................................................................240 A-2 Employment and Social Af fairs: 1999 to 2000 changes ..................................................243 A-3 Internal Market: 1999 to 2000 changes ...........................................................................243 B-1 Data sources for Favorable Publics ..................................................................................244 C-1 Data sources for Salience .................................................................................................245 D-1 Random Effects Poisson regres sion model with time dummies ......................................247 8

PAGE 9

LIST OF FIGURES Figure page 3-1 Detection by source trend ..................................................................................................61 3-2 Stages in infringement proceedings ...................................................................................62 3-3 Number of infringement procedures opened within a year, broken down by stage of the procedure and by Member State ..................................................................................64 3-4 Infringements for which proceedings have been commenced, by legal base, 19982007....................................................................................................................................77 3-5 Infringement proceedings by source of non-compliance and by stage reached, 19982007....................................................................................................................................78 3-6 Infringement cases by Member State, 1996-2007. ............................................................79 3-7 Average rate of directive transposition for the EU15, 1998-2007. ....................................81 3-8 Average number of applicable directives for the EU15, 1998-2007. ................................82 3-9 Stage of the infringement procedure reached for infringement cases being under examination, broken down by sector, 1998-2007. .............................................................83 3-10 Infringement proceedings after the Letter of Formal Notice; Environment, Internal Market, 1998-2007. ............................................................................................................84 3-11 Average transposition rates, Environment, 1998-2007. .....................................................86 3-12 Transposition rates in the Environm ent and Internal Market, EU15, 1998-2007. .............86 3-13 Average transposition rate, Envi ronment, all other sectors, EU15 ....................................87 3-14 Average Environment transpositi on rates, by Member State, 1998-2007. ........................88 5-1 Prisoners dilemma of agency and firm choices ..............................................................137 5-2 Salience of Environmental po licy 1998-2007average, EU15. ......................................141 6-1 The policy process in the EU: the actors and openness. ..................................................155 6-2 Total infringement proceedings by source of detection and the Environment sector, EU15, 1998-2007. ............................................................................................................156 6-3 Total number of complaints by Member State, 1998-2007. ............................................160 6-4 Total number of complaints per capita, 1998-2007. ........................................................160 9

PAGE 10

6-5 Trends in non-compliance and Industr y Employment by country during 1998-2007 .....193 10

PAGE 11

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CONFORMING OR MUDDLING THROUGH: EXPLAINING VARIATIONS IN COMPLIANCE WITH EUROPEAN UNION ENVIRONMENTAL POLICY By Ioannis Livanis August 2010 Chair: David M. Hedge Cochair: Amie Kreppel Major: Political Science European Union (EU) member states have often failed to transpose and implement EU directives into their national legal order. While this is the case in most sectors of policy, the environment sector seems to bear the brunt of non-compliance. The danger is obvious. Failure in a less salient sector such as the environm ent could decrease members belief in European solutions for more salient sectors, and if EU policy is not fully implemented by all member states it risks becoming an empty paper with only a slight effect on environmental quality. Most importantly, non-compliance is an expensive and tim e consuming reality that serves to take up most of what the Commission time. Three waves of literature have failed to produ ce an empirically reliable understanding of the determinants of non-compliance, while there is also little theoretical cumulativeness. This study uses transposition rates of EU envi ronmental directives for the 1998-2007 period, distinguishing between the supranational and in tergovernmental charac teristics of the EU construct, which leads to differe nt sources of theory (in partic ular International Relations and Regulatory Federalism), and di fferent explanations of noncompliance than previously envisioned. 11

PAGE 12

12 A time-specific fixed effects Poisson estimator is used to demonstrate that non-compliance is a function of both domestic and EU level influe nces. This study lends s upport to the ability of supranational institutions to induce compliance using management mechanisms. For instance, it finds that the European Comm issions funding of Non-Governmental Organizations increases compliance. However it also lends support to th e strateginess of the decision to comply by member states. Bargaining power in the Council of Ministers increases non-compliance, and so does the existence of powerful industries and cozy regulatory efficiencies. Finally, from a domestic politics standpoint, policy salience and governmental commitment are shown to increase compliance, while corruption and need for regulation are shown to decrease compliance. These findings suggest that the fu ll picture of non-compliance can be attained if theorists take a complementary rather than di sjunctive look into the explan ations of non-compliance, as implementation in the EU is an intergovernmental enterprise and it should be understood in its own terms, always keeping in mind to look both ways.

PAGE 13

CHAPTER 1 INTRODUCTION 1.1 Introduction The European Union is the worlds largest and richest single consumer market in the World, with a total gross domestic product that exceeds that of the United States, as well as that of the North American Free Trade Agreement, and accounts for 40% of the worlds total international trade. As such stab ility and progress in the EU is very important not only for the EU but also for the rest of the world. The founders of the original European Economic Community (EEC), the predecessor of the modern EU, supported the development of a fully integrated union of European states. However, they were pragmatic in their belief that this goal could only be achieved through the initial promoti on of cooperation and integration in the less salient issues of low politics. Jean Monnet and Robert Schuman firmly believed that the cause of political integration would be served better if the Co mmission would promote cooperation and integration in less salient issues of low po litics such as the environment, in hope that the enthusiasm for European solutions would spillover to the more salient issu es such as common foreign and security policy. However, and contrary to the belief that th e EU is an environmental leader in world environmental politics, implementation of EU environmental policy is very much at the sharp end of the EU policy process. The danger is that failure in a less salient sector such as the environment could decrease members belief in European solutions fo r more salient sectors (i.e. foreign policy), and if EU policy is not fully implemented and enforced equally by all member states it risks becoming a paper exercise with only a slight e ffect on environmental quality. Thus non-implementation does not only have negative consequences on the environment, but also on the integration process, which in turn may have negative consequences 13

PAGE 14

globally, as so much of the EUs present and future seems to be intertwined with the rest of the world. These negative consequences of non-compliance in the environmental sector make the need to explain and model this situati on obvious, especially since compliance with Environmental directives is the policy sector riddled with the bigge st deficit (by far). The European Commission has created a comp lex mix of enforcement and management mechanisms to induce compliance that practically forms a ladder of patrolling. From simple firealarm mechanisms through citizen complaints to more complex enforcement oriented mechanisms that entangle the European Court of Justice (ECJ), to the most extreme deterrence mechanism of financial penalties. However, non-co mpliance is still a possibility, and initial noncompliance a definite reality. An expensive, time consuming, and alienating reality that serves to take up most of what the Commission does as g uardian of the treaties, instead of more important functions like policy innovation. Although the empirical question of implementati on (or the lack thereof) has sparked the interest of academics and attracted the attention of bureaucrats and politicia ns alike, their various attempts to assess the forces that may influence or shape national responses to EU legislation have been largely unsuccessful. Three waves of li terature have failed to produce an empirically reliable understanding of the determinants of non-co mpliance, while there is also little theoretical cumulation. First, in the late 1980s various rese archers focused on the administrative and legal implications of transposition of directives, but their attempts to assess the forces that may influence or shape national re sponses to EU legislation, were largely a-theoretical. A second wave of researchers offered a more theoretical institutionalist theory of compliance. Based on arguments of costs and appropriateness, they focused on the goodness of fit hypothesis as the key to compliance deficit, which pertains to cas es of high institutional incompatibility between 14

PAGE 15

national administrative practices and European re quirements. But the results were largely mixed and inconsistent. The same issues plague the th ird wave of research, which has focused on the domestic politics of compliance. Arguably this last strand of theorists have exponentially increased our understanding of th e determinants of compliance, but the theoretical murkiness created has left researchers still muddling thr ough competitive, complementary and disjunctive explanations. 1.2 Objectives The main objective of this research is to evaluate alternative compliance models. Special consideration is given to the American literature as a guide to non-compliance modeling; especially the literature s on regulatory federalism and bureaucr atic control. I hope the research will contribute to the academic debate here on regulatory enforcement, as well as to a more dynamic modeling of the EU situation. The specific objectives are the following: Determine what factors are likely to account for differences in member state compliance with environmental mandates. Determine what things can (and have) EU officials do (done) to promote state compliance. What considerations both limit and facilitate these efforts? Determine what guidance (if any) principal ag ent theory can provide to EU officials and those who study environmental federalism in Europe. However, inasmuch as non-compliance can be located in the inte rgovernmental dynamics at play in the European Union, then, insights can be provided from the extensive literature of International Relations. Hence, I will also review prominent approaches in explaining noncompliance in the International Re lations literature, distinguishi ng the various theories according to the assumptions they make about the source of non-compliance and lack of enforcement and the hypotheses those entail (free-ri dership, wrong institutions, etc). My contribution consists in: 15

PAGE 16

a) delineating the analytical bo rders of these theories in te rms of explaining non-compliance, and b) proposing new ways of l ooking at the issue that when co mbined with the existing ones will provide for a better understanding of the issue of non-compliance. As such this research should appeal not only to researchers and academ ics interested in EU politics and policy, but also to academics involved in the fields of American federal and state politics, regulatory politics, environmental poli tics, international relations, and cooperation academics in general. Hopefully it will also help create an inter-disciplinary bridge between the much alienated literatures of Public Administra tion and International Relations while offering a powerful explanation of non-compliance in the EU environmenta l policy context. 1.3 Overview The second Chapter in this study will offe r a review of the prominent EU studies approaches in explaining non-compliance. These are distinguished according to the theoretical (or not) stance they take and th e methodological line they follow From simple administrative explanations, to misfits, to domestic politics, the EU studies literature has made considerable steps in enhancing our theoretical and empi rical understanding of compliance with EU directives. However, several criticisms are offe red, in the later parts of this Chapter, which should help explain the reasons why the EU studies literature has collectiv ely failed to offer an adequately reliable understanding of compliance dynamics. The basic background information and relevant actors in the envi ronmental arena are discussed in detail in Chapter 3. Specifically, this chapter offers a small history of environmental governance in the EU, exemplifying the ways di fferent governance approaches came to be and evaluating whether the EU has kept up with its theoretical commitments to new forms of governance. The second part of this chapter provi des a detailed examination of the institutional actors involved in environmenta l policy in the EU, and the specific compliance mechanisms 16

PAGE 17

employed by these actors. The final part of th is chapter provides an extensive look into the purported compliance deficit with EU policy requir ements in general, and specifically with the environmental gap, its magnitude as compared w ith other sectors, and its importance to the Commission and European Union at large. Chapter 4 in this study follows the literature of international relations as it leads through different conceptualizations about the source of non-compliant behavior. From realist to neoliberal institutionalist to more novel approach es like constructivism, this chapter offers a fresh way to look at the issue of non-compliance, delivering important theoretical insights into the ability of power, supranational institutions, an d domestic forces to influence compliance. The same tactic is followed in Chapter 5. Afte r offering a short introduction into how we can perceive the EU and whether it can be compar ed to other federal states, it goes through centralized and decentralized theories of regulatory federalism, offering such insights as the importance of salience and bus iness climate, among others. After offering the conceptualization of co mpliance as the initial stage of policy implementation, Chapter 6 of this study provides an overview of the issues and faults of infringements as the dependent variable in non -compliance research and offers support for the use of transposition rates as the dependent variable. The next two sections cover the operationalization of the independent variables, while the chapter also provides an overview of the different econometric models available give n the count data in th is study. After a short discussion of the attributes of the employed data, the econometric model of choice becomes evident through the use of severa l statistical tests. The timespecific fixed effects Poisson estimator provides the best fit for the dataset of this study, and it is used to assess the empirical strength of the hypothesized relationships in the last section of this chapter. 17

PAGE 18

18 Finally, the conclusion to this study offers a further overview and elaboration on the specific contribution to both theory and methodol ogy offered by this study. Specific attention is paid to the goodness of fit hypothesis that has so far dominated the EU studies literature, making sure to illustrate the faults and spuriousness of the hypothesis using the results of this study. A further discussion is also offere d on the slipups of domestic po litics theorists pertaining to theoretical and methodological issues The last part of this chapter provides a discussion of the enforcement and management approaches to non-compliance as they have developed in both the international relations and regulat ory federalism theories, while it also discusses the limitations of this study and makes several s uggestions for future research.

PAGE 19

CHAPTER 2 THE EU IMPLEMENTATION RESEARCH 2.1 Introduction While European legal and economic integration has made great strides since the mid 1980s, it has exposed a major deficiency within the EU policy-making system. Implementation of EU policy, and more specifically in this analysis environmental policy, is very much at the sharp end of the EU policy process (Jordan, 2002). Non-implementation of EU directives and regul ations, even in low salience policy arenas, is potentially of great importance in terms of the overall functioning of the EU and its significance in world politics. Th e potential scale of the negative consequences of noncompliance highlights the need to investigate, model and explain patterns of implementation and non-implementation in the EU, beginning with low saliency policy sectors. Although the empirical question of implementatio n (or the lack thereof) has spar ked the interest of academics and attracted the attention of bur eaucrats and politicians alike, thei r various attempts to assess the forces that may influence or shape national responses to EU legislation have been largely unsuccessful for different reasons. The purported compliance gap already attracted academic attention as early as the 1980s. The stepping-stone was set was in 1986, when legal scholars Krislov, Ehlermann and Weiler drew attention to the growing problem of compliance (Krislov et al., 1986), while Siedentopf and Ziller (1988), analyzed the implementation of seve nteen directives in the twelve member states. However it was not until the early 1990s that the European Commission focused more on the problem. This was partly due to the advent of the Single Market Program that acted as the catalyst for both Commission atte ntion and the advent of impl ementation studies in the EU context. Researchers have tried to fill-in the gap of compliance in EU policy implementation, 19

PAGE 20

and according to their empirical and theoretical focus they can be categorized into three broad waves of research. This chapter offers an introduction to the Eur opeanization literature, as it has developed under the rubrics of Inte rnational Relations, Comparative Po litics and Public Administration, without necessarily making a distinction of who be longs where. This is because, in their attempt to complement, fix, or expand on previously empirically disconfirmed hypotheses about the source of non-compliance, theorists have moved seamlessly from simple administrative explanations, to adding interna tional relations explanations, to more complex comparative politics ones without regard for model parsim ony or theoretical source. Sometimes, even competing theories are used together to explain variation; management and enforcement theories of compliance are used under the same theory and top-down mechanisms are inter-changed with bottom-up ones. Most of contributions emanate from the disjunction between vol untary and involuntary source of non-compliance as found in the interna tional relations literatur e. Non-compliance can be voluntary (cost-avoida nce) or involuntary (lacking capacit y) and the accompanying logic for influencing this behavior can be enforcement (f or voluntary), or manageme nt (for involuntary). These two dominating perspectives about source and solution are commonly referred to as the enforcement and the management approach (Chayes and Chayes, 1995; Downs et al., 1996). Under the enforcement approach states are conc eived as rational actors that weigh the costs and benefits of alternative c hoices when making compliance deci sions in cooperative situations. Enforcement approaches assume that states vi olate international norms and rules voluntarily because they are not willing to bear the costs of compliance (Borzel, 2002). As such, hypotheses pertaining to the understanding of adjustment co sts fall under this rubric. States will always 20

PAGE 21

choose non-compliance when the benefits of shir king exceed the costs of detection. It becomes clear then that compliance probl ems, under this approach, are best remedied by increasing the likelihood and costs of detection through monitoring and the threat of sanctions (Tallberg, 2002). In contrast, the management approach assumes that states are in pr inciple willing to meet previously agreed upon international commitments but simply lack the wherewithal to do so (i.e. the material resources, technology, expertise, administra tive manpower, financial means, etc.), or are simply confused about their required role due to the ambiguity of international rules. By consequence, non-compliance is best addressed through a problem-solving strategy of capacity building, rule interpretation, a nd transparency, rather than through coercive enforcement (Keohane, Haas, and Levy, 1993; Janicke, 1990). Of course, there is nothing wrong with using all theories or some theories, as they most likely represent different understandings and more than likely look at different aspects of noncompliance. However, the theoretical underp innings of each of the following waves of implementation research must be decomposed, to gain both theoretical and empirical parsimony, while also correcting for the mishaps that happ ened along the way of th eoretical and empirical construction. 2.2 An A-Political Beginning The first wave of scholarship seeking to fill the hole of non-comp liance lacked a strong theoretical framework and focused on compliance mos tly as an a-political process, with legal or administrative barriers in the way of governme ntal will. The main inspiration came from the top-down school of policy implementation that focuses on uniform measures, which must be satisfied in all locations, using a great deal of top-down influence and deterrence(Pressman and Wildavsky, 1973; Bardach, 1977; van Meter and van Horn, 1975; Sabatier and Mazmanian, 1981; Mazmanian and Sabatier, 1983). From this pe rspective, administrativ e barriers comprised 21

PAGE 22

of such variables as internal coordination probl ems (Krislov et al., 1986), the inefficiency of domestic institutions and corporatism (Lampinen and Uusikyla, 1998), and the lack of resources (Ciavarini Azzi, 2000; Dimitrakopoulos, 2001) Additionally, explanations for failed implementation includes legal variables, such as the national constitutional characteristics (Krislov et al, 1986), the national legal cultur e (Collins and Earnshaw, 1992), and the legal complexity and poor quality of directives (Krislov et al., 198 6; Weiler, 1988; Dimitrakopoulos, 2001). However, most of the first wave studies failed to make a distinction between the transposition and the ap plication/enforcement of European directives. Ciavarini Azzi (1988, p. 199) postulates that Community la w, once it has been incorporated, is applied neither better nor worse than national law, which, of course, falsel y assumes that implementers are unaware of the European origins of the law to be transposed. This created an array of counterarguments and some scholars went so far as to characterize compliance as post-decisional politics (Puchala, 1975; Collins and Earnshaw, 1992). It is argued th at governments wish to appear as good Europeans by agreeing to a European directiv e knowing that policy will be eroded at the lower/non-visible channels of state implementation due to the intergovernmental nature of the implementation system (Jordan, 1999). Yet, recent quantitative invest igations of the various lega l and administrative variables suggested over time, have been largely inconc lusive. Lampinen and Uusikyla (1998) found that efficient domestic administrative institutions and political culture play a great role, while Demke (2001) found that organizational, legal and t echnical resources are more important. Mbaye (2001) concludes that there is a positive effect between politic al power and non-compliance, and Mastenbroek (2003) finds administra tive and legal variables to be of credible importance in the 22

PAGE 23

Netherlands. The absence of a political concep tualization of the implementation process might have something to do with the administrative a nd legal theoretical under pinnings of the authors in this first wave of research, but as we will s ee later on in this chapter, implementation research has returned back to these a-political explanations in their attempt to complement failing (empirically) theoretical propositions about th e importance of institutional structures. 2.3 The Misfits The relaunch of Europe in the 1990s (Single European Act) and the resulting revival of grand integration theory encouraged compliance re search to take on a more theoretical stance. Several theorists dubbed neo-institutionalists focused their atte ntion on EU institutions to determine whether they are independent of member state control or not, seeking to explain the degree of influence of the EU policy-making on member states. This which goes a long way in settling the debate betw een supranationalists an d intergovernmentalists1 but narrows the debate to just the impact of EU policies rather than the mechanisms of successful implementation. Focusing mainly on environmental policy, the key hypothesis was that regulatory policies are prone to have administrative impacts and that in cases of high institutional incompatibility between national administrative practices and European requirements, member states will have trouble implementing European directives. This e xpectation was premised on the assumption that implementation would require fundamental changes of core administrative structures (Knill and Lenschow, 1998; Heritier et al., 2001; Borzel, 2003) But generally scholars make a distinction between institutional and policy misfit, with the po licy dimension relating to the content of the 1 Supranationalists and intergovernmentalists hold in common a respect for institutions as shapers of human behavior. Their main disagreement is at what constitutes an institution and the degree of their independence from national principals, with supranationalists affording great influence and independence to institutions of the EU. Proving that institutions act independently from their political principals (the states), would mean that the EU has moved to a new supranational state of affairs, where institutions (agents) in fluence the states (principals). 23

PAGE 24

policies, while the institutional dimension relate s to the regulatory style and administrative practices in a particular po licy sector (Heritier et al ., 2001; Borzel and Risse, 2003). One of the most basic strengths of this approa ch is that it is theo retically rigorous while also being focused on empirical confirmation. However, in this latter strength emanates its largest fault. Empirical investig ation of the hypothesis has been le ss encouraging th an its promise to solve the debate between s upranationalists and intergovernme ntalists. Various empirically oriented researchers, focusing primarily on en vironmental policy, investigated the hypothesis empirical grounds; unfortunately the results were rather dispiriting. Knill and Lenschow (1998), in an analysis of compliance focusing exclusively on the goodness of fit on four environmental directives in Germany and the UK found that their hypothesis was validated in three of the eight cases at hand. Similarly, Haverland (2000) analyzing th e implementation of the Packaging Waste directive in Germany, the Netherlands, and the UK found that the country with the greatest misfit (the UK) adopted more successfully than the co untry that needed only incremental adjustments (Germany), while the latters record was ev en worse than the Dutch, despite the higher adaptation pressure for them (more incompatibility ). The reason for this unexpected result is traced to the existence of institutional veto points (institutional structures th at afford players with the ability to modify and block legislation). While the UK had to dramatically increase its recycling levels and introduce bindi ng legislation which run counter to its tradition of negotiating bilateral solutions, the British i ndustry did not have an effectiv e veto point in contrast with Germany, where the Bundesrat ra ged a two-year battle with the German government. Mastenbroek and Van Keulen (2005) found that the political will of the government was more important than the goodness of fit in expl aining the timeliness of transposition in two 24

PAGE 25

internal market directives in the Netherlands, which points toward the ne ed to bring domestic politics back in explaining complia nce variation, mainly along th e lines of the first wave of compliance research. Finally, even more evidence against the goodness of f it as a stand-alone hypothesis has been offered by Falkner et al (2005), who, in a comparative study on the implementation of six labor directives, report a lack of confirmation, as only 22 % of their cases on non-compliance could be attribut ed to fitness. The authors argue that we must take into account the role of domestic po litics (also see Treib, 2003), an d the culture of compliance (Falkner et al., 2005) of the different member states. It is clear that the main weakness of this se cond wave of research was that the preferences of domestic actors remained largely un-theorized and while theoretical and empirical rigor was offered, empirical confirmation was remote and it accentuated the need to offer an explanation for the deviant cases. The misfit argument was in principle based on th e insights of earlier research on EU decision-making (Heritier, 1995 ), which postulated that EU Member States attempt to export their policy-making attributes to the EU. This of course implied that Member States will try to protect their administrative and legal traditio ns by shirking the implementation of EU legislation. Member States are seen as guardians of the status quo, as the shield protecting national legal-administrative trad itions (Duina, 1997, p. 157). As a consequence governments who failed to upload their own policie s to the EU level woul d try to resist during the downloading process, when the agreed-up on measures were to be implemented (Borzel, 2002). This argument makes a case for the enforcem ent camp of international relations, which postulates that states will resist what does not match their preferences. However, it does not take into account the effect of power in the ability to either upload legislation or shirk 25

PAGE 26

implementation. This study argues that power (d efined as bargaining power in the Council of Ministers or membership in the European Pa rliaments Environment Committee), will mediate the misfit of policies. In additi on, as Treib (2003) argues, it is possible that national actors and Member States may want to change existing polici es and institutions, and they may even use EU venues to accomplish just that. Hence, even the most basic assumption of the misfit theorization can be challenged as misfit may not always be due to an inability to upload but due to the presence and preferences of domestic po litical forces that make or break promises to the EU. A final weakness of second wave of literature ha s been the inability to distinguish between factors that influence transposition and a pplication/enforcement of legislation. Most contributions treated the implementation proces s in a linear manner that tends to ignore the different actors involved and the different processe s that take place in implementation. The misfit of European requirements to administrative traditions refers to the application/enforcement stage of policy implementation of EU po licies by the states. This is not to say that administrative agencies are not directly invol ved in the transposition stage of implementation but transposition is the stage where states pick the right instru ments to implement directives, through a process closely resembling policy formulation (even if the goal is handed down by the EU directives), which entails participation and politics. As such, most goodness of fit explanations need to be supplemented with political explan ations that take into account the presence of different actors in this stage. Or more to the poi nt, that take into account the pres ence of the biggest constellation of actors possible in any of the implementation stages. 2.4 Politics are In These last realizations, along with the limited explanatory value of the misfit hypothesis, have led most goodness of fit advocates to su pplement their models to account for domestic 26

PAGE 27

political influence. Knill and Lenschow (1998) suggested the inclus ion of the degree of embeddedness that national institutions experien ce, while the explanatory value of the policy context becomes important only in situations where the level of institutional embeddedness suggests the more ambiguous pictur e of moderate adaptation pressure and it is with this picture that such notions as policy salience become important (K nill and Lenschow, 1998, p. 611). Borzel (2003), with her push-pul l model incorporates policy mi sfit (as an initial reason for noncompliance), mobilization of domestic actors pr essuring for implementation, and pressure from above where the Commission may initiate in fringement proceedings, as the venues for effective implementation. Knill and Lenschow (1998), treat policy salience as a mediating factor or a secondary casual factor to the constraints created by the institutional framework in which the policy will take place. This study argues that, in cases where th e salience of policy is high, then this will be enough to attract the atten tion of the political principals of agencies, and make them pressure agencies to treat the environmental issue at hand as a crisis and thus w ith an added degree of professionalism, hence increasing compliance. More to the point, the salience of an issue to private actors will create the need for attention to agency workings from politicians, judges, and journalists, which of course in tu rn pressure agencies to be mo re efficient and effective with legislation, regardless of the existence of veto poi nts (more on this below) or change agents. In this case, it makes sense that the existence of parties, which are favorable to the proposed policy change, in legislatures will have a positive eff ect on compliance, since it increases salience and creates a direct link with institutional players. Along these lines, Treib (2003; 2004) argues that the preferences of political parties holding sway over the transpositi on of EU law are important, and that governments may accept 27

PAGE 28

wide-ranging deviations from the status quo if the direction of the re quired reforms is in line with their political party preferences. Similarly, in an empirical case study on the transposition of two directives with significant mi sfit in the Netherlands, Mastenbroek and van Keulen (2006, p. 38) showed that favorable government preferences may work wonders in overcoming misfit. Markus Haverland (2000) takes a different stance and makes a case, using the Packaging Waste directive, for institutional veto points a nd their effect on implementation. The number of institutional veto points centra l governments have to face when imposing European provisions upon their countries tends to shape the pace and quality of implementation regardless of different degrees in goodness of fit. In th is study it is made evident that even when a country has the appropriate institutional arrangements it needs to pay attention to its domestic political conditions as they will decide the quality and timeliness of the implementation, which has been further researched by Mbaye (2001) and Giuliani (2003), who both i nvestigate the effect of the number of veto players. A similar strategy is developed by Risse et al. (2001), and Borzel and Risse (2003). Even though misfit is still a neces sary condition, it is not the only one. Other mediating factors such as the number of veto poin ts in a political system, will make it harder to comply. Also, the political and organizational cultur e will affect whether do mestic actors can use adaptation pressure to induce structural change. Along slightly different lines, Dimitrova a nd Steunenberg (2000) pr esent a spatial voting model of transposition in which various dom estic veto players have to co-operate on transposition. Another line of res earch, formalizes earlier insights which stipulate that domestic politics matter for compliance (S teunenberg, 2004; 2005), and explic itly models the process in which domestic actors must co-operate to transpose a directive into national law (Mastenbroek, 2005). Corporatism is also back in the limeli ght (Lampinen and Uusikyla, 1998; Mbaye, 2001; 28

PAGE 29

Kaeding, 2006). As the argument goes, a high level of corporatism will have an adverse effect on the level of veto players and thus result in increased compliance (Lampinen and Uusikyla ,1998). Conversely, a close and cooperative arrangement between the state and interest groups will increase compliance while an increased interest group involvement (level of pluralism) will lead to non-compliance (Konig and Luetgert, 2008). Fi nally, the level of pa rtisan conflict and polarization within domestic veto players (T reib, 2003; Kaeding, 2006; Konig and Luetgert, 2008) are also hypothesized to affect compliance at the domestic level. A caveat should be offered on the presumed negative impact of veto points on compliance. The existence of veto points should not necessa rily mean that the process of transposition becomes slower. Even the wealthiest and most powe rful of actors (the industry), may not always be opposed to policy that goes against its prefer ences (Wurzel, 2002). For instance, industry may actually be favorable to environmental legislatio n especially in green countries, and push for Europeanization of state environm ental regulations to level the playing field with competition in other states. Hence, the business climate of a st ate should be taken into account as the industry represents one of the stronger players, but the ability of players to use veto points to stall compliance, will also depend on the receptivity of the bureaucracy to behavioral modification (Wood, 1988). Interestingly, regardless of their empirical ri gor, the theoretical insights of the third wave quantitative studies have been rath er inconclusive (as was the case with most misfits). First off, it has been argued that support for European integr ation is an important f actor that facilitates compliance (Mbaye, 2001: public support; Linos, 2006: support by government parties), while others do not (Lampinen and Uusikyla, 1998). And some have even found st atistically significant negative correlation between thes e two variables (Borzel et al., 2004). However, it can be also 29

PAGE 30

argued that support for the EU might not be evident in all polic y domains. After all, it would take a lot of entrepreneurship to create the required linkages between policy domains to achieve normative compliance in all sectors of policy. Cons equently, decisions to comply in one area must not be confused as normative acceptance of compliance in all policy areas as there is a considerable amount of linkage required to di sseminate normative compliance to other policy sectors, which have actors with different interests, ideas and beliefs (Haas, 1998). Second, studies have found that the structural properties of domestic politics, such as the number of veto players, have a significant impact on complia nce (Lampinen and Uusikyla, 1998; Giuliani, 2003; Linos, 2006; Kaeding, 2006; Perk ins and Neumayer, 2007), while others do not (Mbaye, 2001; Borzel et al., 2004; Borghetto et al., 2005), in fact some find an insignificant negative relationship between veto players and non-compliance (i.e. the existence of veto players increases compliance, Mbaye, 2001). Corporatism is found to be insignificant while also shown to have a negative effect on compliance (contrar y to the hypothesized relationship, Mbaye, 2001) or a positive effect on compliance (Kaeding, 2006).2 Pluralism and partisan conflict is shown to increase non-compliance (Konig and Luetgert, 2008), while Kaeding (2006) finds a positive yet insignificant relationship. This back and forth in empi rical confirmation might have to do with comparing different fields of policy or with indicat or strength, but this study offe rs a different explanation. The literature on veto points, and all the subsequent mediating factors, focuses mainly on the ability of citizen groups to use institutional veto points to affect compliance at the domestic level. The ability and possibility of citizen groups circumventing national politics and using the Commissions formal monitoring mechanisms to h it the fire-alarm has been neglected by the literature. Hitting the fire alar m initiates the EUs formidable enforcement mechanisms and 2 For a discussion on quantitative measures of corporatism refer to Kenworthy (2000). 30

PAGE 31

serves to induce compliance from above (in whic h of course the domestic players played a big role). It stands to reason, that veto points or other mediating factors that involve citizen groups are not as important in inducing compliance as the ability of those groups to utilize top-level monitoring and enforcement mechanisms to induce compliance. The only factors that seemed to find support in most quantitative analyses up until 2007 were the various aspects of ad ministrative capabilities (Mbaye, 2001; Borzel et al., 2004; Linos 2006; Borghetto et al. 2005; Berglund et al. 2005). Bureaucratic efficiency was thought to decrease non-compliance (also Haverland and Romeijn, 2007), while Perkins and Neumayer (2007) find that efficiency has an insignifi cant positive effect on non-compliance. The EU studies literature has focused on the effect of governmental capacity on compliance, but the ability of the EU to affect compliance by increas ing this capacity has be en neglected. One would think that since administrative capacity has so far been researched as a determinant of the ability to comply by various quantitative and qualitative studies, that the ability of the Commission to provide states with such management mechanisms would also have been addressed, especially since capacity is, as theorized by the manageme nt camp, one of the major sources of involuntary non-compliance. Additionally, it may also be the case that administrations are efficient in doing what they do, and states do have the capacity to implement legislation. Regardless of the existence of veto points, and the degree of fitness between European requirements, it can be argued that some states are simply focused on doing different things (that serve their preferences), and are, indeed, very efficient in doing those. For instance, it is true that some states (e specially the poor ones) are more focused on private sector development rather than environmental protection; hence their bureaucracies will be efficient in promo ting private sector development rather than 31

PAGE 32

environmental protection. Even if these two goals are not necessarily mutually exclusive, the efficiency focus of the administration and its capacity is pointed toward different goals and values. It should be then more important, to investigate how administrations internalize new values than whether their struct ure precludes change (as with the goodness of fit argument). In this case, the effect of political principals and state commitment should be problematized along with the existence of po werful lobbies that keep the focus of bureaucratic efficiency closer to their own preferences. Under the sp ecific context of this study, we should expect to find that when states are focused toward private development and a cozy relationship exists between business interests, bureaucracies and political principals alike, then compliance with environmental policy will be less (of course, corrup tion should also be impor tant in this respect). Others have abandoned the goodness of fit hypothe sis altogether and claim that the nature of European policies (degree of prescription and flexibility, market -making or market-correcting) combined with the variance in domestic constellations (the degree of lib eralization, a countrys reform capacity, and its dominant belief system) account for the variance in compliance (Heritier et al., 2001). EU compliance is increasingl y modeled along the lines of sociological institutionalism. Such studies start from the assump tion that a rule will be complied with if it is deemed appropriate by the stakeholders (Mastenb roek, 2005). It is argued that compliance will not be automatic and will dependent on a process whereby the rule becomes internalized through socialization, persuasion, or learning (Finne more and Sikkink, 1998; Checkel, 2001; Risse, 2000; Sending, 2002). This approach argues for the i nnate notion of the culture of compliance, which holds that member states differ in their basic proclivity to comply with EU requi rements (Tallberg, 2002, p. 619; Falkner et al., 2005). For instance, a group of scholars who analyzed the implementation of 32

PAGE 33

six directives from the field of EU social policy in the fifteen old member states illustrate that arguments, such as the misfit or veto points hypoth eses, or the first-wave focus on administrative and procedural factors, do not hold across their cases. Rather, they propose that a complex web of administrative, institutional and actor-based fa ctors affects transposition outcomes (Falkner et al., 2002; 2004; Falkner et al. 2005, pp. 277-316). However, these propositions are not unlike mo st quantitative and qualitative studies. The major contribution here is the typology (grouping) of countries according to individual characteristics they may possess that make them stand together in compliance or noncompliance. This typology consists of three world s of compliance. The firs t of these, the world of law observance consists of the Nordic coun tries. Obedience differs. Not all Europeans are equally law-abiding citizens (Waarden, 1999, p. 96). In Denmark, when an act is issued it is obeyed, even if one has opposed its adoption and disagrees with its content, (Biering 2000, p. 959). The presence of a culture of respect for the rule of law among political and administrative actors usually ensures fast and correct transposition (Falkner et al., 2005, pp. 317-341). However, transparency and efficient organizatio n of the administration also help the Nordic countries, to react more readily in compliance conflicts (Sverdrup, 2002; 2003). Greece, France or Portugal are members of the world of neglect, wh ere the absence of a compliance culture in both the political and administrative sy stems leads to long phases of bureaucratic inertia and rather apolitical transposition pr ocesses (Falkner et al., 2005, pp. 317341), which means that administrative factors should be of particular import ance for these states. Finally, in the world of domestic politics a third group of states (Germany, Netherlands, Ireland, UK), may have efficient ad ministrations but the absence of a culture of compliance, means that the transposition of EU law will depend on the fit between European requirements 33

PAGE 34

and domestic preferences (Falkner et al., 2005, pp. 317-341), which means that powerful actors should have a major imp act in these states. However, to group states according to the most dominant characteristic they exemplify is a rather deterministic approach that neglects the value of both the top-down and bottom up influences on compliance. To say that Greece, France or Portugal will always behave the same, regardless of the existence of top-down mechanisms to induce compliance from the Commission (like monitoring, or enforcement), is to miss half of the explanatory power that these top-down mechanisms offer. A typology li kes this seems to suggest that any such mechanisms will be inefficient in raising gove rnmental concern, and thus compliance, through the use of non-governmental actors (which are assumed to be abse nt in these countries). This study argues that both top-down and bottom up mechanisms will be effective in determining compliance, while the presence of domestic actors will be the biggest qualifying factor of success in all member states, regardless of the prev ailing legal culture or ad ministrative style, or veto points. To be precise, the presence of domestic actors or their relative power to affect results can be influenced by supranational institutions as well, and not only by the domestic legal traditions and institutionalized structures. The identification of a Nordic world of comp liance ties in with the findings of Sverdrup (2004) and with a recent study by Perkins and Ne umayer (2007) who found that compliance is significantly different in the Nord ic states than others. However, the later study grouped states according to their legal origin3 which distinguishes between E nglish origin (UK and Ireland), French origin (Belgium, France, Greece, Ital y, Luxembourg, Netherlands, Portugal and Spain), German origin (Germany, Austria), and finally Scandinavian origin (Denmark, Finland and Sweden). As such the above typology of Falkner et al. (2005) is miss-specified as to who belongs 3 The data is available through La Porta et al. (1999, pp. 268-275). 34

PAGE 35

where when it comes to a respect of law and the existence of a compliance culture. To be fair, the Nordic states do have a better transposition record but so do the Netherlands and Austria, and according to Perkins and Neumayers (2007) findings the typology does not hold-up when it comes to the rest of the worlds as countries sl ip in and out of worlds according to their legal origin. Hence, theoretical and empirical investigation, it seems, is not without a sense of irony. At the same time as qualitative studies in the thir d wave have increasingl y accepted the domestic political dimension of compliance, the results of quantitative research seem to point back to the arguments of the early a-politic al research that stressed the importance of efficient and well coordinated administrations (Treib, 2006). 2.5 Quantitative, Qualitative, and Depe ndent Variables: Faults and Caveats A remarkable feature of existing implementa tion research is that neither these novel approaches nor the goodness of fit hypothesis (much less the firs t wave of research) have the ability to explain non-complia nce given their adherence to theoretical and methodological ambiguity respectively. Even t hough the adherents of the go odness if fit hypothesis where methodologically rigorous, their em pirical results were disappoin ting largely due to the use of easy-to-measure variables. These miss the import ance of more interest ing ones, like domestic politics, and for which domestic politics theorizers can account more effectively. From their side, domestic politics (third wa ve) adherents are theo retically rigorous but their methodological stance leaves little room for empirical reliability. The strength of most domestic politics designs is that they allow for controlling key domestic variables, that are broadly agreed to be major determinants of compliance, yet they encounter several shortcomings. First, both qualitative and quantit ative studies seem to suffer from selection bias, as they almost invariably exclude states from the analysis. Al though some qualitative re searchers have covered 35

PAGE 36

all member states (Siedentopf and Ziller, 1988; Demmke, 2001; Falk er et al., 2005) most quantitative and qualitative studies do not. Qual itative researchers tend to focus on countries with a bad record of compliance (see UK), or countries with supposedly good records (see Germany), and disregard such countries as Fran ce (but see Falkner et al., 2005). While small countries such as Austria, Finland, and Sweden, which, as we shall see later on, possess the best compliance records on environmental policy in th e EU are left out of both qualitative and quantitative studies. This could lead to inaccura te conclusions supporting the non-existence of compliance problems thesis (Borzel, 2001; 2003), as spatial variation is more evident if we include the aforementioned compliers in envi ronmental policy specifically. This selection would not present a considerable disadvantage fo r the generalizability of their findings if the reasons for selecting those specific states wher e explained, but in most cases they are not. From their side, most quantitative studies e ither use data from before the 1995 accession, or use data up to 2004 (see Kaeding, 2006) that exclude Austria, Finland, and Sweden from the analysis (for a review of the literature, see Mastenbroek, 2005). The exclusion of the environmental leaders in quantit ative studies creates a gap in the possible hypotheses, as there is much to be said about the pr esence of these member states in European political institutions (like the European Parliament Environment Committee). Any study wishing to include these countries, though, must take into account the newcomer effect, whereby the Commission grants a period of grace to newly accessed count ries for approximately 2 years (Svedrup, 2004; Perkins and Neumayer, 2007). However, the biggest problem does not come from between the qualitative and quantitative divide; it comes from within the quantitative studies. The most important development, in the above third wave of research, was that author s started conducting predominantly quantitative 36

PAGE 37

studies. The Commissions infringement proceedings against me mber states datasets were widely used to gauge noncompliance with EU law (Mbaye, 2001; Borzel, 2001; 2003; Tallberg, 2002; Borzel et al., 2004; Sverdrup, 2004; Beach, 2005) A second strand of quantitative studies used the transposition measures4 that member states officially notify to the Commission (Lampinen and Uusikyla, 1998, Haverland and Romeijn, 2007; Konig and Luetgert, 2008), sometimes also in combination with infringeme nt data (Giuliani 2003), while in a few cases, researchers complement transpos ition data with nati onal legislative sources (Mastenbroek, 2003; Kaeding, 2006). As indicated by the rather inconclusive result s described above, quantitative EU implementation research seems to be ridden with problem s, almost invariably having to do with the dependent variable. First, there is simply no way to know whether the Commission, for whatever reason (resources, strategic-ness), responds to all infringements with the same fervor. Put differently, the Commissions unwillingness or inab ility to monitor and enforce all infringements of EU law introduces a bias in th e dependent variable, and this type of research looks only at the 'tip of the iceberg' of non-compliance (Hartl app and Falkner, 2009, p. 292) Further complicating the reliability of infringements as a measure of compliance is the odd practice by the Commission to change counting rules, eliminate categories altogether fo r years and then bring them back. For instance, complaints were repor ted from 1982-1991, then lumped into a category that doesnt signify whether it is complaints of not from 1992-1997, and then in 1998 complaints reappear with the addition of own investiga tions and non-communication (Borzel, 2001). This makes the use of pre-1998 data highly unreliable, especially for infringement data. For example infringements by state and sector were onl y reported for 1988-1992, while since then 4 These rates, which represent the share of transposed direct ives against all applicable directives at a certain period of time, are regularly reported in the Commissions annual reports on monitoring the application of Community law. 37

PAGE 38

infringements are reported per me mber state or per se ctor, making cross country comparisons in the same sector of policy practically impossible.5 To be fair, transposition rates come with thei r own problems. The data is consistent when it comes to years with only a few changes needed as the Commission switches names on sectors and categories within them. The biggest problem is that the date of notification may differ from the date of actual transposition. Th is should not be an alarming i ssue as states may complement their early transposition with additional measur es. Some researchers have even attempted to complement notification data w ith national legislative sources (see Mastenbroek, 2003; Kaeding, 2006), however under the understandi ng of compliance as conformity offered by this study this is not necessary, as a states notification of implementing measures is enough to designate compliance. A more important but equally not alarming issue with transpositions as indicators of compliance is that the data do not contain an o fficial statement different iating between timely and correct transposition (Hartlapp and Falkne r, 2009). The issue here is that, data do not provide any information on whether the timel y notification of measures implementing a directive, are the correct measures. To worry about correct ness though is to introduce the same selection bias to the data as with relying on the Commissions data on infringement proceedings, as explained earlier.6 All of the aforementioned shortcomings are c onsiderable, but not forbidding, especially if we correct for some of the biases mentioned here. It seems that most of the empirical disconfirmation of otherwise credible hypotheses in all three waves of EU implementation research has to do with two recurring issues. The fi rst, is the aversion of theorists to deconstruct their theoretical underpinning and reconstruct them under the correct premises. And the second is 5 Requests for original datasets to extract per state an d per sector data to both the Commission Secretariat and Environment DG remain unanswered for 6 months now and counting. 6 For more on this and other issues, please refer to chapter 6 of this study. 38

PAGE 39

the aversion of quantitative studies to include important states. More to the point, a large N quantitative study including all the old member states for a seri es of years, has never before been offered on environmental transposition rates. As to the first shortcoming, because most studies are based on different policies, methodologies, and samples (states), it is very di fficult to draw comparable conclusions from them. Even though, the third wave of research ha s added considerably to our understanding of non-compliance, most approaches ar e rather disjunctive than comp lementary to each other. It becomes evident then that we need to model the compliance deficit facing EU policy using a more theoretically integrative and methodologically rigorous design. In this respect I intend to offer a theoretical model of compliance that obt ains the benefits of all three approaches (administrative, goodness of fit, domestic poli tics) and adds control over the potentially important influence of domestic politics as conc eived under the rubrics of International Relations and Regulatory Federalism. Theoretical rigor, and in the e nd, theoretical integration can be achieved if this research speaks to the intergovernmental -supranational divide that lurks in the background of noncompliance scholarship. If the EU is an intergovernmental organi zation then realist theoretical reflexes command the abolition of intra-state res earch. On the other hand, if the EU has more supranational elements than a simple intergovern mental organization, then neo-institutionalist, and domestic politics reflexes command an in vestigation of the effects of institutional structures and domestic opposition. To pro ceed we must, therefore, also identify the intergovernmental and supranationa l characteristics of the EU cons truct. In particular, determine to what extent the EU is the re sult of normal international relations and to what extent it should be understood as a new domestic polity. 39

PAGE 40

In this sense, if we conceive the issue of noncompliance in the EU as an internal political conflict between a central authorit y and a set of semi-autonomous s ub-units; then models of state implementation of federal policies drawn from the American context (in particular theories of regulatory federalism and bureaucratic control), may be particularly useful in explaining domestic patterns of non-compliance in the EU However, inasmuch as non-compliance within the EU can be understood in term s of intergovernmental dynamics, insights may also be drawn from the extensive international relations literature on state compliance with international agreements, regarding the domestic forces at play. Evidence for one or the other model hardly cons titutes conclusive evidence that it is more important in explaining non-compliance. Adding several variables that represent different theoretical approaches to one full econometric m odel allows for capturing the full effect of all variables when taken together ra ther than disjunctively. The benefit is both theoretical and empirical. Adding all approaches to a single model allows the researcher to determine whether variables capturing predictions might lose thei r explanatory power once variables from other theoretical approaches are included. This allows for an evaluation of whether a theoretical approach actually adds to explanatory power. If this is the case, then the approach should be thought as complementing existing theoretica l approaches (rather than providing and alternative), when attempting to gain a fu ll understanding of the bigger picture on what determines non-compliance. In regards to the second problem, mentioned ab ove on including all old member states; as I have already argued, the turn to quantitative goodness of fit approaches came with the use of rather easy-to-measure variables, which miss the importance of more interesting ones, like domestic politics, and suffer from case selection bias (member states). The use of a large N 40

PAGE 41

quantitative approach offers considerable a dvantages in the present context. Econometric estimation techniques allow us to investigate la rge numbers of cases, and even though the timespan examined by this study is relatively short (1998-2007), it comprises and accounts for spatial variation between all states and e nvironmental policies, rather than just the states that perform worse than the ones excluded in previous st udies. As mentioned earlier, the focus on transposition rates should also serve to provide for a better estimation of what determines compliance (of course as compliance is understood in this study). This estimation will not suffer from additional selection biases introduced when using infringement data that are riddled with several omitted (and almost impossible to measure) variables pertaining to the willingness and ability (resources) of the Commission to pursue infringements proceedings in all cases. This study should therefore yield insi ghts that are potentially more generalizable than small N qualitative studies, and missing N qu antitative studies, while also pr ovide insights more credible than infringement based quantitative examinati ons of non-compliance. This is of particular importance in testing theoretical approaches of co mpliance, where there is an added interest in clarifying whether specific casual relationships hold across the full range of countries. 2.6 Conclusion EU compliance is far from being under-studi ed, yet problems stubbornly persist in our understanding of what affects compliance. Research ers have been fascinated with this topic for the last twenty years. First, in the late 1980s various research ers focused on the administrative and legal implications of transpos ition of directives, but their attempts to assess the forces that may influence or shape national re sponses to EU legislation, were largely a-poli tical. A second wave of researchers offered a more theoretical institutionalist theory of compliance. Based on arguments of costs and appropriateness, they focused on the goodness of fit hypothesis as the 41

PAGE 42

42 key to compliance deficit, which pertains to cas es of high institutional incompatibility between national administrative practices and European requirements. Unfortunately, empirical investig ation of this hypothesis showed that a good fit is neither a necessary, nor a sufficient, condition for unobstructed implementation (Knill and Lenchow, 1998; Haverland, 2000; Heriti er et al., 2001; Falkner et al., 20 05). This disappointing result has sparked a renewed interest in bringing domestic politics back in as an explanation of the implementation deficit that accounts for domestic political influence, yet the challenge is to theorize how and under what conditions domestic opposition plays a role. As such, the domestic forces that may influence or shape national respons es to EU legislation remain largely unexplored. This is also wher e the importance of methods co mes in. Previous attempts (administrative, goodness-of-fit, and domestic polit ics) suffer from an a-theoretical posture, empirical disconfirmation, and select ion bias respectively. Th e goal of this research is to begin to address this lacuna through the development, application and empirical testing of possible theoretical models of EU memb er state compliance, using EU environmental policy as the primary forum of analysis, while also correcting fo r selections biases that exclude, arguably, the best compliers of EU legislation.

PAGE 43

CHAPTER 3 EUROPEAN UNION ENVIRONM ENTAL POLICY: INSTITUT IONS, PROCESSES, AND CURRENT CONDITIONS 3.1 Introduction Environmental policy in the European Union has developed from a state of relative obscurity to a voluminous body of law encompa ssing more than 500 measures and six Action Programs (Jordan, 2000). Surprisingly, the Treati es of Rome (1957) made no mention of environmental policy; there was no environm ent bureaucracy, and no environmental groups active in lobbying the, then, na scent European Commission. Given the absence of a formal treaty base and the subsequent need to secure lowe st common denominator d ecisions, environmental policy before the 70s started slow and followed a process that has been frequently characterized as integration by stealth (Weale, 1999). The Commi ssion had to go to great lengths to justify it because it did not appear to have much to do wi th the EEC's core objective, that of creating a common market in economic goods and services. It wasnt until the member states broadly endorsed the European Communitys involvement in environmental policy at the 1972 Paris Summit, and formally embedded it as a policy domain open for europeanization by the Singl e European Act (SEA), that the environment began to escape its unimportant status. Even so, stealthy practices where the norm for decades and Action Programs to come. Any extension in the legal competence of the European Community (EC) in the environmental field fo llowed the Monnet method of packaging the environment into technical regulations and duping states into deeper integration than they might have otherwise accepted (Wallace, 1996). Environmental Action Programs (EAP) pr ovided a broad route map, while the Commission was left to work up specific proposals packaging them in a low politics technical nature legislation (such as standa rds), that required very little input from the public or from sub43

PAGE 44

national actors. This allowed the Commission, to avoid difficult questions regarding who, and at what level of governance, is responsible for a particular environmental issue. This also allowed for Envir onmental Actions Programs to be sufficiently vague and grandiose in intention, without pr oviding for adequate applicability in the field. Goals set forth by EAPs have only marginally been implemented in reality, if at all (H olzinger at al., 2006). The Commissions opportunism with t echnical legislation served it well in not antagonizing the Member States, and integration by stealth was overlooked by th e states as long as it did not encroach on their sovereignty, or as long as stat es did not realize their sovereignty was being encroached. It is exactly these sovereignty issues that derail ed envisioned EAP changes in strategy and environmental instruments from beco ming realized in the field. And it is exactly these sovereignty issues caused by the ascendancy of the subsidiarity principle that reversed the tide of environmental legislation toward the st ates. Adherence to inte rventionist governance models, which encroach on the national scope of action, could hardly continue to be politically legitimized. So long as the European Union preoccupi ed itself with technical environmental legislation and used a strategy of command-and-cont rol interventionism that did not allow or ask for national input, the states were complacent in their ignorance. However, three decades later, with an impressive body of envir onmental legislation in place, w ith the end of the permissive consensus, and with financial s ituations worsening, an unavoidable maturation in EU politics and conceptions of subsidiarity has occurred, that brought European Union (EU) environmental policy to crossroads regarding it s future. The very content and sc ope of environmental legislation is now being reworked and rethought, at both th e national and EU levels, as a response to the 44

PAGE 45

challenges of global competition and demands for deregulation and re -nationalization of environmental policy. 3.2 A Short History of Environmental G overnance: Instruments and Strategies There have been six main periods of environmental policy heralded by the six Environmental Action Programs and roughly corresponding to three main governance approaches. Idealistic as the begi nning was, it also came with a non-idealistic technical focus on command and control instruments, supporting a highly interventionist governance approach (Rehbinder and Stewart, 1985). Without eliminati ng these command and control instruments, the EU moved to a new paradigm on environmen tal governance during the mid-1980s and early 1990s. This new paradigm was underlined by the ma rket approach to the environment (such as environmental taxes, tradable permits, or risk liability schemes), popular in other areas at the time and corresponding to the Single European Acts dedication to complete the single market. A final approach, prevalent since the 1990s, has been primarily focused on context-oriented instruments that aim to allow discreti on and openness in both policy formulation and enforcement (Holzinger et al., 2006). 3.2.1 Idealistic Beginnings, 1973-1982 In June of 1972, the United Nations Conference on the Human Environment held a conference in Stockholm, Sweden. It was the fi rst conference of its kind to concentrate on international environmental issues. The 116 nations that attended discussed such things as the impact of industrialization on the environment, conservation efforts, environmental poisons, and monitoring environmental protection (Emme lin, 1972). Months following the Stockholm Conference, the European Community held its own gathering: the Pa ris Summit of 1972. It called for the creation of a European Community environmental policy, and one shortly followed by means of the First Environmental Action Pr ogram (EAP) in 1973 which argued that the 45

PAGE 46

protection of the environment belongs to the e ssential tasks of the Co mmunity (OJ C112/1 from 20/12/1973). This program would be followed by continued multiannual programs of the same style, which are st ill present today. Environmental Action Programs are medium-term strategic policy documents that reflect the fundamental elements of contemporary envi ronmental thinking and problem perceptions, as well as strategic policy orientat ion (Hey 2006, p. 18). They are li sts of planned activities, but are not binding, and they help to shape envir onmental policy by highlighting the specific legislation needed to improve th e environmental conditions of Eur ope. Each of them has led to the adoption of several series of directives on pr otection of natural resour ces (specifically, water and air), noise abatement, nature conservation, and waste management (Pinder, 1998). The first EAP identified the objectives and fo rmed the foundation for future environmental policies. The focus was on prevention, reduction and containment of environmental damage; the conservation of an ecological e quilibrium; and the rational use of natural resources (Hey, 2006). It also emphasized the need for a comprehensiv e assessment of the impacts of other policies on the environment but left the environment outside th e influence of internal market objectives, thus rendering the environment as its own field of policy. This had mixed results. On one hand, it served to insulate the environment from intern al market bickering in the initial years and conserved the spirit of optimism for far reaching policy change. On the other hand, eventually, and as the internal market became more crucial, it served to frustrate environmental policy and its relevancy to European future development, as it now had to be argued-in into other areas of policy, in which it was an integral part to begin with anyway. The first EAP also highlighted that envir onmental policy should be created with a preventative, aside from fundamentally correct ive, approach. For this reason, it called for 46

PAGE 47

research activities on the causes and effects of pollutants, as well as the criteria for environmental objectives. In light of the rese arch on these pollutants, corrective measures appeared in the form of direc tives. Several specific products we re targeted by these directives, such as the lead content of fu el and crockery, the toxicity of detergents and paints, and the emissions of vehicles. A number of industries also faced scrutiny, such as the chemical, food, metallurgic, and textile industries. The specific sectors of water polluti on, noise pollution, and the pollution created in energy production were also main foci (Barnes, 2000). Through the first EAP, the European Commun ity assigned itself the role of releasing environmental information to the public to increase their awareness and their personal responsibility. It also took on the role of international actor in the environmental arena, agreeing to cooperate and support the work of the Unite d Nations in particular (EU Environmental Information and Legislation Database). In all of these ways, the second EAP (1977-1981) essentially served as a follow-up to the first, continuing to emphasize the need for corrective measures, with a preventative outlook. However, it expanded on the list of environmental concerns of the first. It placed more emphasis on protection of nature, as seen in new sections dealing with specific protection of fauna and flora. Also, as proposed by the Commission, the second EAP included a new section on Environmen tal Impact Assessments (EIA). EIAs would provide a process of examining the environmental impact of proposed priv ate or public projects or developments. They were legally established in Directive 85/337/EC, but the implementation of this directive was unsuccessful, and so it was put aside in the EC Environmental agenda until 1996 (Barnes, 2000, pp. 35-36). A project regarding environmental labels on products was also introduced in the second EAP, and faced the sa me result. The labels were meant to provide consumers with environmental information about specific products, encouraging them to make 47

PAGE 48

more eco-friendly choices by pur chasing items with minimized pollution and waste created in the fabrication process. Yet, this project wa s not adopted until 1993, and was not as successful as anticipated (Barnes, 2000). In terms of a governance approach, the First and the Second Programs (1973-1981) advocated and used, almost exclusively, interventionist instruments such as emissions and quality standards, but also technical specificatio ns prohibitions and other obligations for water and air (Holzinger et al., 2006). Th e quality objectives for drinking wa ter were very strict those for air could be achieved without strong policy intervention. The first and second EAPs were able to form a solid base for EC Environmental Policy because they outlined how environmental issues should be dealt with in the future. They introduced strict quality objectives for air a nd water, developed a number of framework directives for water and waste, and debvelope d many ideas that would flourish in future programs (Hey, 2006). Their successes lead to th e establishment of a separate Environment Directorate-General (DG) w ithin the Commission in 1981. Through this DG, environmental activism grew because the DG provided a channel for environmental organizations to pressure the Commission to pursue greener policies. Th e main criticism about the first and second programs was that their directives were not influential enough. Their approaches to environmental policy were often restricted to trad e interests or to very general provisions that could not target the main sources of pollution. This inability was largely due to lack of enthusiasm during instances of economic r ecession, from 1975 to 1978 and from 1981 to 1983, which prompted a change of focus toward carry ing out environmental pr otection cost-effectively, and using the environment to solve ec onomic problems (Holzinger et al., 2006). 48

PAGE 49

3.2.2 The Internal Market is Coming, 1982-1987 The third EAP found itself in more difficult economic times than its predecessors, but dealt with them in a different way. Mediterranean enlargements, the aftermath of the Arab-Israeli war crisis and the resulting quadrupl ed price of oil, along with the oscillating exchange rates caused by U.S. devaluation of the dollar and massive CAP spending, made the completion of the Single Market even more urgent (Dina n, 2005). Therefore, the top prior ity of the third EAP was to use market principles to achieve generally accepted environmental objectives. Through integration with the area of EC enla rgement, environmental policy was able to emphasize the importance of candidate countries pollution levels. Access ion agreements of a candidate country could come to a halt if it c ould not implement the ECs environmental acquis. Yet, the economic difficulties in sections of the European Union put the financing of environmental policy low on several national policy agendas, including the agendas of candidate countries. Therefore, the European Comm ission proposed limited funding towards the environment for the first time, through the use of Structural Funds (B arnes, 2000). It also proposed that a financial instrument should be set up as a tool of incentive and a catalyst for action. The former was implemented during the span of the third program, but the latter was not put into action until the establishment of the Fina ncial Instrument for the Environment (LIFE) in 1991 (Dinan, 2005). Due to the economic malaise of the time, the third EAP was explicitly called upon to consider the poten tial risks of environmental legislation on the distressed economic sector. Therefore, it worked to empha size the links between the environment and the completion of the single market. Both harm onized environmental emissions and product regulations would be needed fr om the environmental sector to avoid distortion in industry competition and the creation of non-tariff barriers (due to different national product norms). 49

PAGE 50

Another benefit of this integration would be the expansion of the job market, due to the Environmental sectors creation of less-polluting industr ies (Hey, 2006). More importantly, the third EAP presented a marked difference in governance instruments from intervention to economic incentives. In particular, the Commission proposed applying a stricter form of the po lluter-pays principle in the developm ent of the third and fourth action programs (Hey, 2006). The initial definition, alrea dy included in the first action program allowed the polluter-pays principle to refer only to th e costs of avoiding pollution, while also permitted exceptions as long as they did not impair the general market functions (H olzinger et al., 2006). In the third and fourth action programs, the Commissi on expanded the definition of the principle to include social costs emerging from residual pollution, by specifying new instruments, such as environmental taxes and charges, liability a nd funding schemes, as well as international agreements.7 In regard to directives, the third EAP priori tized the areas of waste avoidance, clean-air policies, and noise and risk management for i ndustrial sites. It propos ed formulating emission limit values for both mobile and stationary sources, and introducing better filter technologies for the reduction of emissions at the pipe. At this time, the member states had great influence over the environment in the EC, and they became more concerned and involved than in the past. The German industries and government lobbied fo r harmonized European emissions control policy, while the Netherlands pushed for a tradition of strategic environmental planning, the UK for setting of environmental quality objectives, a nd the Scandinavian countries for reform of chemical policies (Hey, 2006). Under the third EAP, all of these points were taken into consideration and obtained generally successful results. 7 In the fourth action program, the same instruments are listed, expanded to include state aid, tradable permits, and negotiated agreements with polluters (Holzinger et al, 2006). 50

PAGE 51

3.2.3 The Single European Act and the Fourth EAP, 1987 1992 The Single European Act was the first major revision to the 1957 Treat y of Rome, created with the objective of speeding up the completion of the internal ma rket. It was created in support of the White Paper, a document pushed by Jacq ues Delors that identif ied the 279 legislative measures needed to complete the internal mark et with a proposed deadline of December 31, 1992. The SEA included a section on environm ental policy and included an article on environment in its single market provisions. It identified enviro nmental protection as having an important role in the improvement of quality of life, a nd in the process of job creation. Signed in February 1986, it began to have effect on July 1, 1987, just a few months before the start of the fourth EAPs term. For this reason, it served as a great influen ce on the fourth EAP. Environmental policy was more frequently mentioned in Council conclusions, and environmental standards were integrated into the policymaking process of the Commission. The Commission also began to work specifica lly into the areas of atmospheric pollution, marine pollution, waste management, biotec hnology, and enforcement of environmental legislation (Dinan, 2005). The latter was one of the largest prioritie s identified for action in the fourth EAP. Complete and effective implementa tion of environmental measures could only be assured if integration was car ried out at the level of Community policy, the national governments implementation, and in a generalized way so that developments in the private and public sector would meet the environmental re quirements included in both the planning and execution of the proposals (Barnes, 2000, p. 40). Yet, in terms of approach and practice the fourth EAP can be considered more of an extension of the third EAP rather than a radical reorientation. As with the third EAP, the economics of European environmental policies rema ined central to the go vernance approach put forth by the fourth EAP (Hey, 2006). Due to th e Single European Act, the fourth EAP (198751

PAGE 52

1993) found more reason to implement an effect ive environmental education policy. This was accomplished through the declaration of 1987 as th e European Year of the Environment, the adoption of Directive 92/313/EEC on the freedom of access to environm ental legislation, and through the development of the Community Inform ation System on the State of the Environment and Natural Resources (Barnes, 2000). The fourth EAP also changed its methods. While its quality policy and emissions-oriented approaches were successful in fighting probl ems caused by emissions of pollutants from many sources into one medium with little crossove r effect, a better use wa s determined possible through a multimedia approach. Under the fourth EAP, environmental protection was not perceived as an additive, but rather as an inte grated activity within th e whole production process (Hey, 2006). The impacts of transport and of electricity generation proved to be the most problematic factors from the creat ion of the single market, and the chemicals sector proved to be in need of substance-oriented controls. An integrative approach would make possible the introduction of substance-oriented controls, the targeting of th e most threatening industries within a sector, and the use of integrated risk assessment procedures (Barnes, 2000). Yet, this integrative approach would also come with a tightening of environmen tal standards and an introduction of more stringent measures. For a second time, economic instruments rather than interventionist approaches to governance were used to promote change. The fourth EAP reiterated the use of such instruments as taxe s, and subsidies (initiate d under the third EAP), expanding them with tradable emissions permits state aid, and negotiated agreements with polluters (Holzinger et al., 2006). Th is new regulatory approach was a better fit with the global trend of neo-liberal macroeconomic philosophies about market liberalization and deregulation; 52

PAGE 53

based on which, marketing the environment seemed sensible and desirable as a requirement for sustainable and unobstructed economic growth. The use of a new approach, sector anal ysis, and new environmental instruments emphasized a shift in the frame of EAPs, from trade-oriented to sustainability-oriented. Environmental policy was gradually perceived as an integrated part of economic decisionmaking and the term sustainable development slowly became a normative reference for environmental policy in the EU from 1990 onwards. Sustainable thinking was seen as a tool that could improve the environment, social efficiency, and economic competitiveness simultaneously (Hey, 2006, p. 21). Influenced by this sustainable movement, and outside factors such as the emergence of new global thre ats, a new wave of environmentalism in Europe, wider support for economic instruments, and prep arations for the United Nations Conference on Environment and Development (UNCED) conf erence in 1992, the Commission proposed a strategy to stabilize emissions through efficiency standards, fiscal inst ruments, and research. Additionally, the climate change debate began to take hold and its nature required a long-term perspective of environmental po licy in the European Union. It meant that with the aim of CO2 reduction, many sectors, such as energy, transport, agriculture, a nd the chemical industry, would have to undergo changes. The result of this new focus was the CO2/Energy tax (proposed in 1992), that advocated energy efficien cy and fuel substitution (Hey, 2006). 3.2.4 Roll-Back and Context-Oriented Governance, 1992-2002 Even though economic incentives, as an instrument of environmental governance, were widely accepted (and requested by many of the states), the proposal for an environmental/energy tax was not received with the same equanimit y. In fact the period af ter the UNCED conference 53

PAGE 54

was characterized by a more global reform wave where such catchwords as privatization and new public management advocated a re-natio nalization of EU environmental policy. Sustainability remained on the agenda and in fact was strengthened as a Community target in the Amsterdam Treaty (1997), but the ideas developed in the fifth program reflected a major departure from traditional tools of environmental governance. Interven tion and incentive based governance types were complemented and some times replaced by what became called contextoriented governance (Knill and Lenschow 2005). The Commission launched several new initiatives8 but most lacked commitment and were ultimately frustrated by the demands from member states about the competitiveness of indus tries and the decentralization of environmental policies. This new agenda partly contradicted th e ambitious ideas of the fifth EAP and therefore little progress could be achieved as the instruments used by the Commission were even sometimes self-canceling. For instance, the proposal for an energy tax coming out of the economic incentives governance approach, was wa tered down during two years of negotiations and eventually suffered re-nationalization, which transferred responsibility for introducing such a tax to the states. Several other existing and proposed pieces of legislation were attacked under this push. The Drinking Water, Packaging Waste, and Environmental Impact Assessment (EIA) directives faced serious opposition from member states and even from other di rectorates-general within the Commission (Hey, 2006). In short, the ambitious new elements of the fifth EAP caused a nationalization backlash by the stat es. The subsidiarity principle, established in EU law by the 1992 Treaty of Maastricht, had a large part in both awakening the states and promoting the development of new governance models more open to national input. 8 Like the Cardiff Process, whereby s ectoral Council formations were asked to identify the key problems of their sectors, to define objectives and to formulate activities in order to meet the objectives (Hey, 2006). 54

PAGE 55

The new regulatory approach focused on procedural requirements, framework directives, voluntary agreements and self-regu latory information and management tools (Hey, 2006). As the argument went, effective governance is best guaranteed by collaborati on between public and private actors at the various levels of policy fo rmulation. This shared re sponsibility, benefits not only the content of legislation (as it becomes mo re context-specific), bu t also its successful implementation. This meant that both states and non-governmental actors ha d to be included in the process. States where brought to the process with the use of broad objectives and avoiding detailed specification, thus enha ncing flexibility and allowing the states to choose their own instruments to achieve these objectives. Nongovernmental actors where afforded new and revised/strengthened civil society rights. Most notably the three Aarhus pillars (1998): freedom to information,9 participation rights,10 and access to justice11 (Directives 2003/4, 2003/35 and CEC Directive proposal 2003/624). The consensus built by public and private actors assured that the correct legislation would be formulated and that domestic opposition would be minimal, as the use of information a nd publicity also raised e nvironmental awareness. To summarize, the fifth EAP st arted out with interventionist and economic instruments of governance, but ended up with a patchwork of diffe rent, partially contradictory trends, with both economic and context-oriented environmental po licy approaches being promoted simultaneously Hey, 2006). This confusion and constant change coul d not be different in the environment than it was for the entire EU. The 1990s were a period of constant change, w ith two major treaties signed and ratified (Maastricht, Amsterdam) a nd one more in the works (2001 Treaty of Nice), 9 The access to environmental information directives exis ted much earlier (90/313/EEC), but was transformed to a new information directive (2003/4/EC. 10 Arrangements are to be made by public authorities to enable the public affected and environmental nongovernmental organizations to comment on, for example, proposals for projects affecting the environment, or plans and programs relating to the environment. Amending 85/337/EEC and 96/61/EC. 11 This proposal grants citizens the right to initiate admini strative or judicial procedures against acts or omissions that do not comply with environmental law, but this instrument has not yet been adopted. 55

PAGE 56

with the completion of the single market and pending introduction of the Euro, and with the EUs biggest enlargement looming in the futu re. Moreover, states found themselves maturing into the EU, and new additi ons (Finland, Sweden) brought concerns about openness and participation. All these transf ormed the environmental policy horizon radically, and accounted for the lack of consistency within this period. 3.2.5 The 6th EAP and the Them atic Strategies, 2002-2012 The sixth EAP was written shortly after Sweden took over the presidency of the European Union Council of Ministers from France in 2001. Sweden, as one of the most environmentally concerned EU states, set sustainable development, employment, and enlargement as the priorities of its agenda. The sixth EAP liste d its four main areas of concern as nature and biodiversity, environment and health, natural re source and waste, and climate ch ange, while in its third year, the concept of thematic stra tegies, a modernization of EU environmental policy-making, taking a broader, strategic approach, was introduced (EC-Environment: 6th EAP). These thematic strategies where nothing more than a fr amework of general princi ples and objectives on key issues. In short, the EU identifies themes and principles and fu rther specifies them by strategies, which are partially frameworks for further frameworks (Hey, 2006). It is clear from the above sentence, that the sixth EAP took an unclear approach to environmental legislation in the decade of enlargements. It se ems that the EU was wary of potentially contentious and controve rsial political decisions, and chos e to leave it up to the states and non-governmental actors to figure out the specifics, while it assumed a role of manager rather than initiator (Hey, 2006). This was partially due to un certainty as to the scale and solution of certain environmental problems (suc h as chemicals). But mostly it was due to the realization that if envi ronmentalists, consumers, businesses, industries, and interest groups would give their input on large issues more effective strategies could be created, and these 56

PAGE 57

organizations would naturally become more involved in enforcement. Also, if information on the environment was more readily available and si mplified for the general population, as through eco-friendly product tags, the memb er state citizens would furt her their involvement by making new eco-friendly decision s on a daily basis. Environmental policy in the sixth EAP is thus left to the hands of cooperative arrangements and the increasing influence of expert communities. This development may save the EU time and money and political capital (that it doesnt have), but it al so runs the danger of becoming too expertly defined and too narrow and short-sighted to address environmental problems given the limited horizons of member stat e politicians. With subsidiarity in the hands of the states, as per the procedur e set-up by the Lisbon Treaty of 2009,12 the cooperative management of the policy process can prove to be ve ry demanding, especially for smaller states that lack the resources and staff (or the dedication to environmental protection for some states) to deliver comprehensive answers to environmental problems. However, it is far from evident that contex t-oriented instruments deliver more than interventionist and economic in centive approaches. As Hey (2006, p.27) aptly remarks holistic and integrated approaches promise to tackle a nd balance everything with everything at the same time, and that in the end they amount only to fi ne rhetoric on principles and little action. Hey couldnt be more right, but thats just on the rhetoric part, as that is what most economic and context-oriented approaches have been. In a study of whether the adve rtised instruments in EAPs matches what is actually used in practice by the EU, Holzinger et al. (2006) find that regardless of the fanfare behi nd the introduction of novel ideas like economic incentives and cooperativeness, interventionist instruments still claim the lions share in mechanisms used to 12 Any national parliament may flag a proposal for EU action which it believes does not respect this principle. This triggers a two-stage procedure, which makes it politically di fficult to continue with a proposal if a majority of national parliaments objects to it. 57

PAGE 58

conduct environmental policy. In fact, between 1967 and 2000, interventionist instruments amounted to 85% of the instruments used, while the context-oriented ones were only present 11% of the time. Economic instruments, the ones advertised for two whol e EAPs (the third and fourth one) amounted to a measly 4% of legislati on (Holzinger et al., 2006). It is true, however, that EAP declarations did find themselves being tran sferred in the practice of EU legislation, just not at the amount advertised, and certainly not in the amount requir ed to produce any real change in governance modes. All this, points to the reality that governance ideas in Environmental Action Programs remain just that, ideas devoid of practical use. Certainly, EAPs are non-binding agreements but they represent what the states, the Commission, and the societal conditions of each time demand. And even though states, society, and theori sts demanded a more open method of policy formulation and implementation in later years, the Commission agre ed and then continued to use the same centralized deterrence system of command-and control inst ruments used since the beginning. The next section deal s specifically with them. 3.3 The EU Institutions and Compliance Mechanisms: Cooperation and Deterrence The European Union uses an integrated syst em of governance characterized by a mix of deterrence and cooperation (Gormley, 1998). Even though in later years the focus and rhetoric has switched to more cooperative mode ls and decentralization in the policy making process, the Commission still uses deterrence in struments by a large extent (Holzinger et al., 2006). The primary centralized enforcement me chanisms (monitoring, sanctions) are indeed complemented with management tools seeking to enhance the capacity of states to comply and prevent involuntary non-compliance, but the weig ht is largely on monitoring and sanctions. 58

PAGE 59

3.3.1 The Commission and Monitoring As is so often the case in Europe, the Commission was the first of central government bodies to initiate legalization (i nclusion in Treaty clauses) on environmental policy. Nevertheless no one can refer to the Commission as the body that deals with the environment as it is made up of very different bodies with very different vi ews on the environment. As mentioned earlier, the internal market DG played a negative role in va rious occasions, using the collegiality principle to block or otherwise delay environmental legi slation (e.g. drinking water, packaging waste directives in the early 1990s). The Environment DG is one of 41 Directorates-General (DGs) and specialized services which make up the European Commission. Its main role is to initiate and define new environmental legislation and to ensu re that measures, which have been agreed, are actually put into practice by Member States. The Environment DG is based largely in Brussels and has around 750 staff,13 something that is striking sin ce for example only 15 officials are charged with the supervision of chemicals, while 500 officials are so charge d in the United States Environmental Protection Agency (Sbragia, 1996). The Commission is the sole body responsible for proposing legislation and acts as guardian of the treaties ensuri ng that treaty obligations are fu lly met by individuals, companies and Member States under Article 155 of the TEC employing the procedure laid down in Article 226 of the TEU. In practical terms this entails ch ecking that transposition measures are notified and that they implement directives properly, whil e monitoring the application of regulations. The Commission carries out these tasks based on its own initiative, complaints from citizens, nongovernmental organization (NGOs), or businesses, questions from members of the European Parliament and petitions received by the Europe an Parliament exposing possible infringements 13 Accessed online at: H http://ec.europa.eu/dgs/environment/index_en.htm 59

PAGE 60

of Community law (Nugent, 2001). Hence monitori ng, as a deterrence control mechanism is a sufficiently decentralized mix of deterren ce and cooperation mechanisms for the EU. The Commission is ill-equippe d and too underfunded to detect violations by policepatrol mechanisms alone (McCubbins and Schwartz, 1984). Collect ing and assessing information on state compliance th rough its own initiative is, t hus, complemented by reliance on non-governmental organizations, institutions and othe r actors (such as states) to hit the fire alarm by filling official complaints or by parlia mentary petitions and questions. Any individual or non-governmental organization may submit a formal complaint to the Commission,14 or petition the European Parliament15 (EP), about a states purported failure to comply with Community legislation, while the sa me can also be done by a Member of the EP (through official questions). Finally, since 2008, the Commission has also deci ded to use more cooperative scheme for fire-alarm detecti on, whereby citizen or NGO complaints are referred directly to the implicated state (an informal stage), allowing th e state to resolve the issue privately (without Commission involvement). However, the Commissi on does still have the option of taking further action, if it so chooses, by launc hing infringement proceedings.16 Figure 3-1 shows the trends in detection over the last 10 years: It is evident, from the figure below, that even though the number of detections is generally increasing, the source of these detections is not what one would expect given the rhetoric about an incr ease in cooperative (fire-alarm) mechanisms of monitoring. Both the 5th and 6th EAP declared a dedication to cooperative governance instruments, which of course entails a cooperative monitoring instrument. 14 Can do so online at: H http://ec.europa.eu/community_law/your_rights/your_rights_forms_en.htm 15 The Petitions Committee of the EP assesses the valid ity of these petitions and transfers responsibility for investigation to the European Commission. 16 This is a pilot project active only in the EU 15. The goal seems to be to reduce the administrative workload of the Commission, while also enhancing cooperation between national and non-governmental actors. Further information can be found online at: H http://ec.europa.eu/environment/legal/law/complaints.htm 60

PAGE 61

0 500 1000 1500 2000 2500 3000 Total Complaints Own initiative Non communication Figure 3-1. Detection by source trend Holzinger et al.s (2006) findi ngs on the gap between declarations made through EAPs and actual practices are evid ent through the increase of police-pa trol monitoring mechanisms used by the Commission (own-initiative detection). Conversely, police patrol mechanisms of monitoring should be decreasing or at least remain the same, whereas complaints (as a cooperative monitoring mechanism) should be steadily increasing, especially since the Commission has declared its dedication to making it easier for citizens to complain and making its presence less intrusive (e.g. the 2008 pilot pr oject), however, this is not the case. Still, monitoring is half of the semi-centralized enforcement mechanisms utilized by the Commission. 3.3.2 The Commission and the European Court of Justice As guardian of the EU Treaty, the Commission has powers to take legal action against Member States not following their obligations (under Article 226). To induce compliance the threat of sanctions has to be credible and even though the Commission does not like 61

PAGE 62

antagonizing the states, the formal stage of infr ingement proceedings can be a fearful yardstick for the states. The infringement procedure has bot h formal (judicial) and informal stages where deterrence and cooperation play an equally influential role. Both unofficial and official stages of these proceedings of infringement are portray ed in Figure 3-2 (Borzel 2001, p. 807) bellow and explained in detail afterwards: Figure 3-2. Stages in infringement proceedings If the Commission considers that there may be an infringement of Community law that warrants the opening of an infringe ment procedure, it addresses a Letter of Formal Notice to the Member State, requesting it to submit its observations by a specified date, usually two months (Borzel, 2001). This instrument is less form al that the subsequent ones, and is mostly used by the Commission to weed-out simple cases of misunderstandings and cases of involuntary non-compliance, as such it is highly cooperative in nature. In the light of an unsatisfact ory reply, or absence of a repl y, from the Member State, the Commission may decide to issu e a Reasoned Opinion (or second written warning). This instrument is clearly more formal and it is where the Commission clearly and definitively explains the reasons why it considers there to ha ve been an infringement of Community law and 62

PAGE 63

calls upon the Member State to comply within a specified period, normally two months (Borzel, 2001). This stage serves to turn up the pressure against the state, as referral to the court is eminent, but it still falls under the cooperative instruments desi gned to induce compliance with a mix of eminent threat and assi stance (informing states why and how they are not complying, and what they need to do to avoid referral). If the Member State fails to comply w ith the Reasoned Opinion, the Commission may decide to refer the case to the Eur opean Court of Justice. This is one of the most important stages as the states are normally eager to avoid being named and shamed (Tallberg, 2002, p.617). Still, the Commission is not quick to refer cases to the ECJ (especi ally with bigger states), as both sides feel the pressure of costly litigation and reputations are on the line. This environment of shared interest (or fear) to avoid liti gation creates a fertile ground for negotiation and alternative forms of dispute settlement. In f act, the Commission has in stitutionalized compliance bargaining in a procedure that invol ves direct negotiations with memb er states. It should be of no surprise that most infringement proceedings end here. Looking at Figure 3-3 below, lett ers of formal notice and reas ons opinions, as instruments that deliver a mix of cooperative and deterren ce instruments to induce compliance, work very well in most cases. In fact out of all the infringement proceedings opened for 1996-2007 only 26% reached the stage of reasoned opinions, and a meager 9% were referred to the European Court of Justice. However not all states have the same respect for the Commissions carrots and sticks. States like Denmark, Netherland, Sweden and Finland, already identified as environmental leaders, retain their scepter even in dealing with infringement proceedings in a timely and cooperative manner. In contrast, others, such as Greece, France, Italy, and 63

PAGE 64

Luxembourg seem to defy Commission mechanisms and end up going to court frequently as a result. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% BELDENGERGRESPAFRAIRLITALUXNLAUSPORSWEFINUK 60 85 63 58 65 58 63 57 57 68 67 62 84 77 68 30 13 27 30 25 30 27 31 30 23 25 31 13 18 25 10 2 10 11 10 12 10 12 13 8 7 7 3 5 71996 2007 LFN RO ECJR Figure 3-3. Number of infringement procedur es opened within a year, broken down by stage of the procedure and by Member State But of course, sometimes referral to the ECJ is inevitable. The European Court of Justice is not a real venue for negotiations though, and referrals to the cour t almost always end up in favor of the Commission (Tallberg, 2002), as such Court judgments are a definite deterrence tool in the EUs arsenal. However, it does have some cooperative functions as it serves to clarify legislation and provide precedents, which clearly reduces the legal uncertainty about the meaning of rules, and hence decreases involuntary non-compliance due to ambiguity. In the past, after the ECJ ruling, and if th e state involved refuses to comply with the Courts decision, the only measure available was to renew the infringement proceedings. However, after the 1992 Treaty of Maastricht, an article 228 proce dure may be initiated, which 64

PAGE 65

may end in a daily fine for as long as the state fa ils to effectively comply with requirements. This procedure consists of the same stages as the first infringement procedure (under Article 226), but this time it can end with the imposition of fi nancial penalties. Tabl e 3-1 below provides a description of the closure of cases from 1998 until 2007: Table 3-1. Closure decisions 1998-2007, by stage reached State of the procedure Total Closures Percentage Before sending of formal notice (art. 226) 8211 38.21% Before sending of reasone d opinion (art. 226) 8354 38.88% Before referral to Court of Justice (art. 226) 2580 12.01% Before lodging the application before the Court (art. 226) 927 4.31% Withdrawal 651 3.03% Before sending of formal notice (art. 228) 624 2.90% Before sending of reasone d opinion (art. 228) 184 0.86% Before referral to Court of Justice (art. 228) 117 0.54% Before lodging the application before the Court (art. 228) 24 0.11% Withdrawal (second referral) 10 0.05% After judgment of Court of justice (art. 228) 6 0.03% The situation is clearly positive as regards to the ability of the mix of cooperative and management instruments used by the European Commission to induce compliance. Out of 21,500 cases closed within a decade as many as 94 % of the cases where resolved before lodging the application before the Court. Approximately 40% of the cases are resolved before the Commission sends a reasoned opinion, and another 12% before the Commission refers the case to the Court of Justice, which is where negot iations and agreements between the Commission and the states are most likely to be pursued to avoid costly litigation for both sides. The ECJs infringement judgments also fulfill the managerial function of reducing the legal uncertainty of EU rules, by clarifying the treaties and providing precedents for future disputes. The above table also illustrates that states are less inclined to defy the Commission when the threat of sanctions becomes more real. Even after the Court has pa ssed judgment on a case, th ere is still room to 65

PAGE 66

maneuver since the article 228 procedure for fina ncial penalties has not started yet. As the Commission goes down the ladder of inducing co mpliance, from the more cooperative article 226 formal notices to the more threatening instru ment of a second referr al; a final 3% of the cases end before sending a fo rmal notice under article 228. Financial penalties as an in strument were clearly designe d to induce compliance in the absence of a police force to enforce Court deci sions, and as the above discussion indicates, they are very successful in that respect. However, for these penalties to work, they must be sufficiently high to deter further non-compliance or repetition of the behavior in other areas and cases. They must also negate the possibility of a state benefiting from its own failure to conform to Court decisions. It is expect ed then that the Commission would make use of this new power with adequate thought and stri ctness. In fact, during the pe riod 1997 it proposed penalties in 21 cases with amounts ranging from 6,000 to 264,000 euro per day (Tallberg, 2002). It is true, though, that even if the penalty is sufficiently high, states may still choose to defy the Commission. In the first ever fine imposed by the European Union on one of its members,17 Greece defied the penalty of 20,000 Euros per day fo r almost five months (and several letters from the Commission demanding payment. It was only after the Commissi on warned Greece that it would withhold money dues from Treaties if m oney was not received by the end of 2000 that Greece finally complied. This is clearly an extreme case, not at all representative of the success the management and enforcement mechanisms of the European Commission enjoy. Even so, compliance was achieved (in the form of fine payments), with an instrument instituted by the 1997 Treaty of Amsterdam (instituted with human rights and Eastern Euro pean enlargement in mind). 17For failing to comply with a 1970s directive on waste dumpi ng near Hania on the island of Crete, first ordered in a with an ECJ judgment handed down in 1992. Judgment was made on July 4th 2000, first payment came on December 22nd of the same year. 66

PAGE 67

The aforementioned measure is a last resort d eterrence mechanism, used very few times with mixed success.18 However, the Commission has several other strategies that fall under the cooperative theme for inducing compliance. For instance, the LIFE+ Environment Policy and Governance instrument co-finances technological proj ects that offer signifi cant environmental benefits, for example process or efficiency improve ments. This part of LIFE+ also helps projects that improve the implementation of EU environm ental legislation, that build the environmental policy knowledge base, and that develop enviro nmental information sources through monitoring. Another strategy is the LIFE+ Information and Communication instrument, which co-finances projects that spread information about environmental issues, such as climate change and conservation. This strand of LIFE+ can also support environmental awareness and training campaigns. Finally, the Environm ent DG provides operating grants to European environmental Non-Governmental Organizations (NGOs).19 This includes only non-profit making and independent environmental non-governmental orga nizations active at a European level (have activities and members in at leas t three EU Member States). To summarize, the Commission has created a complex mix of enforcement and management mechanisms to induce compliance th at practically forms a ladder of patrolling. From simple fire-alarm mechanisms through comp laints (cooperative), to the more complex semi-cooperative mechanism of formal notice, to the more enforcement oriented mechanism of referral to the European Court of Justice under article 226, to the most extreme deterrence mechanism of financial penalties under article 228. This mix of soft and hard instruments covers 18 The Austrian Christian Democratic Party formed a coalition with the far-right Freedom party in the 1999 elections. To signal their displeasure, the other member states cited the Amsterdam treaty and unofficially imposed mild sanctions, but backed down when the Austrian government threatened to call a referendum on EU membership unless the sanctions were lifted. 19 The legal base for the program is the LIFE+ Regulation which provides for funding of operational activities of NGOs that are primarily active in protecting and enhancing the environment at European level and involved in the development and implementation of Community policy and legislation. 67

PAGE 68

almost the whole gamut of patrolling (suggest ed by several theories on compliance), but of course non-compliance is still a possibility, and initia l non-compliance a definite reality. An expensive, time consuming, and alienating realit y that serves to take up most of what the Commission does as guardian of the treaties, in stead of more important functions like policy innovation. 3.3.3 The Basic Principles of EU Law and the National Courts A decentralized system of compliance has developed alongside the centralized system described above. At its core, nationa l courts serve as the agents of the European Court of Justice and individuals hit the fire alarm by going ei ther to national courts or the ECJ (through preliminary ruling requests). The origins of the syst em can be traced back to the establishment of two basic principles of EU law, direct effect and supremacy. The notion that community law might have a direct effect in the legal orders of the member states is not present in the TEC, although regulations are direc tly applicable. Direct effect means that individuals can rely on community laws as such, without a requirement for national implementing legislation (only rega rding Treaties). The development of direct effect began with the Van Gend en. Loos v. Netherlands (1963) case (a Dutch transport firm brought a complaint against Dutch customs for incr easing the duty for a product im ported from Germany, thereby infringing the TEC, which spoke of no new duties, or raising of duties.). The ECJ made clear the radical basis of the doctrine. It claimed that the community represented a new legal order, and that the states had limited sovereign rights by becoming members of the community. The article of the Treaty in question had direct effect because it cont ained a clear and unconditional prohibition, which did not require further intervention at the national or community levels. Community law therefore imposes obligations on i ndividuals but is also in tended to confer right 68

PAGE 69

upon them. It allows individuals to take advant age of Community law regardless of whether national law exists or not.20 Surprisingly the initial development of the principle was more or less uncontested by the member states. But the transforma tion of EU law that ensued the application of the principle was not as important as its extension to secondary legislation, which funda mentally altered the community policy process. The extension of direct effect to directives started by the Grad v. Finanzamt Traunstein (1970), when it ruled that a directive had direct e ffect if it contained a clear and unconditional obligation on a member st ate and had not been implemented by that state within the time prescribed by th e directive. In the Van Duyn cas e (1974) the Court argued that a directive could produce direct eff ects, because it would be incompa tible with the binding effect of the directive to argue that in prin ciple they have no direct effects. Furthermore, in Johnson v. RUC (1984) the EC J declared that the right to a judicial remedy is a general principle of EC law. In Francovich and Bonifacy v. Italy (1991), the court held that in certain circumst ances, individuals are entitled to sue governments for damages sustained as a result of the governments failure to implement a directive within the prescribed period. In Haim (2000), the court extended the scope of Francovich type li ability to public law bodies legally independent of the state. The principle of direct effect would have had little impact if Community law did not supersede national law. This was about, which se t of rules should be supreme if a directly effective community law contradi cts provisions of a national law. Again the TEC was unclear on the issue, and the ECJ ruled over supremacy of community law in the Costa v. ENEL (1964) case. The Court pointed out that member states ha d definitely transferred sovereign right to the 20 Private citizens may not sue one anothe r on the basis of an EU directive, as directives are addressed to the Member States. 69

PAGE 70

Community and that Community law could not be overridden by domestic legal provisions without the legal basis of the Community itself being called into question. The ECJ expanded on the primacy of community law in Simmenthal v. Commission (1979) when it ruled that every national court must apply community law in it s entirety and must accordingly set aside any provisions of national law wh ich may conflict with it. Direct effect and supremacy, t hus, provide the framework for the decentralized system of compliance to work without EU involvement, but without the ability of the European Union to interact with national courts th e system would not be fruitful. Requests for preliminary rulings from the European Court of Justice complete th e system. Under article 234 of the TEC, if an individual argues before a nationa l court that a national law or policy conflicts with EU law, and if that court is unable, or unwilling to resolve the dispute itself based on previous EU case law, that court may seek authoritative guidance from the ECJ by making a preliminary ruling request. The parties involved, as well as EU institu tions and national governments, may submit legal arguments to the ECJ, and based on the as sessment of these arguments, relevant case law, and treaty provisions, the ECJ issu es a ruling, which the national cour t then applies to the case in question. Requests for preliminary rulings came slow ly at first, but accelerated in the 1970s and 1980s, and are about 250 a year nowadays (Dinan 2005). The ECJ gives requests for preliminary rulings a higher priority that other cases becau se national courts must await a result before proceeding with the case in question. Under article 243 lower national courts may seek guidance, but the highest national courts must do so. The pr eliminary rulings procedure is of fundamental importance to the proper functioni ng of the legal and economic syst em of the EU, and effectively gives the ECJ the power to review national law (thereby turning it into a supreme court), while it 70

PAGE 71

also sets up a system whereby the ECJ decide s and the national courts enforce. More importantly, in Kobler (2003), the court ruled that individuals we re entitled to compensation in cases where the highest national courts had not sought a preliminary ruling or had disregarded the Courts interpretation in a preliminary ruling. Taken together, all these instruments make fo r a formidable force for inducing compliance in a decentralized manner. Specifically, the system allows private individuals to sue their own governments for non-compliance, and provides for financial compensation when the government has failed to enforce community directives. This sh ifts the costs of litigation to the national level but also permits closer monito ring of non-compliance (since natio nal level players know national issues better). Additionally, since national cour t rulings are not necessarily media events, the decision has less coverage than ECJ decisions an d thus the climate of opposition and alienation between the Commission and member states is minimized, while the de cision is more likely to be followed since it was made at the national level. Finally, requests for preliminary rulings serve a largely managerial function as they reduce ambiguity, one of the main sources of involuntary non-compliance (Tallberg, 2002). 3.3.4 The European Parliament The European Parliament (EP) plays a complex role in the policy process but its impact is not as crucial as that of th e other actors. Under Article 155 of the Treaty of Rome the Commission alone is empowered to set the agen da and propose legislati on. However after the 1979 introduction of direct elections for the Eu ropean Parliament, its members came to adopt procedures that would allow them to forward draft proposals for legislation to the Commission (Judge, 1992). Any member may draft a proposal, wh ich is then referred to the appropriate committee for consideration; under Rule 63 repo rts the EP can bring up a new issue on the policy agenda for Commission Communication and encourage action from the Commission 71

PAGE 72

(Judge, 1992). The TEU made formal provision for the Parliament to invite the Commission to present a legislative proposal, thereby formalizing its ability for legislative initiative (Peterson and Stackleton, 2002, p. 99). However, when a proposal comes under Article 130s the Parliament is least effective as its responsibilities conclude in ba sic consultation with the Commi ssion. This is why the EP has strongly supported the Commission s use of Article 100a as the le gal base for the issuance of directives concerning the envi ronment (internal market), which triggers the codecision procedure. Indeed from 1989 to 1992 out of 29 envi ronmental directives 15 of them have been based on Article 100a (Judge, 1992). Additionally, as mentioned above, the European Parliament can initiate the infringement pr oceedings by questions and petitions. Parliamentary petitions are a traditional instrument of control. Any citi zen of the EU, may peti tion the EP Petitions Committee on a matter which comes within the Communitys field of activity and directly affects the citizen. Thus citizens can comp lain on both the Commission and the European Parliament. 3.3.5 The Council of Ministers The Council of Ministers is the sole deci sion-making body of the EU when it comes to solely environmental matters (unconnected to the internal market). Although the Commission plays a big role in proposing legislation the C ouncil of Ministers for the Environment are the ones who negotiate the final arrangements and vot e on its transformation to law. Typically the Council of Ministers weakens the restrictions proposed by the Commission because a great deal of intergovernmental bargaining is needed to reach a compromise. The battle used to be decided between the six green members (Germany, Holl and, Denmark, Austria, Sweden and Finland) otherwise called the leaders and the remaining members called the laggards, but the accession of ten new members in 2005 may serve to tip this balance. Germany is the odd one out 72

PAGE 73

from the environmental leaders as it implement ation record proves the opposite, but then again this may just be the case we are looking for, as an environmentally oriented state that finds it hard to follow up on its own initiatives. This leader-laggard dynamic of the EU politics, explains most of how an agenda reaches the hands of the Commission. Typically an enviro nmentally progressive state such as Germany passes national legislation more st ringent than that found in the EU generally, and the pressure for Europeanization begins. Although memb ers certainly pressure the Commission for environmental issues pertaining to their own dom estic politics, it is th e adoption of national legislation that triggers the Europeanizat ion of domestic regulations (Sbragia, 1996). This indirect pressure is based on the right given to them by Article 130s (the one not connected to the internal market), but green members make the case for Europeanization due to inequalities between their firms and other countr ies firms. The possible restriction to trade brings the Community into the game by default as firms in green countries, where a progressive environmental restriction has been approved are anxious to avoid being put at a comparative disadvantage with businesses else where in the EU (Sbragia, 1996). Nevertheless, the EUs institutional process does not allow for one member or a troika for that matter to control the agenda. As soon as the piece of legislation reaches the EU bodies it becomes part of a complex framework of codecision (Article 100a, if a market claim was made) and transformation in to be acceptable to the ma jority of the EU states (whose Environment ministers will have to vote on it), and the de sires of the EP (generally a pro-integration institution). Most of the time though the progressive piece beco mes an EU directive, which means countries that even countries did not agree with it, they w ill eventually have to implement and enforce it. 73

PAGE 74

3.3.6 The European Council and European Institutions Even though the European Council has rarely preoccupied itself with environmental policy matters, its role in the policy process of the EU has become rather critical especially since the TEU. Its formal functions include that it shall provide the Union with the necessary impetus for its development and shall define the general po litical guidelines thereo f (Hayes and Wallace, 1997, p. 160). It brings together he ads of state or of government (see France) of the member states and the President of the Commission, while the head of state or of government that holds the Presidency of the Council wi ll also hold the chairmanship of the European Council for its meetings. The fact that might make the European Council a part of the EU hierarchy is that it gives political guidance (sets the agenda) to the Union on controversial matters. In a sense, since the European Council is a meeting place for the head s of state or of government (the political principals of the individual Mi nisters of the Council) it holds the Council liable for respecting its decisions (conclusions of the presidency); and si nce the president of the Commission is present the same effect is expected of the Commission. St rictly speaking, conclusi ons of the presidency are not legally binding, but in f act they constitute a form of soft law which the Commission and the Council have to take into account and resp ect (Peterson and Stack leton, 2002, p. 30). In fact presidency conclusions have rarely21 been without any real impact on the life of the Union. 3.4. The Compliance Gap Even if the Commission were to initiate legislation on some area of policy (like the environment), there is no guarantee that it woul d pass un-bruised throug h the myriads of actors involved in the legislative process. This reality is made worse by the fact that as soon as 21 ECOFIN is notorious for circumventing presidency conclusions, for more information see Philippe de Schoutheete, The European Council in Peterson and Stackleton, eds, (2002), The Institutions of the European Union Oxford University Press, p. 21-46. 74

PAGE 75

legislation leaves the hands of the legislators, it is in the hands of the member states to transpose it in a timely manner and proceed to its implem entation and enforcement. The Commission has set up a series of enforcement and managemen t techniques to induce and ensure compliance, yet non-compliance is possible and fo r some sectors even probable. 3.4.1 The Gap in the EU The number of infringement proceedings agains t EU member states al legedly breaking the blocs environmental directives is continuing to rise despite th e Commissions efforts to improve implementation and enforcement of legislation. Because of its limited resources the Commission depends on external sources for information about domestic activity in th e member states. Onsite visits and other spot check s by Commission offici als are of limited value; they are usually time-consuming, politically fraught, and can easily be blocked by member states that are under no legal obligation to co operate (Jordan, 1999). The Commissions reliance on external actors to hit the fire-alarm is abundantly represented in Table 3-2 below, which portrays the source for detection of cases from 1996 to 2007. This table clearly shows that the Commissi on relies on fire-alarms more than policepatrols, reflecting of course its own resource limitations a nd abhorrence in antagonizing the states on its own initiative. The Commissions reliance on external actors as a source for detection of non-compliance is ex emplified by the fact that out of 29,045 total detected cases of non-compliance for 1996-2007, 47% came from co mplaints, while 38% came from noncommunication, and only 14% were the result of the Commissions own initiative. More importantly, taking 2001 as the year of refere nce, 111 of 272 cases in itiated by the Commission itself in 2001 were related to th e environmental sector (40.8%); it received 587 complaints out of 1300 total (41.3%), and had 113 non-communicati on cases (18.6%) out of 607 cases for all 75

PAGE 76

sectors.22 These percentages illustrate that the Comm ission spends most of its time addressing issues that have to do with environmental non-compliance. Table 3-2. Total number of newly detected in fringement cases, by year of detection and by origin Year Total Complaints Cases detected by the Commission Non Communication Total (Of Which) Parliamentary Questions (Of which) Petitions 1996 2155 819 257 22 4 1079 1997 1978 957 261 13 4 760 1998 2134 1128 396 18 7 610 1999 2270 1305 288 16 10 677 2000 2434 1225 313 15 5 896 2001 2179 1300 272 5 1 607 2002 2356 1431 318 30 20 607 2003 2709 1290 253 23 20 1166 2004 (EU15) 2146 1080 285 23 13 781 2004 (EU25) 2993 1146 328 23 13 1519 2005 2653 1154 433 16 11 1079 2006 2518 1049 565 18 5 904 2007 (EU25) 2345 917 488 7 5 940 2007 (EU27) 2666 958 512 7 5 1196 Source: Statistical annex (full) [SEC(2008) 2855], by 25th annual report on monitoring the application of Community law [COM(2008) 777].23 According to the Commissions own classificat ion there are five re lative categories of noncompliance with EU law. Non-communication is the most prominent type of non-compliance with EU law. Directives are not directly app licable, as a result of which they have to be incorporated into national law. Me mber states are left the choice as to the form and methods of implementation. Generally each new directive sets a time limit (usually two to three years) for members to amend their law in line with the di rectives provisions. Member states must notify transposal measures by this dead line. Non-compliance manifests itself in a total failure to issue 22 Source: Annex 1, Detection of Infringement cases, by DG XI, 19th Annual Report on monitoring the application of Community lawCOM(2002)324. 23 Available online at: H http://ec.europa.eu/community_law/infringements/infringements_annual_report_25_en.htm 76

PAGE 77

the required national legislati on (Borzel, 2001). The delays can sometimes be attributed to the institutional and administrative structures of the state, but also, in extremely technical fields, such as chemicals and biotechnology, states wi th limited resources experience problems (Annual Survey, 2001).24 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1998199920002001200220032004200520062007 Treaties Regulations Decisions Bad Application Non Conformity Non Communication Figure 3-4. Infringements for which proceedi ngs have been commenced, by legal base, 19982007 As demonstrated by Figure 3-4, non-communica tion claims the lions share with 61% (in the 1998-2007 period), as the legal base for cases in which infringement proceedings have been initiated. Non-communication is easily detected, as it comes under the Commissions automatic reporting mechanisms where no flag need be raised by any Commissi on official or nongovernmental actor to detect compliance. As suc h, one would expect that states would be more careful and non-compliance under this mechanism would be minimal, however this is definitely not the case. Additionally, non-communication as a source of detecti on of non-compliance for 24 I have included in exact, the Commissions comments and justifications as they will serve as a basis for comparison to what I find as the reasons for interstate vari ation and non-compliance. 77

PAGE 78

the Commission is 38% compared to the 47% accounted for by complaints (from Table 3-2 above), and complaints are more likely to come later in the implementatio n process than at the stage of legal transposition. Yet, the Comm ission ends up opening many more infringement proceedings for non-communication rather than for any other type of non-compliance (where complaints are prominent). In fact, as dem onstrated in Figure 3-5 for the 1997-2007 period, noncommunication accouned for 61% of all infringeme nt proceedings (followed by bad application at a distant 18%), while 66% of non-communication cases reach the letter of formal notice stage in infringement proceedings, as compared to 60% for bad applicati on. Non-communication is, thus, followed to the letter by the European Commi ssion and 8% of these cases reach the referral to the European Court of Justice stage, which is less of a percentage than the other sources of non-compliance, yet the sheer number of cases (9 20 cases) compared to th e other sources of non compliance is staggering (52% of all referrals). Th is of course means that valuable resources and time are spent disproportionately on non-communication cases, and any theoretical and empirical analysis of non-compliance should focus mainly on th ese cases, and explanations of this initial non-compliance (at the pre-implementation stage). 020004000600080001000012000 Non-Communication Non-Conformity Bad Application Treaties Regulations Decisions Non-CommunicationNon-ConformityBad Application Treaties Regulations Decisions LFN 7228 775 1924 1333 RO 2855 499 946 708 ECJR 920 213 354 274 Figure 3-5. Infringement pr oceedings by source of non-compliance and by stage reached, 19982007. 78

PAGE 79

As soon as a directive is transposed, the bur den of implementation and enforcement falls on the individual Member States (and the Commission through complaints). By focusing on the politics of implementation, one misses the crucial importance of the compliance (transposition) stage pertaining to EU directives, as that is th e stage where most of the push and pull happens between governments, domestic groups ( opposition or not) and European principals (Commission, European Parliament). The cross s ectional situation on infringements is portrayed below, where the theorized leader -laggard dynamic is exemplified. 0 500 1000 1500 2000 2500 DENNLFINSWEUKIRLLUXAUSSPAGERBELPORGREFRAITA Figure 3-6. Infringement cases by Member State, 1996-2007. Another type of non-compliance is non-conformity The transposition of Directives may be wrongful. Member states are not only required to adopt measures to implement directives, they have to ensure that these measures comply with Community law. Non-compliance takes the form of either incomplete or incorrect incorporation of Directives into national law. Parts of the obligations of the Directive ar e not enacted or national regulations deviate from European obligations (Borzel, 2001). Problems of non-conform ity arise for a variety of reasons. First, the different levels of responsibility required to be allocated between different levels of government 79

PAGE 80

within a member state. Second, difficulties may arise in amending national law because of the spill-over effect of environmental provisions on other areas of state activity (Annual Survey, 2001). Non-conformity is not a big problem for the EU, however, it is the type of noncompliance which leads to comparatively more cases reaching the Court ( 14% of non-conformity cases as compared to the Treaties, Regulations and Decisions that foll ows next with 12%). Bad application is yet another form of non-comp liance. Even if the legal implementation of a Directive is correct and complete, it still may not be practically applied. Non-compliance involves the active violat ion of taking conflicting national measures or the passive failure to invoke the obligations of the Directive. The latt er also includes failures to effectively enforce Community Law, that is, take positive action against violators, both by national administration and judicial organs, as well as make adequate remedies available to the individual against infringements which impinge on their rights (Borzel 2001). Bad application is the second largest type of non-compliance (after non-communicat ion) on all counts and commands 20% of all referrals to the ECJ. Two final two types of non-compliance are: Violations of Treaty Provisions, Regulations, and Decisions and non-compliance with ECJ judg ments. Treaty Provisions, Regulations, and Decisions are directly applicable a nd, therefore, do not have to be incorporated into national law. Non-compliance takes the form of not enforcing or incorrectly applying and enforcing European obligations (Borzel, 2001). Under this type, member states find themselves entangled with the Commission for almost 13% of the Commissions enforcement effo rts (compared to the 61% of Commission efforts on non-communication). As for non-compliance with ECJ judgments, once the European Court of Justice finds a member state guilty of infringing Community Law, the member state is finally obliged to remedy the issu e. Non-compliance under th is type refers to the 80

PAGE 81

failure of member states to execute Court judgments, whic h establishes a violation of Community Law (Borzel, 2001). However, a disc ussion on this type of infringement has preceded this discussion. What remains is a disc ussion on the temporal and cross-sectional levels of non-compliance and the sector where non-compliance seems to be worse. 3.4.2 The Gap in the States and the Environment The Commission publishes regular reports on the implementation of EU law25 and keeps regular track of the stat es transposition records,26 as such the tools necessary for the name and shame factor to induce compliance are more th an adequate. However, as the above discussion clearly illustrates, non-c ompliance is a serious issue. The mo st important type of non-compliance (for this study) is non-communication, which repr esents 60% of all infringement proceedings instigated by the European Co mmission. Figure 3-7 below tracks th e evolution of the rate of directive transposition (the source of non-co mmunication infringement proceedings) from 19982007. 92.00 93.00 94.00 95.00 96.00 97.00 98.00 99.00 100.00 1998199920002001200220032004200520062007 Average Compliance Figure 3-7. Average rate of direct ive transposition for the EU15, 1998-2007. 25 Available online at: H http://ec.europa.eu/community_law/infringeme nts/infringements_annual_report_en.htm 26 Available online at: H http://ec.europa.eu/community_law/direc tives/directives_communication_en.htm 81

PAGE 82

It becomes clear from this figure, that the rate of transposition is quite high for the period in question, yet one must not forget that there an averageof 2100 directives each year applicable in all sectors for this decade, and states comp lied with 1955 of them, which creates a backlog of 10-15 directives on average each year, that states have to implement on top of the approximately 200 a year average applicable directives. Taki ng into account that th e EU produces about 100 directives each year, the 10-15 di rective backlog is not negligible and certainly, the eleven some thousand infringement proceedings commenced on non-communication cases are not negligible either (an average of 1,200 a y ear). Arguments, such about the increase in the number of directives applicable intr oducing a bias in the data, as states have to comply with more legal acts than in the past (Borzel, 2001) seem to not apply since compliance actually increased for 2003 (from the above Figure 3-7), while the number of applicable directives in that year rose almost by a thousand directives. This situation is accurately depicted by following Figure 3-8. 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 1998199920002001200220032004200520062007 Number of Applicable directives Figure 3-8. Average number of appli cable directives for the EU15, 1998-2007. Regardless of the fact that compliance has actu ally increased in years where the number of applicable directives also increased, the growth in the number of applicable directives is logically expected to strain the Commissions enforcem ent and management mechanisms for inducing 82

PAGE 83

compliance. The Commissions own resource limita tions is what created the need for a mix of management and enforcement practices. However, as applicable directives increase, the Commission will reach the ceiling of its capacity to follow up on infringements with the same fervor as in the past, and non-governmental actors will also find it increasingly difficult to follow every piece of legislation. Thus, it is also logical to assume that th e purported increase introduced on non-compliance by the increase in applicable directives will be offset by the limitations of the detection and follow-up mechanisms. It is true, that ther e is considerable variation in infringement proceedings between sectors of policy. Accordin g to the Commission the Environment sector is the one were member states have continuously been misbehaving th roughout the last two decades of policy. The following Figure 3-9 provides an accurate depiction of this fact. 01,0002,0003,0004,0005,0006,0007,000 Environment Internal market Energy and transport Enterprise and industry Health and consumer protection Taxation and customs union Employment, and social affairs Justice, freedom and security Information society and media Agriculture Fisheries Budget Competition Other Sectors LFN RO ECJR Art 228 Figure 3-9. Stage of the infringement proced ure reached for infringement cases being under examination, broken down by sector, 1998-2007. 83

PAGE 84

It is evident, from Figure3-9, that the Co mmission spends a considerable amount of its time dealing with environmental infringements. In fact, 25% of all infringement proceedings, within the 1998-2007 time period, were related to environmental policy, with the internal market a close second (21%), and most other sectors re cording much smaller rates. More importantly when comparing the two most deviant sectors, another aspect of non-compliance comes to the fore. After issuing a letter of formal notice in the environment, 90% of the infringement cases will move on to the next round of reasoned opinions while a staggering 25% of the cases will be referred to the European Court of Justice. Comp ared to the next worst sector as regards to infringement cases, environmental policy is clear ly problematic for the Commission, as almost 7% of infringement cases ultimately reached th e article 228 stage, compared to just 3% for internal market policy. The situation is presented in the Figure 3-10. RO, 58% ECJR, 26% Art 228, 7%Environment RO, 51% ECJR, 16% Art 228, 3%Internal market Figure 3-10. Infringement proceedings after the Letter of Formal Notice; Environment, Internal Market, 1998-2007. 84

PAGE 85

While 7% Article 228 proceedings does not s ound like an impressive figure, it becomes more impressive when one realizes that this represents more than 45% of all article 228 infringement proceedings for the 1998-2007 time period ( a whopping 249 cases in all). It is evident, then, from the above discussion and illustrations, that the Commission spends a considerable amount of its time using its enfor cement and management mechanisms to induce compliance in the environment policy sector. As Figure 3-11 demonstates, transposition rates for the environment are not as bad as one may think, but the Commission uses most of its time enforcing this non-communication type of infringements (61% of all infringement cases, followed by 18% by bad application). In fact, th e 924 ECJ referrals on the environment sector (for the ten year period), constitute 52% of all referrals by source of non-compliance (noncommunication, bad application, etc). Therefore understanding the dynamics behind non-communication cases would go a long way towards decreasing the Commissions administ rative burden, while clarifying the reasons for non-compliance by the states. Bearing in mind these figures on infringement proceedings, and the above discussion on the importance of non -communication as the biggest source of noncompliance, is non-communication comparatively wors e in the Environmental field than it is in other policy areas? To put things in to perspective, out of a total of 1505 applicable directives for 2001, only 123 (8 %) of them were related to the environment (compared to 121 for the internal market; 452 for enterprise). But even though th e applicable directives were only 8% of all applicable ones, overall compliance of Member States reached 97.4% that year, while compliance in the environmental sector dropped to 92.4%. The situation in the environmental sector is depicted in the following figure. 85

PAGE 86

84 86 88 90 92 94 96 98 100 1998199920002001200220032004200520062007 Figure 3-11. Average transposition rates, Environment, 1998-2007. As we can see, compliance with environmental directives demonstrates a peculiar trend in which, compliance deteriorated for the 2000-200 2 years and then rapidly resumed its normal levels. One could argue that the final stages for the completion of the internal market project in the previous year finally hit home with compliance overload, since states surely would have developed a backlog on environmental compliance, trying to transpose directives in sectors perceived more important at the time (like the internal market). However a comparative look at compliance records between the two sector s indicates a similar trend for both. 84 86 88 90 92 94 96 98 100 1998199920002001200220032004200520062007 Environment Internal Market Figure 3-12. Transposition rates in the E nvironment and Internal Market, EU15, 1998-2007. 86

PAGE 87

However, if one notices the Energy and Trans port sector, compliance there spiked almost 10% more than pre 2001 levels for the 2001-200 2 period, possibly due to the ramping up of Commission and state activity in preparation for the Kyoto Protocol Emissions agreement ratification (with directives for greenhouse gases falling under this sector). Again, one should not be fooled by the relatively hi gh levels of compliance (communi cation of measures implementing directives), as non-communication is the largest source of opened infringement proceedings and the environment claims the larg est share of those (by 25%). Non-compliance is comparatively worse in the environment sector, Figure 3-13 below demonstrates the temporal component of th e problem and highlights the relative gap in environmental compliance as compared with rates of compliance with all other sectors of policy. 84 86 88 90 92 94 96 98 100 1998199920002001200220032004200520062007 Environment All but Environmnent Figure 3-13. Average transposition rate, Environment, all other sectors, EU15 87

PAGE 88

As the figure clearly illustrates compliance in the environment sector used to be much higher than all the other sectors in the early years of this stud y, but as compliance with other sectors began to increase (starting in 2001), co mpliance with the environment lagged behind, reaching its lowest rate of 90.4% in 2002. It is true, however, that compliance with other sectors dipped along with the environment in 2002 (contrary to its upward trend), but not to the lowest level it has been in th e 1997-2007 decade. Regardless, of this environmental compliance is lower than compliance in other sectors for 2001-2002 by 4%, in stark contrast to what is happening in other years. The reasons for this dive in 20012002 in the environment sector are difficult to explain and pertain to a series of hypotheses offered by this st udy, certainly though, it can be argued that states were preoccupied with othe r sectors than the environment. However, a depiction of interstate differences is important to grasp the spatial character of the problem. Figure 3-14 below offers such a depiction. 91% 92% 93% 94% 95% 96% 97% 98% Figure 3-14. Average Environment transpos ition rates, by Member State, 1998-2007. 88

PAGE 89

In short, if we accept trans position data as valid indicators of member state noncompliance with EU environmental law, it is clear that interstate differences exist and in some cases are extreme. Clearly, there is a laggard group of states w ith an average compliance rate between 92-95%. A difference of 3% between th e leader and laggard groups might seem negligent, at first glance, but according to the above discussion on the share of the environment in opened infringements (at firs t place with 25%), and the share of non-communication cases in infringement proceeding (again at first place with 61%), there is a clear reason to worry. One of the most striking features of the above figur e is the difference betw een the leaders and laggards on opened infringements and transpositions rates. The United Kingdom, a state cons idered a laggard in terms of transposition rates, is actually part of the leader group in opened infringement proceedi ngs with the fifth best place (less opened infringements). This oxymoron coul d be attributed to the Commission being wary of antagonizing bigger states, but that would not explain why Fr ance is second worse in opened infringement proceedings. More appropriatel y, a case can be made that the UKs internal political conditions and the existence of multiple veto points for actors, could account for the delays in transposition of directives. As the issue gets politicized a nd settles to a winning coalition in the transposition stage, implementati on goes more smoothly afterwards. It could also be that directives give little time for trans position, or that the UK is unusually averted to infringement proceedings, but th ese explanations seems less credib le. Politics, is a feature of transposition and not of implemen tation, after law has been trans posed into national legislation, political actors can do very little to re sist its implementation and enforcement. On the opposite side, there is a leader group of states with an average compliance rate of over (or close) to 97%, mostly comprising of the 1995 enlargement memb er states (Austria, 89

PAGE 90

Finland, and Sweden). Most studies dealing with environmental non-compliance either use data from before their accession or exclude them from th e analysis (for a review of the literature, see Mastenbroek, 2005). This limits the possible inferences and introduces a bias in the data as it eliminates interstate variation, clearly, an im portant component of environmental policy. The exclusion of the environmental l eaders also creates a rift in th e possible hypotheses as there is much to be said about the presence of these member states in European polit ical institutions (like the European Parliament Environment Committee). Based on the figures offered here, and on the fact that the Commission spends a considerable portion of its time dealing with non-compliance (disproportio nate with the portion that environmental directives repr esent each year), it is easy to understand why previous research on EU compliance issues has focused so much on the environmental sector. The environmental sector is important, as it is the sector where non-compliance (uncharacteristically) occurs the most, but it is not the only sector with non-comp liance, the internal market follows somewhat closely. Surely, failure in a less salient sector, such as the envi ronment, could decrease Member State belief in, and support of, European solutions for more salient sectors, which is potentially of great importance in terms of the overall func tioning of the EU and its significance in world politics, but failure in high sali ence sectors, could be even wo rse for the future of the EU. Previous researchers have met with considerable shortcomings as they tend to select easy to measure variables and l eave out possibly the most im portant ones (goodness-of-fit research), and by focusing on cer tain countries and disregardi ng major compliers and other sectors (domestic politics research). As such any model of non-compliance should not only be more theoretically integrative bu t also methodologically rigorous. 90

PAGE 91

For these reasons this analysis will offer a theoretical model of compliance based on all three approaches to non-compliance (administrative, goodness-of-fit, domestic politics), while using a mixed-method design (Mastenbroek, 2005) that obtains the benefits of all three approaches and adds control over the potentially important influe nce of domestic politics. The primary forum of analysis will be environmenta l policy, reflecting: its importance as a less salient sector; the fact that the Commission spends a considerable portion of its non-compliance efforts on it; the fact that it is the sector with the poorest compliance record; and the fact that most previous research has focused on it. Yet, to avoid the pitfalls of previous research, the model offered here will need to be tested not only on the environmental sector, but also on all sectors (a task for future reference). A model stemming from environmen tal policy and tested on all sectors of policy will be able to speak to previous research (illuminating its shortcomings), but it will also illuminate its own generalizability to explain non-compliance in the EU at large. 3.5 Conclusion There have been many ups and downs in Eu ropean environmental policy making. Through six Action Programs the EU has tried to influe nce the environmental policy plateau, sometimes successfully and sometimes failing to deliver what it promised. Parallel with the introduction of new Action Programs, and given the political climat e of the times, the EU has introduced several different and complimentary types of governance. Fr om the use of interventionist instruments in the first two Environmental Action Programs (EAP s), to the use of economic incentives in the next two and the advertised use of context-orient ed instruments in the last two EAPs the EU has attempted to change the envir onmental governance field considerab ly. However, this change, as suggested earlier, has largely been rhetoric, with limited practical change in employed instruments (Holzinger, 2006). 91

PAGE 92

In contrast, change did occur in the way the environment is administered in the EU, due largely to the end of the permissi ve consensus utilized to expa nd EU competence in previously national areas of policy. This was unequivocally taken away with the Danish no vote to the Treaty of Maastricht. The honey-moon years of central decision-making ended for the EU, and nowadays environmental protection happens at the na tional level. The role of the EU is to set minimum standards, thereby ensuring continuity whereas the responsibility for achieving those standards remains in the hands of the states This development, coupled with the limited resources of the Commission, has forced the Co mmission to expand its strategies for inducing compliance. The commission has moved from using deterrence instrument s to utilizing a mix of deterrence couple d with cooperation with the much needed non-governmental actors and their ability to hit th e fire alarm and the much needed cooperation of the states in coenforcement, using both centralized and decentral ized mechanisms, such as capacity building, and national courts. Regardless of this change in strategy t hough, non-compliance with EU legislation still remains an issue. Directives are often not tran sposed on time, badly inco rporated into national law, and poorly enforced. National courts are do ing a bad job providing governments with clear objectives of law to reduce ambiguities, and non-govern mental actors find it hard to get the right information needed to push governments to comp ly. Non-compliance is manifest in most areas of policy, while for some sectors like the internal market non-compliance has become an ongoing issue. For others, like the environment, non-compliance amounts to a crisis. Backlogs are getting big, and non-compliance is starting to bite both stat es and the Commission, whose resources have reached their limits. Some types of non-compliance are easier to deal with than others though. Non-conformity a nd bad application pertain to th e implementation side of the 92

PAGE 93

93 policy process but their numbers are not alarming enough to warrant increased attention. In stark contrast, non-communication is alarmingly the hi ghest source of infringement proceedings, straining the already limited enforcement resour ces of the Commission. More importantly, the transposition stage of the policy process is where all the politics take pl ace and should therefore be the focus of any non-compliance study.

PAGE 94

CHAPTER 4 ACCOUNTING FOR THE COMPLIANCE GAP: INTERNATIONAL RELATIONS 4.1 Introduction The nature of the European Union (EU) has been studied intensively in the past years, but there is no universal agreement as to what type of institution it most closely resembles. It is not quite a federation, not quite a confederation, and not quite an international organization. However, if the EU is to be analyzed, there is a distinct a dvantage in distinguishing between intergovernmental and intra-governmental features, especially since non-compliance may very well be rooted in the intergovernmental dynamic dyna mics at play in the European Union. If the EU is understood as an internati onal organization then Internationa l Relations theories may offer some initial insights on non-compliance. On the one hand, if the EU is an internat ional organization, real ist and neo-realist explanations focus mainly on rational action and systemic influences that fail to look inside states. On the other hand, if the EU has more s upranational elements than a simple international organization, then neo-institutionalist and domestic politics explanations require an investigation of the effects of institutional st ructures and domestic politics. To proceed, therefore, it is necessary to id entify the intergovernmental and supranational characteristics of the EU construct. In particular, it would be useful to determine to what extent the EU is the result of normal international relations and to what ex tent it should be understood as a new domestic polity. More importantly, though, the concept of compliance must not only be clearly differentiated from policy implemen tation, but also be clearly define d as differing concepts result in differing hypotheses as to the forces at play. The notion of policy implementation is tied to what has been called the textbook conception of the policy process (Nakamura, 1987, p. 142). This conception assumes that the 94

PAGE 95

policy cycle may be divided into several clearl y distinguishable phases, ranging from problem definition and agenda-setting to policy formulation, policy implementation, evaluation and finally to policy termination or re-formulation. Policy implemen tation thus refers to what happens after a bill becomes a law (Bardach, 1977) or, as one schol ar has put it, the process of translating policy into ac tion (Barrett, 2004, p. 251). A similar, but slightly different concept is that of compliance. It has been prominent in international relations research among scholars studying the domestic fulfillment of international agreements (for an overview, see Raustiala a nd Slaughter, 2002). Here, a prominent concept is that compliance can be said to occur when the actual behavior of a gi ven subject conforms to prescribed behavior, and non-compliance or violation occurs when actual behavior departs significantly from prescribed behavior (Y oung, 1979, p. 104; Raustial a and Slaughter, 2001). Thus, the compliance perspective also starts from a given norm and asks whether the addressees of the norm actually conform to it. Hence, it focuses more on the process and less on the outcome of implementation. Moreover, comp liance can occur wit hout implementation. Conversely, implementation does not necessarily have to result in compliance but may be incomplete or contrary to the pres cribed goals (Raustiala 2000, pp. 391). Following this understanding of compliance unde r the international re lations context, noncompliance within the EU can be understood in terms of intergovernmental dynamics. In this case, insights may be drawn from the extensiv e International Relations literature on state compliance with international agreements. Hence, a review of the prominent approaches in the International Relations literature is necessary. Making sure to distinguis h the various theories according to the assumptions they make about the source of non-compliance and lack of enforcement and the hypothese s those entail (fre e-ridership, wrong in stitutions, etc). 95

PAGE 96

Theorists differ in their identification of th e causes or reasons why states might choose to comply or not. Most theories or research programs in international relations apply systemic level analyses to understanding unit leve l behavior (Caporaso, 1992; Hans enclever et al., 1996; Krause and Williams, 1996). The most prominent contempor ary efforts involve realist and neo-realist efforts that stress the systemic distribution of material capabilities, whic h provide the basis for interstate leverage, while neo-liberal instituti onalists' focus on the formal organizational rules that guide strategic behavior. International relations (IR) approaches are much the same in their essential theories of the state. Realists and Institutionalis ts regard states as unitary ra tional actors, whose behavior and choices may be understood in terms of the array of incentives and choices available to the states. Domestic politics is generally treated as a residual category fo r IR scholars, although there is growing recognition that domestic politics potentially plays an increasingly powerful role in shaping state choices in the in ternational sphere (Katzenstei n, 1978, Keohane and Milner, 1996; Evans et al., 1993). Even in the area of domestic politics, the state is generally treated as a unitary actor that is dealing w ith a pluralistic society. Varia tion in state choices, from the domestic perspective, lies largely with the ability of diffuse domestic interests to forge dominant coalitions with which to pressure the government. 4.2 Why do States Comply with International Commitments? International Relations theories are primarily concerned with explaini ng state behavior and scholars have not given up on the goal of develo ping generalizable claims about the source of non-compliance with international commitments. It can be voluntary (cost-avoidance) or involuntary (lacking capacity) and the accompanying l ogic for influencing this behavior can be enforcement (for voluntary), or management (for involuntary). These two dominating 96

PAGE 97

perspectives about source and solution are comm only referred to as the enforcement and the management approach Chayes and Chayes, 1995; and Downs et al., 1996). Under the enforcement approach states are conc eived as rational actors that weigh the costs and benefits of alternative c hoices when making compliance deci sions in cooperative situations. Enforcement approaches assume that states vi olate international norms and rules voluntarily because they are not willing to bear the costs of compliance (Borzel, 2002). States will always choose to non-comply when the benefits of shirking exceed the costs of detection. It becomes clear then that compliance probl ems, under this approach, are best remedied by increasing the likelihood and costs of detection through monitoring and the threat of sanctions (Tallberg, 2002). For Neorealists the only way to alter the pa y-off matrices and ensure compliance in the absence a credible contractual environment is through the use of a hegemonic state (Downs et al., 1996; Fearon, 1998). On the other hand, advocates of Neoliberal Institutionalism subscribe to the importance of international institutions as substitutes to the enforcement powers of hegemonic states. For them, international instit utions can (and do) provide mechanisms for monitoring compliance and for coordinating sancti ons against free-riders (Victor et al., 1998; Weitsmann and Schneider, 1997). Compliance, under neorealism, may be a matter of state choice, compliance, adherence, and cooperation all turn on the political calculation of member countries that it is best to comply with international commitments (Gourevitc h, 1996, pp. 363-4). The EU cannot rely on the legitimate monopoly of force to bring about complia nce, unlike its member states. In some cases state adherence to international commitments w ill be relatively easy, especially when compliance follows a states best interests and little opposition makes itself an obstacle in the process. In most cases, however, the choice is potentially much more difficult. This is because compliance 97

PAGE 98

entails committing scarce resources to ever expa nding responsibilities, whereas the distribution of those resources will certainly be highly political. Regardless of whether it is in a states best interest to sign and comply with commitmen ts, the politics of co mpliance will alter the anticipated process and probability of success radically. If states do not anticipate resistance from domestic social forces for failing to comply w ith an obligation, they may very well commit to obligations they know they cannot possibly meet, or which are crafted so ambiguously that their obligations are not univer sally interpreted in the same manner (Haas, 1998). In reality, there are potentially several self -interested motivations behind the oxymoron of signing onto EU commitments without any reason able expectation of compliance. States may recognize that they are unable to comply, and co mmit out of a hope that the EU will help them comply at home, or they may want to signal their commitment in related areas of national importance that where part of the deal, or to strengthen a leader's political potential for implementing at home later, or because signing is part of a broader diplomatic culture associated with the West with which leaders wish to be associated (Haas, 1998). Compliance, according to Haas (1998) is also a function of the capability to comply with commitments regardless of initial purpose. This capabi lity is related to the political and technical factors associated with the decision to comply. More to the point, it is re lated to the effect of domestic resistance and the degree of behavioral change expected by political actors. Some states may lack the political will to comply; lacking the political wherewithal to induce behavioral change on its citizenry. Social choice theorists and the new institutional economics would suggest that self-regulation is less likely to yi eld compliance, because of the enormous potential for self-interested shirking. On the other hand, pluralist views of state-society relations might presume that obligations worked out through antagonistic state-so ciety relations would be more 98

PAGE 99

difficult to enforce than those worked out within the state (Haas, 1998). It stands to reason then that states may find it easier to influence the beha vior of the private sector than public activities. But the opposite is also true, given the amount of political will needed to change the behavior of domestic actors. It is true that most envir onment activities subject to EU legislation are conducted by the private sector, although some of the regulated activities apply to both public and private actors (such as air a nd water quality standards), hen ce one may find different states behaving differently depending on what is harder to change; themselves or their citizenry. Taking this line of reasoning ev er further, it can be argued that compliance is also likely to be affected by anticipated gains from domina nt coalitions at ho me (Haas, 1998). One key potential source of variation has to do with the issue being regulated and the dominant coalitions/interests active in that issue area (Milner, 1988). It may make sense that the industrial sector, with its political repres entation and concentrated infl uence on domestic politics, would encourage political regimes against environm ental legislation and compliance since they would accrue additional costs from compliance. However, it may also be true that states with an industry active in pollution c ontrol technologies would enco urage further environmental legislation and compliance, anticipating market opportunities. This intricacy would suggest a differential commitment to, and therefore comp liance with, environmental legislation depending on the dominant coalitions at home. Several liberal theorists have chose to open the black box of the state, and the effect of domestic actors, of ten in alliance with international non-governmental organizations, has been proposed as significan t (Keck and Sikkink, 1998; Risse, et al., 1999). Finally, regardless of a states in ability to change itself or it s citizenry, and the effect of domestic coalitions on policy, compliance problem s may also be exacerbated by the lack of capacity to keep commitments. The lack of competence to develop and enforce technical 99

PAGE 100

regulations consistent with inte rnational commitments is prevalen t is most developing countries (Haas, 1998), which of course may be attributed to the level of administrative development and capacity, as well as the mode of administration (enforcement or management), with less developed states having fewer resources readily available for enforcement. This lack of capacity has also been called involuntary defection (Putnam, 1988; Chayes, Chayes, and Mitchell, 1998; Chayes and Chayes 1993), and falls under the management approach to compliance. It assumes that states ar e in principle willing to meet previously agreed upon international commitments but simply lack the wherewithal to do so (i.e. the material resources, technology, expertise, ad ministrative manpower, financial means, etc.), or are simply confused about their required role due to the ambiguity of international rule s. By consequence, non-compliance is best addressed through a problem -solving strategy of capacity building, rule interpretation, and transparency, rather than through coercive enforcement (Keohane, Haas, and Levy, 1993; Janicke, 1990). As with the enforcement approach, interna tional institutions ar e crucial for ensuring compliance under the management approach. But, as Young (1992, p.183) proclaims the effectiveness of international institutions varies directly with the capac ity of the governments of members to implement their provisions. So, it make s sense that the role of institutions must be to provide financial and technical assistance fo r states with weak implementation capacities thereby helping to reduce the costs of complia nce. As mentioned earlie r, another form of involuntary compliance is rule ambiguity, wh ere more often than not there will be a considerable range within which parties may reasonably adopt differing positions as to the meaning of the relevant treaty language (Chayes and Chayes, 1995, p. 11). Under the 100

PAGE 101

management approach, the role of international institutions should be then to offer procedures that clarify obligations, such as rule interpretation, and transparency. 4.3 Realist Thought and Non-ComplianceEnforcement An environmental crisis is different from ot her crises, because while it is all around us, many important aspects of it elude us. We do not n ecessarily receive reminde rs of the crisis in our everyday life, thus governments perceive or misperceive environmental problems not as urgent as a military or diplomatic crisis (T homas, 1992). This has detrimental effects for the environment, as discovering the importance and severity of the problem constitutes half of its resolution. On top of this, IR theory in its realis t form does more harm than good. Realist recommendations for self-help policies work against environmental well being. States are instructed to use whatever means possible to increase their advantage over others. This competition leads states to adopt short-term po licies, which at best disregard environmental consequences and at worst use the environment as an expendable resource to be exploited. This brings up the issue of collective action to ensure public goods. But in Europe the environment may or may not be perceived as a public good, depending on the state. For instance, some European states (Ireland, Spain, Greece, Port ugal) find themselves in more of public goods situation as they exhibit less vulnerability to the detrimental effects of pollution due to their geography or ecosystem characteris tics (wind patterns). Other states, in contrast find themselves in more of a Common Pool Resources (CPRs) problem (Sweden, Norway, Netherlands, Finland) as pollution, domestic and imported, threatens the very existence of their ecosystems (Connoly, 1999). The difference between the alternate perception European states have on the issue can be delineated by explaining the difference between public goods and CPRs. Public goods, are both nonrival and nonexcludable, meaning that one st ates consumption of the good does not detract 101

PAGE 102

from the benefit other states can obtain by us ing it, and once the good has been provided it is impossible to exclude anyone from its use. In contrast, Common Pool Resources, while being nonexcludable, are rivalrous in nature, in the sense that consumption by any actor reduces the amount of the good others have access to by th e amount of that consumption (Barkin and Shambaugh, 1999). Free-riding by non-affected or, ev en more importantly by affected states, becomes even more dangerous to the environment as they can not only shirk in the provision of the common pool resource, but also threaten to overconsume th e CPR thus gaining concessions in the form of different standards. The second Su lfur Protocol provides evidence of this type of behavior, as Spain and the UK were allowed to make smaller cu ts in the Sulfur dioxide emissions than most member states (Connolly, 1999). Free-riding by nonaffected states, give n their geographical position poses few ecological dangers for neighboring countries. However, free-riding from affected states (as was the UK in the Sulfur Protocol), may result in potentially serious environmental degradation. Nevertheless, the reality of EU environmental legislation renders realist theory an unsatisfactory explanation of non-compliance. Namely, the leader-laggard dynamic that characterizes EU legislation turn s realist theory of power on its head. Generally, small states whose geographic position and sensitive ecosystems make them victims of negative externalities from imported pollution by powerful polluting states have been able to influence and ultimately pass legislation that limits the powerful states abili ty to indulge in self-h elp policies. Therefore power in the realist sense has not helped thos e who posses it, except in rare occasions where concessions were given as a resu lt of their threat to overconsum e, as briefly mentioned above. The environmental leaders have been able to us e instituted regimes to their advantage, at odds 102

PAGE 103

with realist theory that notes the inadequacy of institutions to constrain powerful actors against their will. Another tenet or extension of realist theory, the hegemonic stability theory,27 which extends realist propositions on conflict avoidance when a balance of power is not possible, seems at odds with European reality as well. While most neo-realists presume that states are loath to cooperate out of anticipation of fr ee riding, the presumption is that all states would only comply if they were compelled. Realists suggest then that compliance will only occur if a dominant countrya hegemonexercises some degree of pr essure on a country to comply. This can be done either through rewards for compliance or threatened sanctions for violations. The suggestion seems to be that compliance will vary depending the existence of a hegemon and its ability to offer rewards or impose sanctions (Haas, 1998). Hegemons, or even rational actors, according to this approach, provide public goods either benevolently (because they benefit disproportio nately from the production of the good and are thus willing to absorb the costs of free-riders), or because they are able to use their power to coerce smaller states into paying for a part of the production (Connolly, 1999). In fact as Mancur Olson (1971) suggests, this is what rational actors do, or better yet, this is the only way rational actors should engage in collectiv e action. In the European case, however, the distinction between perceptions of the environmen t as a public good and as a CPR works against the hegemonic solution. Hegemons are powerless to control outcomes in common pool resources situations, as one state cannot single-handedly produ ce clean air benefits for all. Surely a supposed decrease in pollution can benefit all as they would import less pollution from the hegemon, but still the collective good of cl ean environment cannot be provided without 27 See Gilpin (1981), and Wohlforth (1999). 103

PAGE 104

the participation of all members. Then there is the problem of the existence of this hegemon.28 Unilateral efforts at enforcement are improbabl e as state economic power is in rough parity between France and Germany, there is no concentrated military dominance, and no country commands deference from the rest. On top of th at, the EU is not an institution that commands sufficient resources or respect from member states to be able to unilaterally enforce compliance (Haas, 1998). And again, EU environmental legi slation is often led by small comparatively powerless states that do not have the ability to coerce other stat es into paying for the production of a cleaner environment; rather they use established regimes to do so far disproportionately to their power resources (Connolly, 1999). More pragmatically, it is also possible that the role of power in compliance may be directly related to the way in which the commitment wa s made. Moravcsik (1998) stresses the role of power in completing international agreements. However, if power was used to extract the agreement then power must be used to extract compliance. Self-interest calculations seem to work well when international agreements are bein g discussed, as un-interested states may choose to opt-out of several or all of the provisions included in the final draft of the agreement,29 and they may very well use their power to accomplis h that exemption. However non-compliance, as defined before, refers to events of states choosing to not comp ly with obligations they have already committed to. But this line of reasoning is not completely false. It may very well be that since more powerful states have an easier task in ne gotiations, their power may help them pursue agreements closer to their pr eferred outcome. Fearon (1998) argue s that studies of compliance should begin with the negotiation of the agreement th at is to be enforced, and that bargaining is a 28 See Keohane (1984) and Snidal (1985). 29 As happened with the UK and Denmark in the 1992 Tr eaty of Maastricht, and agai n in the Treaty of Lisbon. 104

PAGE 105

shortcut to understanding implementation. When st ates have the power to bargain an agreement close to their interests and situations, then those states will have no problem implementing that agreement as it is closer to their self-interest. This explanation resembles the goodness of fit argument offered by EU studies literature. Heritier (1995) argued first that states try to minimize the cost s of adaptation by Europeanizing their own policies. Several res earchers took this hypothesis even further. Knill and Lenschow (1998) provide the notion of adaptation pressure, which is defined as the degree of institutional incompatibility between national administrativ e structures and pract ices, and European requirements,30 while still others make a distinction between institutional and policy misfit (Borzel and Risse, 2003; Green Cowl es et al., 2001; Heritier et al., 2001) focusing on the content of policies rather than the inst itutional dimension. However, it may also be that bargaining power allows states to preempt this goodness of f it by Europeanizing already fitting policies. Conversely, powerless states may have a hard time bargaining for agreements close to their own interests and, as such, will not comply as well or as often. It should be evident then that both the number of votes in the Council of Ministers and the importance of countries matters in bargaining. In line with this argument, it is easy to understand why directives adopted under the unanimity rule in the Council are transposed more swiftly than deci sions under the qualified majority rule (Mbaye, 2001). This is because under unanimity rule, self-interested strategic member states can veto any proposal that does not satisfy their preferences. Under qualified majority voting, member states will not be ab le to veto a decision that goes against their preferences and administrative tr aditions. For this reason, member states outvoted in the Council will have an incentive to delay implementation (Falkner et al., 2004). 30 See also Duina 1997; Duina and Blithe 1999; Borzel 2000, 2003. 105

PAGE 106

However, a similar explanation may also be the case. States with a lot of political clout (as denoted by the number of votes in the Council) may be less inclined to comply given their power status, not out of sp ite for losing in the C ouncil vote but because they simply do not feel compelled and it goes against their interests. It becomes evident he re, that a hypothesis is needed to account for both lines of reasoning above esp ecially since power can work in these two ways. On the one hand, power can bring goodness of fit and thus easier compliance. But on the other hand, given the qualified ma jority rules in the Council, power may be used to shirk compliance. This is not to say that the below hypot hesis is infallible; it does suggest though that if power cannot be used to attain goodness of fit then it can be used to shirk compliance. Of course, this suggestion would go a long way in se ttling the goodness of fit debate by pointing out that fitness does not matter in determining compliance as much as the bargaining power needed to attain it. Thus, the following hypothesis is credible: H1: Bargaining power in the Council of Ministers affects compliance. While realism and neo-realism may account fo r some degree of compliance outside of Europe, within the EU compliance is much more likely to be a matter of the exercise of institutional channels of influence and of na tional conviction (Haas, 1998). The fact that small states can coerce powerful states to adopt unw anted legislation that abates environmental degradation using instituted regimes, points to the miraculous abilities that liberal institutionalists accord to institutions. 4.4 Neoliberal Institutionalism: Coopera tion and EffectivenessManagement Neoliberal institutionalism can be classified under the rubric of contractarian approaches to regime formation and effectiveness. Presented as the alternative to realist thought when it comes to explaining cooperation, it also goes even fa rther to suggest ways of assuring cooperation between rational unitary actors. Promoting liberal democracy and liberal economic ideals (free 106

PAGE 107

trade) through institutions seems to be a good way to preserve peace, based on the assumption that democracies do not fight each other, or more generally that ideological, procedural, and cultural resemblance between states will negate the reasons to fi ght (ONeal and Russett, 1999). In addition, promoting liberal ec onomic ideals creates a system of interdependence of state economies that makes states more sensitive to each others needs and less likely to collide over interests. In essence, one state s interests are intertwined with ot her states interests and selfish attempts to promote those interests are negated b ecause they become ineffective, if not counterproductive (Keohane, 1998). The argument made by neoliberal institutionalist s about the environment is that it is an issue of interdependence (Keohane and Nye, 1977) that can be conceptualized as the management of cooperation in a system of soverei gn states lacking the kind of central authorities necessary to ensure hegemonic coercion or induce ment. The value of regimes is concentrated in their ability to reduce incentives to cheat, by reducing transaction costs and enhancing the value of reputation in the internationa l system (Keohane and Nye, 1977). Regimes are also said to be able to create the conditions for orderly multilateral negotiations by increasing the symmetry of information between governments, thus facilita ting trust and predictability. By facilitating communication and negotiation between states they also create interactions and interdependencies, effectively reducing the incen tive of states to selfis hly pursue their national interests for reasons of reputation and reciprocity (Keohane and Nye, 1977). In essence, cooperation is promoted as stat es act to maximize their cooperative absolute gains rather than relative gain s. The environment is a sector where degradation can be better abated through cooperation because it is a phen omenon that does not recognize state borders and actions in one state pose threats in others. Th e need to cooperate l eads to the creation of 107

PAGE 108

institutions defined as p ersistent sets of rules that prescrib e roles, constrain activity, and shape expectations (Keohane et al, 1993 ). These institutions are said to be able to set agendas, coordinate policy and even more importantly generate or alter st ate behavior. Thus, they limit the ability and negate the reasons actors have to selfishly free-ride. Research on international institutions and their potential in fluence on national choice has iden tified three principal analytic functions performed by internati onal institutions: enhancing the contractual environment within which state choices are made (i ncluding voting rules, suffrage pr ovisions, number of parties, frequency of meetings, etc.), building state c oncern, and building nationa l capacity (Keohane et al. 1993). According to Keohane et al. (1993) governmental concern must be sufficiently high to prompt states to devote scarce resources to reso lve the problem despite the opportunity cost this entails. Regimes can offer rewards or punishments contingent on state policy in order to increase governmental concern. They can also generate ne w information that alters perceptions of the consequences of state actions or of the ecologi cal vulnerability of st ates. Finally they can increase concern by providing information to spec ific domestic actors that will in turn press for action. Information may affect gove rnmental concern by publicizing state actions to potentially critical domestic (and foreign) audiences. It may also affect capacity by giving governments more and better information with which to act. This argument is partly in line with the EU studies literature on the effect of national actors in obstructing or facilitating compliance (Duina, 1997; Haverland, 2000; Mabye, 200 1; Giuliani, 2003; Kaeding, 2006; Perkins and Neumayer, 2007; Konig and Luetgert 2008). Starting with Markus Haverland (2000) the ma in argument is that the degree of compliance may be determined by the preferences of domestic coalitions, mediated by institutional structures 108

PAGE 109

such as veto points. Several mediating e xplanations have been offered, expanding on Haverlands argument, such as the level of co rporatism (Lampinen and Uusikyla, 1998; Mbaye, 2001; Kaeding, 2006) or the level of pluralism (K onig and Luetgert, 2008), or finally, the level of partisan conflict and polari zation within domestic veto pl ayers (Treib, 2003; Kaeding, 2006; Konig and Luetgert, 2008). However, the effect of the principal (EU through the Commission) in enhancing compliance by tending to the needs of these domestic forces or recognizing their effect and using it to ensure compliance has larg ely been neglected in the literature. As the preceding discussion suggests, regimes can increa se governmental concern, and thus compliance, by doing exactly that (usi ng domestic players). Relatedly, even if the effect of EU enfor cement mechanisms (infringement proceedings) has been well documented (Borzel, 2000, 2001), the effect of management mechanisms has also been neglected by the literature. A hypothesi s on the EUs ability to enhance governmental concern (and thus compliance) by using manage ment mechanisms to empower domestic actors has never been offered, and given the above discussion it is more than warranted. To be precise, this study argues that while the ability of dome stic coalitions to use veto points to block compliance is well documented; the ability of the structure (EU) to use those same institutional channels to its advantage should be taken into account as it can mediate the power of these domestic coalitions and change their preferences to better matc h the compliance requirements of the EU. That is to say, veto points are not as impo rtant in determining compliance, as is the EUs ability to affect actor preferences and even utilize the same institutional channels to ensure compliance by enhancing the power of actors or domestic coalitions who stand closer to its own preferences. This line of reasoning warrants the following hypothesis: H2: Increasing governmental c oncern increases compliance. 109

PAGE 110

When it comes to the European Union most of what Keohane et al (1993) suggest on affecting compliance by increasing governmental c oncern seems to be already in place. The EU is one of the world's strongest international institutions, partic ularly with th e introduction of qualified majority voting under the Treaty of Maastricht of 1992.31 Few institutions outside the EU have been designed to be poten tially so influential and it has pr oved that it values the effect of information by the creation and strengthening of NGOs using the Environment Directorate Generals program for providing operating gr ants to European Environmental NonGovernmental Organizations (ENGOs),32 which begun operations and funding of ENGOs since 1997. Additionally, through the use of such di rectives as the Access to Environmental Information directive, which was adopted in 199 0, the Commission has sought to facilitate the formal role that domestic actors play in monitori ng the implementation of legislation in its stead, and as outlined earlier most of what the Comm ission detects comes from the domestic actor formal complaint procedure. However increasing concern will be fruitless if states fail to fulfill their obligations especially since public pressure may decrease if people believe that the is sue has been addressed (DeSombre, 2002). Hence, a reasonable contractua l environment must be in place, one that assures monitoring of each others behavior at moderate cost so that reciprocity can be followed. Under the principle of reciprocity, a states action confers the ability to other states to act in the same way. Based on this principle a state might refrain from a particular course of action, expecting that in the future other states will reci procate (Krasner, 1983). Regimes according to Keohane et al (1993) can a nd do provide these contractua l environments by generating 31 And its expansion/strengthening with the Amsterdam (1997 and Lisbon (2009) treaties. 32 The legal base for the program is the LIFE+ Regulation which provides for funding of operational activities of NGOs that are primarily active in protecting and enhancing the environment at European level and involved in the development and implementation of Community policy and legislation. 110

PAGE 111

information about potential zones of agreement and by monitoring compliance which will assure states that cheating will be de tected and punished. It seems l ogical then that a reasonable contractual environment would serve well in increasing compliance. Accordingly: H3: Providing a reasonable contractual environment will increase compliance. As with governmental concern, the EU studies literature is lack ing a measure of how effective the enforcement mechanisms of the EU are in inducing compliance. A few theorists have considered the EUs abil ity to provide for a reasonable contractual environment as an enforcement mechanism to assure complian ce (Haas, 1998; Tallberg, 2002), but it has never been quantitatively addressed. A few other resear chers use infringement proceedings (the major enforcement mechanism of the EU) as a depe ndent variable to designate non-compliance (Mbaye, 2001; Perkins and Neumayer, 2007). This study argues that using an enforcement mechanisms proceedings in measuring compliance not only represents a statistical impossibility but it also confounds the importan ce of those mechanisms in ensuring compliance. It could be that European Court of Justice (ECJ) cases i nduce compliance by showing member states that non-compliance will not go unpunished and thus reduci ng the risk of free-r iding and shirking in implementation. Rather than using those cases as a measure of non-compliance (Mbaye, 2001), one should use them as an indicator of the pow er of EU mechanisms to induce compliance. The EU has monitoring mechanisms in pl ace, both bottom-up and top-down. The Commission has a twofold responsibili ty: first, monitor the legal tr ansposition of directives, and second monitor their practical implementation. And while the first task does not seem all that difficult, the second one seems almost impossible. This is because the Commission is almost entirely dependent, upon Member States report ing back on what they are actually doing, upon costly and time-consuming consultancy reports, or on whatever national environmental groups 111

PAGE 112

and citizens choose to submit through the formal complaints procedure (Jordan, 1999) as its ability and resources to perform on-site checks are extremely limited, and can easily be blocked by member states. Yet, when the Commission is able to detect non-compliance, the ultimate sanction -referral to the ECJis a relatively effective instrument. The ECJ can rule that members are in breach of EU environmental law, and it has the power to en force its decisions especially since the advent of article 228 fines (Jordan, 1999)33 thus rendering the Commissions power to promote cooperation through punishment an effective mech anism of control. This reality seems to counterbalance the aforementioned weakness on monitoring eff ectiveness, and without the existence of the ECJ the potential of free-ri ding by member states would go unchecked.34 Finally, states must possess the political and ad ministrative capacity to make the domestic adjustments necessary to implement their commitmen ts in the international arena. By political and administrative capacity Keohane et al (1993) refer not only to the states ability to make and enforce regulation, but also to the broader ability of actors within the society to participate in policy-making and implementation. Regimes can help in this respect by tr ansferring information, skills, and expertise necessary for effective domestic programs and by building coalitions of support for environmental regulati on, and aiding monetarily the pro cess. It is therefore probable that increasing governmental capacity shoul d have positive effects on compliance. The EU does provide eligible member states with financial support from the Cohesion Fund. Only those member states with a per capita GNP of less than 90% of the EU average will be entitled to such support (Wilkinson, 1992). However there is no stipulation for states supported by the Cohesion fund in the implemen tation of environmental policy relating to 33 There are members that still have not complied to decisi ons dating back to 1990s, Greece being one of them. 34 Chapter 3 of this study offered an extensive descriptio n of this enforcement power and its effectiveness. 112

PAGE 113

compliance or compliance levels. Additionally, instead of transferring information and skills the Commission is more about checking the implementation of directives. And while a directive is binding as to the result to be ach ieved, it leaves it to the national authorities to choose the form and method of the implementation. This often leaves states muddling through the provisions of sometimes extremely technical directives. Th e opportunity to acqui re technology, training, financing, and more general reso urce transfers may help with compliance. Conversely, the fear that such resources will be withheld as a result of non-comp liance, such as occurs through conditionality, could also enc ourage states to comply. Not unlike the previous hypotheses, the EU studi es literature has focused on the effect of governmental capacity on compliance, but the ab ility of the EU to affect compliance by increasing this capacity has been neglected. One would think that since administrative capacity has so far been researched as a determinant of the ability to comply by various quantitative and qualitative studies (Mbaye, 2001; Mastenbroek, 2003; Sverdr up, 2004; Falker et al., 2005; Perkins and Neumayer, 2007), that the ability of the Commission to provide with such management mechanisms would also have been addressed. Especially since capacity is, as theorized by the management camp, one of th e major sources of involuntary non-compliance. The recognition of capacity problems by EU institu tions involved in compliance is evident by the provision for structural funds to compensate for l ack of resources under the Treaty of Maastricht (1992), which also instituted the article 228 procedure that provid es for financial penalties for non-compliance with ECJ judgments (thus enhanc ing the contractual environment mentioned earlier). Hence, as neoliberal institutionalists propose increasing capacity s hould play a positive role toward compliance. Accordingly: H4: Increasing governmental capacit y will increase compliance. 113

PAGE 114

Clearly, some degree of institutional design s hould improve the likeli hood that each of the aforementioned factors could contribute to state decisions to comply. The central idea behind this approach is that institutions enhance coopera tion by coordinating beha vior around equilibrium points, raising the costs of defection, increasi ng transparency, lowering transaction costs, and increasing member state capacity (Keohane 1984). However, even though neoliberal institutionalists postulate that regimes can increase cooperatio n through the three Cs (concern, contractual environment, capacity), a well estab lished and authoritative institution such as the European Union (or the Commission for that matte r) is not capable of doi ng so successfully or independently. Taking the suggestions of neoliber al institutionalists abou t increasing the three Cs role of the regime can be productive; however there are good reasons to question this way of thinking. Relaxing the assumption of states as unitary actors al lows a more detailed look on internal processes that influence co mpliance (i.e regime effectiveness). 4.5 Social Constructivism: a Normative Approach to Non-Compliance It is true that both theories analyzed above treat preferences of states (and thus action) as given; as readily deduced from objective condit ions and material charac teristics of the state (Finnemore, 1996, p. 8) and fail to look inside the state for the influences that might account for the creation of those pref erences. It can be said though that Neoliberal institutionalism, relaxes this assumption to explain why cooperation is ma intained (Sterling-Folker, 2000). It actually stops keeping preferences constant and allows for institutions to play a major role through the three Cs (Keohane et al., 1993) as outlined above However, it considers the structures ability to influence the behavior of the agent, and ev en though it suggests vari ous processes by which the states preferences can be ch anged internally (using local constituencies, increasing capacity etc.) it misses a very important poi nt about the ability of the local elites to influence compliance as well. 114

PAGE 115

Social constructivists have a different notion of the state. They assume that state preferences do not come from a ra tionalist calculation of interest. In fact, Constructivists assume that states are incapable of reaching the maximizi ng choice, and even worse, it is quite hard to clearly anticipate how national interest will be affected by a policy choice when this choice will have to become internalized into national law an d coincide with other policy domains. Under this bounded rationality, states are expected to satisfi ce and rely on previously developed cognitive frames to attach interest into policies bei ng internalized. Following processes of standard operating procedure, collective understandings, gu ided by norms, beliefs and ideas are become the source of state choices, as these norms, ideas and beliefs are the elem ents that account for interests (Beach, 2005; Faure and Lefevre, 2005) Norms also regulate behavior and specify appropriate behavior on related i ssues. In a sense norms (or struct ures) define who the actors are and how they will behave in certain circumstances (Jepperson et al., 1996, p. 54). The principal mechanism by which such norms are developed and disseminated is by wellplaced individuals with entrepreneurial skills, wh o can often turn their individual beliefs into broader, shared understandings (C heckel, 1998). Collectively called the civil society (Cardenas, 2004; ONeil et al., 2004) these networks of policy entrepreneurs are purp ortedly successful in turning individually held ideas into broader normative beliefs and do mesticating norms of compliance, when so-called policy windows ar e open. When members of the epistemic community acquire influential positions in national administrations and international institutions like the Commission, given their authoritative st atus, they can induce compliance using the developed norm of compliance they bring to the table. An additional process of social learning and socialization serves to further norm cr eation and dissemination (C heckel, 1998), and thus 115

PAGE 116

compliance happens almost automatically since legalized norms are internalized, and become institutionalized as the new wa y to achieve societal goals. Even though this practice of compliance through standard operating procedure can be affected by policy entrepreneurs, it can be argued that it is actually the publics response and internalization of compliance norms that will drive compliance levels. The policy entrepreneurs can do so much to assist in the dissemination of transnational ideas and beliefs about the benefit of EU policy, publics have to come to terms with accepting this policy as legitimate and as the law of the land (Beach, 2005, p. 124) for complian ce with EU policy to occur as a normative obligation (something that is not open to interpretation and debate). It becomes evident then that states with more pro-EU publics are expected to have an easier task in implementing EU policy since the public normativ ely accepts EU policy as legitimate. According to this line of reasoning, the following hypothesis is warranted: H5: Compliance in states with favorable publics on EU environm ental policy will be higher. However, it can be also argued that these pol icy entrepreneurs might not be successful in all policy domains. After all, it would take a lot of entrepreneurship to create the required linkages between policy domains to achieve no rmative compliance in all sectors of policy. Consequently, decisions to comply in one area must not be confused as normative acceptance of compliance in all policy areas as there is a considerable amount of linkage required to disseminate normative compliance to other policy sectors, which have actors with different interests, ideas and beliefs (Haas, 1998). Severa l studies in the EU implementation literature (Lampinen and Uusikyla, 1998 ; Mbaye, 2001; Kaeding, 2006; Perkins and Neumayer, 2007) have used public approval for the EU (not always making a construc tivist argument) as an 116

PAGE 117

indication of the publics acceptance of EU policy as legitimate. In this respect, this study is in line with previous research. However, a line must be drawn between policy domains to gain explanatory power, as what works in one area of policy may very well work in others, but it does not have to. As such, it is the publics approva l of EU environmental policy that this study focuses on and not approval of the EU at large in determining what affects compliance. 4.6 Conclusion Non-compliance under the guidance of mainstream international relati ons theory can have two sources and two remedies. If one ascribes to the enforcemen t approach, then the source of non-compliance is self-interested voluntary choi ce. The best way to achieve compliance is by monitoring and sanctions, but in the absence of a hegemon or a true internati onal monopoly of legitimate force, the coercive strategy of monitoring and sanc tions is seldom available or effective. Sanctioning authority is rarely grante d by treaty, rarely used when granted, and likely to be ineffective when used (Chayes and Chayes, 1995, pp. 32-33). If one ascribes to the management approach, then the source of non-co mpliance is largely involuntary (capacity and ambiguity issues), and the best remedy is cap acity building and rule interpretation. In an influential contribution, Chayes and Chayes (199 5, p. 22) emphasize that i f we are correct that the principal source of noncompliance is not will ful disobedience but the lack of capability or clarity or priority, then coercive enforcem ent is as misguided as it is costly. It becomes evident then, that most of th e potential causes of non-compliance have two lines of reasoning, ultimately painting a pictur e relatively duplicitous and complex. For this reason, it should be clear that theories expl aining non-compliance must not be treated in a disjunctive way as the complexity and duplic ity of an issue might be overlooked. When mainstream international relations theories are complemented with a focus on non-rational interest calculation and constructivist approaches are considered in a contingent manner they 117

PAGE 118

118 help provide a better understanding of the interp lay between domestic and international politics (Keohane and Milner, 1996; Evans et al., 1993; Gourevitch, 1986; Ka tzenstein, 1977; RisseKappen, 1995). In this sense, when research seeks to unde rstand a states choice to comply by focusing only on the effect of international/systemic pressu res, it may miss the intricate interplay between systemic and domestic forces. A state may si gn onto commitments only to realize that its capability to comply depends on its ability to change the behavior (preferences) of its citizenry using inducements provided by the system or by teaching compliance through the creation of consensual understandings of new interests and how those are to be ach ieved best through the EU. More importantly, it may also realize that the lack of suppor t from below, or the lack of policy entrepreneurs working to embed and domesticate norms of compliance, will make it harder or even impossible for such inducements to work.

PAGE 119

CHAPTER 5 ACCOUNTING FOR THE COMPLIANCE GAPREGULATORY FEDERALISM 5.1 Introduction Although the European Union is often viewed as a unique type of polity, giving rise to debates between neo-functionalists and realists, supranationalis ts and intergovernmentalists, there is no need for a sui generis theory to understand its particular nature. If we conceive the issue of non-compliance in the EU as an internal political conflict betw een a central authority and a set of semi-autonomous sub-units; then mode ls of state implementa tion of federal policies drawn from the American context (in particul ar theories of regulatory federalism and bureaucratic control), may be par ticularly useful in explaining patterns of non-compliance in the EU context. To proceed, one must, therefore, identify the federalist characteristics (if any) of the EU construct, which can assist in determining whether such a comparison and extrapolation from the U.S. literature to the EU is possible and/or effective. 5.2 American Federalism and the EU, Comparable? The United States have been used as a comp arison point in many studies of federations, obviously because it is the first modern federation, the result of the failure of a confederal form, and the most enduring federation in the world (Bondari, 2003). American federalism is of particular interest for the EU because of its dual character with divided sovereignty, based on two independent levels of decision-making. One cannot easily determine the similarities be tween the U.S. and EU of today, given the degree of internal and external di fferences. However, if the EU is compared to the path of U.S. Federal development the similarities become more obvious. According to Watts (1998, p.121), a Federation refers to a specific species within the genus of fede ral political systems. It is a compound polity that combines a general gover nment with constituent units, each possessing 119

PAGE 120

powers delegated to it by the people through a cons titution, each empowered to deal directly with the citizens in the exercise of a significant porti on of its legislative, ad ministrative, and taxing powers, and each directly elected by its citizens. In contrast, Confeder ations have generally been distinguished from federations as a species of federal system in which the institutions of shared rule are dependent on the constituent governments, being composed of delegates from the constituent governments and therefore having only an indirect electoral a nd fiscal base (Watts, 1998, p.121). It is also true that the dist ribution of constituti onal powers between gove rnments is a major feature of all federal systems. Federations can be thought of as being diffe rent than decentralized unitary systems since they constitutionally guarantee the autonomy of constituent governments in the responsibilities they perform. Thus it is preferable to describe federations as noncentralized, on the grounds that decentralization implies a hierarchy with power flowing from the top or center, according to their will, or can conduct policies according to their own national interests within their states (Watts, 1998, p. 124). The EU is certainly not a c onfederation, where member stat es can enter and leave on a whim. It has a legitimate power directly over it s citizens and commands -especially since the Treaty of Lisbona considerable amount of orig inal jurisdiction on policy-making not afforded to the states alone. However, there is a certain de gree of flexibility, as far as the distribution of powers is concerned, which deviates from constitutionally defi ned powers for both levels of government. The EU is not an international orga nization either, as its members are intricately bound together on policy issues, under a system of credible threats of punishment for noncompliance, where European Court of Justice (ECJ) judgments may come with daily penalty payments. If the EU is an international orga nization, then why all the preoccupation with 120

PAGE 121

democraticness and transparency (Majone 1998; Chryssochoou, 2000; Moravscik, 2004; Follesdal and Hix, 2006). So if the EU is not quite a federation, not quite a confederation, and not quite an international organization, then what is it? The EU has been studied intensiv ely in the past years, but the increase in literature has not led to an agreement about the nature of the beast. Most see the EU as a unique type of entity. However, ther e is little consensus on which characteristic of the EU matters or is dominant, the federal or the confederal. The EU has been called an incomplete federation or an overweening confederation (Kincaid, 1999, p. 37). Confusing matters ever further, it has also been called a confederal federation, wh ich means a confederal order of government that operates in a federal m ode within its spheres of competence (Kincaid, 1999). The EU is also analyzed as a regulator y state with many faces (Knill and Lenschow, 2003). Some others have called it quasi-federal or decentralized unitary state, or a system of confederal governance, or a confederal consociation (Mckay, 2001; Abromeit, 1998; Bogdandy, 2001). Or finally, an eme rging federation, a quasi-federal polity with a system of governance based on constitution-like treaties (Bor zel and Risse, 2000). Whatever it is, the EU certainly provides a test for the limits of distinction between confederal and federal systems of governance (Bondari, 2003). The EU today has many features of a federa tion (even if there are conflicts on this definition), these are:35 -The EU has at least two levels of governme nt existing autonomously and having a direct effect on the people. At the same time there is not one supreme sovereign power, which governs the whole. Sovereignty is shared or divided be tween the different levels of government, rather 35 On this issue see: Burgess (1993), Sbragia (1993), Abromeit (1998), Watts (1999), Borzel and Risse (2000), Hueglin (2000), Friedman Goldstein (2001), Mckay (2001), Chopin (2002). 121

PAGE 122

than exclusively located at one level (Bondari, 2003). The Treaty of the European Union (TEU) has formally accepted the principle of subsidia rity, which ultimately m eans that the members concede some policy areas as the exclusive competence of the EU especially when matters of efficiency arise and it is more sensible for the EU, rather than the member states to act. -Judicial relations between Co mmunity law and national laws are comparable to what many federations experienced, most notably the USA. The conflicts between the EU and the member states are regulated by the European Court of Justice, which interprets the EU treaties, as well as any rules adopted unde r its authority, as the enforc eable highest law of the land. Community law, not onl y has direct effect,36 where the ECJ claimed the community represented a new legal order, and that th e states had limited sovereign ri ghts by becoming members of the community, but also supremacy over national law,37 where the Court pointed out that member states had definitely transfe rred sovereign right to the Commu nity and that Community law could not be overridden by domes tic legal provisions without the legal basis of the Community itself being called into question. -The gradual erosion of unanimity voting re presents a movement away from a decision rule that is typical of confed erations or supranational organi zations. Qualified majority voting (QMV) in the Council of Minister s is the typical decision rule except for matters of defense, enlargement and taxation. Member states cannot always advance their will, but at the same time minority representation is guaranteed (Bondari, 2003). -The EU has a directly elect ed Parliament (since 1979), whic h over the last decades has managed to significantly increase its role in the EUs institu tional procedures with full 36 Van Gend en Loos v. Netherlands (1963), Grad v. Finanzamt Traunstein (1970) 37 Costa v. ENEL (1964), Simmenthal v. Commission (1979) 122

PAGE 123

codecision powers where the Council of Minister s decides under qualified majority voting (under the Amsterdam Treaty of 1997, and Lisbon Treaty of 2009). Existing federations are characterized by varying degrees of centralization. The EU displays a high degree of centrali zation in certain areas (such as central controls over fiscal matters, including VAT rates and controls over national borro wing) while it has only limited power in others (such as defense). The issue of defense, however, coupled with the issue of centrally provided social insurance, makes the EU deviate from the trad itional definition of a federal state, as both activities are provided at the national level in other federal systems. According to Florentia Bondari (2003) there are several ot her features of the EU that de viate from traditional definitions of a federation: member states hold the exclusive power to amend or change the constitutive treaties of the EU based on unanimity, and natio nal ratification is required; the European Council, which has a very important agenda setting role, is confederal in nature and represents states; the distribution of powers among various levels of governance is negotiable and contested under the Treaties; taxing and spending power is controlled to a large degree by the states;38 the EU lacks an essential element of democratic control in that th e EUs executive (Commission) is not yet determined by the EUs citizens, either directly, through the electi on of a president, or indirectly (Bondari, 2003). However, the historical experience of other federations shows that these issues were prominent with them in their respective be ginnings as well, and as such, should not be considered insurmountable or forbidding on the EUs path to federalism. The more serious issue is what David Mckay (2001) calls the stateness pr oblem pertaining to lack of allegiance to the 38 The EU spends less than 2% of the GDP of the whole EU area, in contrast EU member states spend on average 55% of their GDP on public spending. 123

PAGE 124

Union by its diverse population. A federation is based on a certain degree of balanced homogeneity (Bondari, 2003), wher e the population wants the union, but not the unity, and thus keeps its allegiance to the federal government w ithout being too close to it (in which case the federation would lose its part icularity and be considered a state). Conversely, when the population is too close to their individual national parts, the federation is in danger of dissolving (as happened with the USSR). To resolve the stateness problem McKay (2001) employs a scheme of three (ultimately fou r) variables that should help delineate what constitutes this problem in the EU. First, the origins of the founding constituti on. at their inception constitutions depict varying degrees of centralization, some of them even create di stinctions between central and provincial responsibilities. The EU was largely established as an economic integration process rather than a response to defense problems, which meant that such distinctions were inappropriate. The Treaty of Maastricht (TEU ) granted the EU most of its state-like characteristics, like the transf er of macroeconomic functions to the EU, the creation of a European citizenship, and the symbolic importance of the change in the name from EC to EU. However the TEU delineated the distinctiveness of the European experience, since, unlike other federations, policy typically precedes institution bu ilding in the EU. This fact means that the clear delineation of responsibiliti es between different levels of government is not one of the main features of EU treaties (its constitution). Second, how constitutions have been adapted over time to accommodate centralist and decentralist forces and the role of political par ties in the process. Constitutions do not always accurately reflect the practice of federalism, t hus McKay (2001) gives an essential role to political parties. Political parties are the main agents for articulating interests including those 124

PAGE 125

based on provincial distinctivene ss. The structure of the party system is often an effective measure of the strength of federalism. Decentra lized parties reflect a decentralized polity in which state/regional loyalties are stronger than national loyalties. The main indicators of party centralization are whether the same parties operate at the national and state levels and the degree of party discipline applicable at the national level. In the EU no union-wide parties exist that are worthy of the name, thus contributing to a decentralized federalism. In addition, national political parties remain largely disassociated with Eu ropean identity polit ics (Mckay, 2001). Third, the role of parties as legitimizi ng agents in achieving an acceptable balance between central and provincial power. Political parties are th e main means whereby provincial grievances are aired but also whereby centralis t and decentralist trends are legitimized. An increase in central activity into areas consider ed the domain of a state can reduce the legitimacy of the federal government and increase stateness problems. The aforementioned absence of genuinely European parties couple d with the absence of European identification by its citizens, places limits on the legitimacy of the EU government. The main constraint on the ability of ethnically diverse states to centralize political authority is the streng th of provincial/state loyalties in relation to national l oyalties. Where provincial loyalty is strong relative to the center stateness problems again emerge (Mckay, 2001). Fourth, the Fiscal Dimension. Distributional issues are ofte n at the heart of conflicts between central and state governments, and of thos e none is more important than fiscal relations. The advent of the EMU in the EU facilitates centralization of macroeconomic policy, and has increased the power of the EU over national/state control of monetary policy. However the EMU redistribution may not be able to broker asymmetry in economi c performance between members 125

PAGE 126

and thus potential stateness problem s might arise since states have to abide with strict standards of debt, borrowing, and ta xation (Mckay, 2001). The increasing policy responsibil ities of the EU have given new breath to the democratic deficit literature because there is an insuffici ent sense of European identity to legitimize the moral authority and the scope of new EU responsibilities. The aforementione d show that to avoid stateness problems, institutional arrangeme nts should accommodate not only the prevailing pattern of complex self-identification but al so the balance between regional and national identification. All in all, the European Union remains a sui generis39 polity. Many have challenged the use of this term, because it limits the use of comparison (Abromeit, 1998; Weiler, 1999; Newman, 1997). To say that something is unique, is to suggest that it cannot be compared with anything else, and theories used in other contexts will not apply to it. Also, it is true that the EU is in a state of constant change, which makes it ha rder for theorists to build generalizable theories for policy phenomena (or for the character of th e EU at large). The 2009 Treaty of Lisbon has radically, once more, changed the institutional char acter, the voting mechanisms in a variety of policy areas, and the principle of subsidiarity, which defines what the EU can and cannot do. If the EU is to be analyzed, it must first be must be understood on its own terms. The member states do not have internal autonomy in all areas, but at the same time there is a certain degree of flexibility in the distribution of powers. On the one hand, the EU is not an intergovernmental organization as there exist an assortment of supranational institutions, which have the power of co-deciding with the governments of the me mber states in an increasing number of competencies. On the other hand, it is not a federal state e ither, because it has not 39 Tsoukalis considers the term an admission of defeat regarding the taxonomical standards of profession, or even as a sign of agnosticism (Tsoukalis in Weiler, 1999: 13 1). For Hay the sui generic approach is an unsatisfying shrug (Hay in Weiler, 1999: 132). 126

PAGE 127

submitted to constitutional law nor do its institu tions constitute a strong center with supreme powers, as the federal authorities in a federati on (Sbragia, 1993). States continue to play an important role in decision-maki ng. Therefore the EU does not qualify as a federal state or as an intergovernmental organization, and many scholars have come to the conclusion that it is a hybrid of intergovernmental relations and domes tic polity. This makes it necessary for one to compare the processes, dynamics, and locus of power within individual areas of policy, to determine whether a comparison is possible between the EU and U.S. This is offered in the following section. 5.3 U.S. Regulatory Federalism and the EU Turning then to the domestic polity side it is true that the U.S. and the EU have had different experiences in developing institutional arrangements for environmental policy. In the U.S the federal governments role in the 1960s was limited to mana gement of public lands while water and air pollution were cons idered local issues (Jeppesen, 2002). This changed in the 1970s as increasing public concern abou t the environment led to the creation of the Environmental Protection Agency (EPA) together with amendments to the Clean Air and Water acts. The EPA was created by executive reorganization under an order that unified 15 different programs into one single jurisdiction. The EPA grew as an organization of substa ntial authority and independence, regulating through its 10 regi onal offices (Wood, 1988). During the Reagan administration, though, the EPAs responsibilitie s and budget were severely challenged in response to the deregulation wave. However, th e EPA managed to overcome these difficulties and now has the sole responsibility to enforce and/or oversee the implementation of federal regulation in the U.S.A. Contrary to the U.S., the EU institutiona l arrangements have evolved mainly through treaties. There was no formal-legal base for e nvironmental regulation addressed in the founding 127

PAGE 128

Treaty of Rome, nevertheless a considerable amo unt of legislation found its way to the states (Sbragia, 1996). The Single European Act (SEA) of 1987 was the first to formally address the issue and assign responsibility for environmenta l policy to the European Commission. The legal basis was further strengthened by the Treaty of the European Union (TEU) in Maastricht, 1992, and the Amsterdam Treaty in 1997 (Jeppesen, 2002). The last two treaties changed the envir onmental policy horizon radically. The SEA introduced qualified majority vo ting (as opposed to unanimity, with veto powers) and increased the role of the European Parliament, which now participates in the co-operation procedure for Environmental legislation linked to trade and economic issues (Article 100a) and has some veto powers. Regulations directed towards the protec tion of the environment, but unconnected to the single market, still require unanimity from the C ouncil of Ministers (Article 130s), but allow for stricter national standards in countries (such as Germany and the Netherlands). The EU has thus preempted substantial policy making powers from the states. Even though the two systems originated in quite different circumstances, substantial similarities exist. The main similarity is th at central government a nd member states share responsibility in the environmental policy field. The U.S. uses a technique called partial preemption, under which the federal government a llows states to assume primacy in the implementation of programs under the condition that state programs will be approved by the central government and will abide by minimu m standards and goals set by the federal government and its EPA. The EPAs role is to enfo rce federal statutes within the states that elect not to apply for primacy, and monitor the enforcem ent in states that have assumed primacy. If a state is found not to be in compliance with minimum standards and procedures, the EPA can revoke its grant of primacy (Grotty, 1987). 128

PAGE 129

This mechanism relieves the federal governme nt from the responsibility of providing financial and staff resources to enforce federal statutes, while ensuring that minimum standards are met by the states. Of course financial incenti ves in the form of grants-in-aid are used to induce state governments to assume primacy, but as Grotty (1987) observes, the federal government has actually reduced the amount of fede ral funds it gives to the states for pollution control, hence the level of federal government intrusion has lessened. This preemption of state authority is found with little to no differences in the EU case also. According to the principle of subsidiarity th e community shall take action relating to the environment up to the extent that objectives can be attained better at Community level than at the level of individual member states (Jepessen, 200 2). This principle included in the SEA and reinforced in the TEU has been the basis for en vironmental policy making at the EU level. It basically has to do with the e conomies of scale in EU le gislation and its comparative efficiencies. However, the body responsible fo r proposing such legislation, the Commission, does it in the form of directives. And while a directive is binding as to the result to be achieved, it leaves it to the national au thorities to choose the form and method, and the implementation,40 which subjects the Commission to a role of monitoring rather than enforcing. The two procedures, primacy and directives, are thus rather similar, but there is one basic difference. In the U.S. system the state has th e ability to opt out of the implementation and enforcement of regulations, while European me mber states must implement and enforce or else.41 40 The initial legal stance of a directive was a letter that provides a general direction toward a common goal, which had no legal binding for the member states. The European Court of Justice, change d this by recognizing the supremacy and direct effect doctrines of EU legi slation not included in the Treaty of Rome. 41 For an extensive discussion on enforcement mechanisms of the EU, please refer to section 3.3 of chapter 3 in this study. 129

PAGE 130

5.3.1 Centralized FederalismPrincipals and Agents Administrative theories are primarily concer ned with explaining administrative behavior whether it comes from agencies as a whole or th eir agents. On the level of state intervention two contrasting all-encompassing images of public bureaucracy are often distinguished: one image focuses on central controls over ag ency structure, tasks and emphasizes the importance of elected officials in determining bureaucratic behavior, while the other im age focuses on the nature of the policy task and the organizational characteristi cs of the bureaucracy (Scholz and Wei, 1986, p. 1249). Both images are rooted in the classical di stinction between centralized and decentralized forms of regulatory federalism and their respective competences, and although both can contribute to our understanding of member state compliance, it can be argued (as happened with International Relations theories) that none of them is able to provide us with a full set of hypotheses regarding the phenomenon of non-complia nce with EU environm ental directives. When regulation is centralized the focus is on uniform measures, which must be satisfied in all locations, using a great d eal of top-down influence and deterrence. When regulation is decentralized the focus is on local choice and cooperation leading to non-uniform implementation and enforcement of regulati ons. According to William Gormley (1998) a deterrence model seeks to punish regulated firms that violate ru les, while a cooperative model seeks to persuade them to improve their performa nce. All models trying to explain difference in compliance come from one of these theoretica l approaches to envir onmental regulation. Of course, there are mixed approach models that in corporate all considerations into one full model (see most notably Thompson and Scicchitano, 1985; Scholz and Wei, 1986; Hedge, Scicchitano, and Metz, 1991; and Wood, 1992). An approach consistent with the centralized federalism view is the principal-agent model proposed by John Chubb (1985). According to this model federal structures are a two-tiered 130

PAGE 131

principalagent hierarchy. Within the top tier are elected offici als (principals) that exercise power to influence their agents (federal bur eaucracy), and in the s econd tier the federal bureaucracy becomes the principal and tries to influence the behavior of subnational bureaucracies (agents). The ability of the principal-agent model to explain intergovernmental relations lies on two premises that have to be proven. Hedge, Scicch itano and Metz (1991) requ ire that an analysis must first demonstrate federal re gulators are responsive to federa l political principals, and second that federal regulators are able in turn to condi tion behavior at state ag encies. In the European context, when speaking of a federal bureaucracy we mainly refer to the European Commission. However, in the context of state agencies the plot thickens. One has to distinguish between state agencies and state governments. This is mainly because for a principal-agent model to work the second relationship has to be at least viable. The first problem with this, when juxtaposed to the EU case, is that the Commission does not make a distinction between sa te agency (ministry) and the state when it brings infringement proceedings (punishment) upon it. The proceedings are against the state and not against its environment ministry. This poses the question of accountability and control for the agency responsible for such infringements. In a way the state agency in charge for the environment is never responsible in the eyes of the Commission, or the Treaties for that matter. Passing the buck is frequently exercised in domestic politics when it comes to such situations, when the ministry blames the parliament for late decisions or bad laws, and the parliament blames the ministry for bad implementation, whichever is th e case. An additional difficulty is introduced, especially in parliamentary democracies, when th e head of the Ministry for the Environment is almost always a member of the governing body of the parliament. Domestic politics may in this 131

PAGE 132

sense infringe on the Commissions ability to locate the perpetrator of the infringement and much more hold him/her/it accountable. The second issue pertains to the abilities of the European Commission to condition the behavior of state agencies (by monitoring a nd enforcement). The Commission has a twofold responsibility regarding two non-compliance gaps. First, monito r the legal tran sposition of directives, and second monitor th eir practical implementation. And while the first task does not seem all that difficult because states are supposed to report back to the Commission on transposition measures and the Commission can refe r them to the European Court of Justice for non-communication or bad application, the second one seems almost impossible. This is because the Commission is almost entirely dependent, up on Member states reporting back on what they are actually doing, or upon cos tly and time-consuming consulta ncy reports, or on whatever national environmental groups and citizens choos e to submit through the formal complaints procedure (Jordan, 1999).42 Since the Commission is unable to control st ate agencies, task oriented models of regulatory enforcement (Gormley, 1998) and mo dels of mixed adaptation (Scholz and Wei, 1986; Hedge, Scicchitano and Metz, 1991; W ood, 1992; Hedge and Scicchitano, 1994) are correct to assume that the pr esence of active environmental groups increases monitoring and enforcement even through fire-alarm mechanisms such as complaints, much more so in the EU than in the U.S. As illustrated in Chapter 3, the percent of infringements opened that were initiated by complaints was 47 % during the 1996-2007 period. This reality verifies the ability of groups to induce enforcement through the formal EU complaint procedure, but it also limits the explanatory power of principalagent models of compliance since the Commission requires the presence and activity of citizen groups for its monitoring and enforcement mechanisms to work. 42 For a discussion on the monitoring mechanisms of the EU, please refer to sections 3.3.1 and 3.4.1 of Chapter 3. 132

PAGE 133

This argument is partly in line with the EU st udies literature on the e ffect of national actors in obstructing or facilitating compliance (Dui na, 1997; Haverland, 2000; Mabye, 2001; Giuliani, 2003; Kaeding, 2006; Perkins and Neumayer, 2007; Konig and Luetgert, 2008). However, this literature focuses mainly on the ability of ci tizen groups to use inst itutional veto points (Haverland, 2000), or other mediatin g explanations, such as the level of corporatism (Lampinen and Uusikyla, 1998; Mbaye, 2001; Kaeding, 2006) or the level of pluralis m (Konig and Luetgert, 2008), or finally, the level of partisan conflict and polarization within domestic veto players (Treib, 2003; Kaeding, 2006; Konig and Luetgert, 2008) to affect compliance at the domestic level. The ability and possibility of citizen groups circumventing national politics and using the Commissions formal monitoring mechanisms to h it the fire-alarm has been neglected by the literature. It stands to reason, that mediating factors that i nvolve citizen groups are not as important in inducing compliance as the ability of those groups to utilize top-level monitoring and enforcement mechanisms to induce complia nce. Thus, a first hypothesis is warranted: H1: States with more active citizen (e nvironmental) groups will comply more. A second difficulty for the principal-agent m odel regarding the EU case comes also from the issue of monitoring and enforcement. McCubbi ns et al. (1987) suggest that by themselves, rewards and punishments do not deal directly with the problem of asymmetric information. If non-compliance is the case, they suggest that principals invest resources in monitoring, and since monitoring is costly and sometimes ineffective on its own they supplement their suggestion by investing in procedural requirements. The suggestion of procedural requirements ha s its own problems in the EU context. The main tool of environmental policy in the EU is the directive. Directives, are binding as to the objective to be achieved but leave it to the states to decide rega rding how to achieve the required 133

PAGE 134

outcome. EU law is thus deliberately flexible to allow for adjustment to national circumstances. This is contrary to a hierarchical models c ommand and control procedure. Not surprisingly member states almost always prefer directives to other forms of regulation and almost always pressure the Commission for this form of regul ation. In this sense, again mixed adaptation models, which incorporate both top-down and botto m-up influence to compliance, are correct in predicting that the federal agency is subject to bottom-up influen ces, and pressures for individual adjustments. That being said, this study now turns to th e higher tier of the principal-agent pyramid. Admittedly there isnt much to be said about the influence and control elected officials are able to exercise on the Commission as it is almost minimal on environmental directives because it depends on whether the Commission is going to issue legislation pe rtaining to Article 100a (on internal markets) for which the codecision procedure is activated, or Article 130s for which the consultation procedure is followed.43 Given this reality, it is no t surprising that the Commission has tried, in the later years, to use Article 100a as the legal base for the issuance of directives concerning the environment. The fact that the Eu ropean Parliament (EP) has strongly supported this attempt, which triggers the codecision pr ocedure, means that the Commission is becoming more responsive to the needs of its political principal, even if this happens for its own considerations (i.e more integrat ion). Additionally, the EP has only recently been able to attract more power, namely its Environmental Committee and ask for more inclusion in the policy process (Rule 63 reports), and can invite the Commission to present a legislative proposal, but its monitoring capabilities do not exis t in the sense the principal-agent model would suggest. 43 For a discussion on the effect of political principals on the Commission, please refer to section 3.3 of chapter 3 in this study. 134

PAGE 135

The existence of parliamentary petitions and que stions in the pre-legal/administrative stage of the infringement proceedings might be considered and additional power, but these have radically decreased in later years.44 Therefore we would expect state committee membership and ideology to play some role in increasing en forcement but only through stricter scrutiny of Commission action and certai nly to a lesser degree th an that encountered in the U.S. literature. It is not far-fetched, however, to argue that members of the EP would use whatever clout they have at their disposal to attain favorable regulatory outcomes back home. There is some evidence to suggest that agencies do, in f act, respond to member pressures45 in the U.S. regulatory enforcement literature, and the EU studies literature has, indeed, considered the possibility that goodness of fit can be induced using EU institu tions to minimize the costs of adaptation by Europeanizing policies closer to pre-existi ng state policies and administrative styles.46 However, the effect of environmental leaders (countries that push fo r legislation) through the European Parliament has largely been neglecte d in the literature. It is reasonable to assume that environmental leaders with membership in the EP Environment Committee should be able to promote their agenda down to the Commission and comply more with directives since they were the ones pushing for them. It is true, though, that the existence of differential decision making procedures (as mentioned earlier), might se rve to decrease the usefulness in membership to the EP Environment Committee as a tool to control the Comm issions product (directives), and how close those products will be to the desired ones. Nevertheless, membership in the EPs Environment Committee should have a positive effect on compliance back home, even if that 44 From Table 3-2 in section 3.4.1 of Chapter 3 in this study, Parliamentary Questions/Petitions as a source for detection of non-compliance is negligent and constitutes a mere 1% of all detected cases. 45 Scholz and Wei (1986), for example, discover that states with liberal congressional delegations received higher rates of OSHA enforcement during the ea rly 1980s. Hedge and Jallow (1990) find that states with membership on the House Interior Committee were so mewhat more likely to receive higher levels of regulatory resources and enforcement during the Carter years. 46 On this issue see: Heritier 1995; Duina 1997; Knill and Lenschow 1998; Duina and Blithe 1999; Green Cowles et al. 2001; Heritier et al. 2001; Borzel 2000; Borzel and Risse 2003. 135

PAGE 136

effect pertains to the shame factor of being in the committee and not complying with legislation the committee was involved in. Accordingly: H2: Leader states with membership in the relevant (Environment) Committee of the EP will comply more. Finally, the effect of the Pres ident of the Council of Minister s should be considered when assessing top-down influences but only through its chairmanship of the European Council. Presidency of the Council of Ministers bestow s agenda setting capabilities to its presiding member state. A clear example of this remains the 1988 adoption of the directive dealing with the issue of acid rain, which was initially proposed in Germanys 1983 presidency and was adopted in 1988 again under German presidency (the President of the Council has the ability to set the agenda). However this is less effective to day as with 27 member states in the EU, each serves in the presidency just once every thirteen and a half years. Therefore presidency should not be considered in assessing top-down infl uence as the U.S. literature would suggest. In a sense, the states are the real princi pal of the Commission, which means that the Council of Ministers should be able to monitor and affect the behavior of the Commission (as the representatives of individual memb er states). However this can onl y be done in the realm of the European Council, where heads of state or of government (the principals of the Council of Ministers), the Commission president, the fore ign ministers of the me mber states, and the president of the Council have an informal, yet considerable ability to stack the deck for the Commission. As we have seen though presidency conclusions can, and have been circumvented, which points to a possible, yet seri ous gap in the ability of the principals to influence the behavior of the agent. 136

PAGE 137

However, a discussion on cooperative and dete rrence models should not end here; further elaboration is needed to determine which of th e two the EU currently uses under the assumptions of Scholz (1991). His study on regulatory enforcem ent and administrative effectiveness makes the case for using the model of cooperation, but cau tions us regarding the perverse behavior of policy supporters, who in the end oppose effect ive administration. Using a game theoretical framework, and obviously having already calculate d the equilibriums in another study he sketches a prisoners dilemma between the Agencys enforcement choices and the Firms respective compliance choices, as shown in Figure 5-1 below. Agencys Choice Flexible Enforcement (a1) Maximal Enforcement (a2) Firms Choice Flexible Compliance (b1) 3,3 Voluntary compliance equilibrium 0,5 Harassment Minimal Compliance (b2) 5,0 Capture 2,2 Deterrence equilibrium Figure 5-1. Prisoners dilemma of agency and firm choices Based on this illustration, the apparent choice of social importance is the top left cell of the voluntary compliance equilibrium where both actors get the greatest rewards. A related problem (to the aforementioned) regarding the implemen tation and especially the enforcement of EU directives by the Commission is that the process seems to be reactive rather than proactive (firealarm vs. police-patrol). Generall y each new directive sets a time limit, usually two to three years, for members to amend their law in line with the directives provisions. Member states must notify transposal measures by this d eadline. So the Commission, according to Scholz (1991), uses the flexible enfor cement strategy as its initial re sponse. However Scholz cautions against the possibility of agen cy capture; member states already know the Commissions choice, 137

PAGE 138

which is flexible enforcement, thus they would be better off in the short run making the minimal compliance choice. Hence the process is driven to th e least socially desirable result of capture. As already noted, the Commission waits until the end of the two-year period to pursue enforcement and punishment. After the end of th e time limit the game moves on gradually to the deterrence equilibrium (assuming the member stat e has not yet complied with the directive) as the Commission initiates infringement proceedi ngs. After the soft type of infringement proceedings (letters of formal notice, and reasoned opinions) ha ve ended and the member state (ridiculously) still has not complied, the Commissi on moves itself to the harassment level as it refers the case to the ECJ. Thus in all cases the Commissions behavior induces the socially undesirable outcome. Only if we assume that st ates act in good faith, and choose the flexible compliance position from the beginning, is the social ly desirable result achieved. However, this is not likely to happen for all states. The data suggest that member states will minimally comply and only begin moving up the ladder of comp liance when threatened by infringement proceedings. The case in point being that the unique Nash Equilibrium of the game is the deterrence model (2, 2) which constitutes the best response of each player to the other players strategies.47 The fact remains, however, that due to this flawed process the environment is not as protected as it should be according to adopted EU law. It is clear then, that the pr incipal-agent model offers a limited understanding of interstate differences in EU enforcement and compliance. Ho wever, its contribution remains warranted in a regulatory enforcement model as it helps us und erstand some of the top-down components of federal bureaucracies behavior. It is true, however, that the extr emes of the two views, top-down 47 If player A (agency) plays strategy a1, the best response from player B (states) is b2. If player A plays a2 then player Bs best strategy is again b2. Thus strategy b2 is the dominant response for player B (states). Following a similar analysis for player A we can see that there is a unique Nash equilibrium of the game, which is (2,2). However the social optimum is (3,3). 138

PAGE 139

and bottom-up, can be reconciled under a principal-agent framework if more emphasis is given to the agent side of relations (Wood, 1988). 5.3.2 Bottom-Up Approaches Up until the 1980s most research on politicalbureaucratic relations focused on the image of bureaucratic imperviousness to control. Bureaucrats have more knowledge, experience, intergovernmental ties, and time than political principals (Heclo, 1977); they have a monopoly over information about agency resources and act to maximi ze budgets (Niskanen, 1971; 1975); discretion provides agencies with the opportunity to alter original policy prescriptions during the implementation stage (Aberbach and Rockma n, 1976; Downs, 1967; Rourke, 1984;). As Wood and Waterman (1991) suggest, two factors assi sted in a fundamental change in out understanding of this relationship: an economic theory of political-bur eaucratic relations, and empirical support for that theory. This theory saw central political institutions (Congress, President, Courts) as molding the preferences of public bureaucracies, wh ich are situated lower in the hierarchy, using various political tools of control. Most notably, fire alarm oversight and the use of the Administrative Procedures Act48 (McCubbins and Schwartz, 1984; McCubbins, Noll, and Weingast, 1989). However, the aforementioned image of the burea ucracy as a mindless tool in the hands of elected officials does not take into account, or at least unders tates the importance of, several internal factors that might a ffect bureaucratic behavior such as professional norms, agency socialization and historical legacy (Golden, 2000). Another criticism of the way we understand tools of political control is made by Hedge and Scicchitano (1994). Th ey suggest that the 48 Enacted in 1946, this act affords the following protections to private parties being regulated : Publication of all agency rules and procedures in the Federal Register. Proposed changes in substantive rules must also be published in the Federal Register, while an opportunity for response by interested parties must be granted, and such responses are to be taken into account by the agency. Adjudicatory procedures within the agency must include the opportunity for aggrieved parties to be heard. Finally, a right to judicial review is required for a person suffering legal wrong or adversely affected by agency actions. 139

PAGE 140

intergovernmental setting in which regulation takes place may also pl ay a role in determining the extent of bureaucratic co ntrol. Control is less likely to occu r when the forces of federalism are strong; in a federal system bureau cratic agents are pulled in multip le directions by a multitude of political principals, who in turn experience thei r own pressures and compe ting role expectations (Hedge et al., 1990, p. 1075). Similarly, success of political control depends on the degree of coincidence of national and sub-national preferences (Hedge and Scicchitano, 1994). As Gormley (1998) found, the enforcement st yles of inspectors depend on whether a regulatory agency deals with issues that are lo w in salience and low in complexity or not. Low salience and complexity creates an environment in which agencies may not be quite as responsive to their principals, a nd may lack a strong element of professionalism. These agencies may also be more dependent on the preferences and inclinations of individuals (called taskoriented agencies). In contrast, agencies char ged with highly salient issues receive the lion's share of attention from politicia ns, judges, and journalists (ca lled crisis-oriented agencies, Gormley, 1998). It is true that environmen tal regulatory agencies, for the most part, exemplify the characteristics of task-oriented agencies given th e salience of the environmental issue. However, even though the environment has become a fairly salient issue in our society, environmental regulation is generally low in salience in some European countries. To some extent, this reflects the fact that environmental implementation and enforcement is handled by the states themselves rather than by a more conspicuous federal agen cy such as the Commission. As seen below, the environment has become salient in some states while normally it is low in salience when compared with different areas of policy. 140

PAGE 141

0 2 4 6 8 10 12 14 PORSPAGERITAIREGREBELAUSNLLUXUKFRAFINDENSWE Figure 5-2. Salience of Environmen tal policy 1998-2007average, EU15. Governmental policy, has always been, and w ill always be, a function of crisis (Kingdon, 1995, p. 95). In practical terms, it can be argued that compliance will be quicker in states where the environment is more salient (and elevated to crisis-oriented agency status), since agencies will be inundated with intereste d political forces (public and private). Also, in some cases, public and private officials may actively exert pressures on state agency behavior by working through the states representative institutions, the cour ts, or by directly influencing senior agency officials. In contrast, where the environment is less salient it will lead to less compliance due to fewer political interventions, which lead to more a more relaxed (task-oriented) style by agencies. The salience of an issue in domestic politics has largely been neglec ted in the EU studies literature, with a few notable exceptions (K nill and Lenschow, 1998; Versluis, 2003). Knill and Lenschow (1998), treat salience as a secondary casual factor to the constraints created by the institutional framework in which the policy will ta ke place. In short, it is argued that policy salience will be an insufficient explanation of what drives compliance, as the institutional framework (goodness of fit between European re quirements and regulatory styles) will define 141

PAGE 142

whether salience will be effectiv e or not. However, only three of their eight cases confirmed their hypothesis and as an afterthought they suggest that in situations where the level of institutional embeddedness49 suggests the more ambiguous pictur e of moderate adaptation pressure50, the explanatory value of the policy context beco mes important (Knill and Lenschow, 1998, p. 611). Since salience was expected to be high, Germ anys troubles with th e Drinking Water and Environmental Impact Assessment (EIA) directives could only be attributed to institutional embededdness. However, from Figure 5-2 above, the salience the environment in Germany is not as high as purported in their st udy, which of course may have le d to the wrong conclusion as to what matters in non-compliance. The quantification of salience proves beneficial in this way, and in fact, salience has never before been qua ntitatively addressed in the literature. It is, then, sensible to argue that policy sali ence is not a mediating factor or a secondary casual factor to the constraints created by the institutional framework in which the policy will take place. According to the pre ceding discussion, in cases where th e environment is elevated to a crisis status by interested po litical forces (public and private) then agencies will be pressured to treat the environmental issu e at hand as a crisis and thus with an added degree of professionalism, thereby increasi ng compliance. A similar argument is made by Versluis (2003), where issue salience is showed be related to the existence of ch ange agents which pressure to initiate change. However, this study argues that, the salience of an issue in private actors will create the need for attention to agency worki ngs from politicians, judges, and journalists, which of course in turn pressure agencies to be more efficient and effective w ith legislation, regardless of the existence of change agents. 49 Embeddedness is defined as the extent to which national in stitutions are deeply and/or widely institutionalized. i.e. the extent to which administrativ e arrangements are ideologically rooted in paradigms (Hall 1993). 50 Moderate adaptation pressure refers to cases where EU le gislation is interpreted as demanding changes 'within the core' of national administrative traditions but does not challenge core factors themselves. 142

PAGE 143

To be precise, change agents pressuring for le gislation can use the sa lience of the issue as a tool, but the two (salience, and change agents ) are not mutually exclusive in the sense that salience will not work without the change agents. This study has already addressed the existence of change agents,51 what remains is to hypothesize on the purported relationship between salience and the political principals of agenci es, as the salience of an issue will motivate politicians to actively pressure agencies from the top. In this case, it makes sense that the existence of green parties in legislatures will have a positive effect on compliance, since this existence gives salience an added strength and creates a direct link with institutional players. It is also true though, that green parties may not necess arily possess the strength to affect legislation, but their existence should ma tter nonetheless because they represent a voi ce in national legislatures and given legislative competition and the salience of the issue, it might be wise for ruling coalitions to go with the flow rather than disregard the issue and risk losing more votes to those green parties. Two hypotheses can be formulated in this case: H3: The greater the salience of an issue in domestic politics the greater the compliance. H4: As the proportion of special interest politi cal party (green party) membership in the national legislature increases, so will the rate of compliance. It is also true, that although EU legislators are not formally accountable to state officials, the decisions of EU officials are likely to reflect the states politic al climates. More to the point, much of the argument for the upstream pressure of green members, besides the need for regulation (Scholz and Wei, 1986) due to environmental degradat ion, relies on their attempt to protect national regulated industry from the pitfalls of stricter regul ation. It is a fact that when green members adopt stricter pol icies that those mandated by EU legislation in their attempt to protect their environment, they impose co sts on their industry which now becomes less 51 With hypothesis 1 in this chapter, on the existence of active environmental groups. 143

PAGE 144

competitive than other less regulated industries in other countries. It follows that the industry in that state will pressure its government for Eu ropeanization of the new standards. However, industry clout (Hedge et al., 1990, Hedge and Sci cchitano, 1994) may work in another direction as well. Powerful economic interests may press fo r less legislation and enforcement especially in countries where industry is advanced and comp rises of a big portion of that states Gross Domestic Product (GDP). States where the environmental problem is severe (largely due to industry emissions) should normally be expected to work harder in pollution abatement and implementation of directives and invest more resour ces. However, it is true that if the environmental problem is big, then the resources need to alleviate it will al so be considerable (Zito, 2000). This makes it necessary for states with a big pollution issue to invest more resources than other states, but it does not necessarily mean that states with a greater need for regulation (more pollution) will automatically invest more resources. On one hand, a states commitment to envir onmental regulation should have a positive effect on compliance as investment in implementation and pollution abatement will rise. In this sense, environmental expenditures not only show th e level of commitment but also the capacity of the state to implement such legislation. Bu t on the other hand, any such response from states will have a negative effect on indu stry. It is, thus, also expected that the industry will push for less legislation or for more non-compliance with EU directives. Even t hough state regulatory agencies can never completely eliminate pol itical opposition, they can attempt to minimize political costs by shirking implementation when faced with business opposition. As such, states with a greater need for regulation may actually comply less with environmental directives, not necessarily because of the need for larger investments, but because the industry in heavily 144

PAGE 145

polluted and manufacturing intensive states wi ll mobilize against environmental regulation. These lines of reasoning s uggest the following hypotheses: H5: States with a greater overall commitme nt to environmental regulation will comply more. H6: States where industry clout is great will comply less. H7: The greater the need for regulation (sever ity of the problem) the lower the compliance with environmental directives. The importance of business opposition in us ing their strong and well organized lobby groups to undermine effective implementation is well documented in the EU studies literature and elsewhere (Grant et al., 2000; Weinthal and Parag, 2003; Perkins and Neumayer, 2007). In fact several studies have some predictor rega rding the issue of imposing adjustment costs on domestic stakeholders, and the effect this will have on implementation (Underdal, 1998; Tallberg, 2002; Beach, 2005). However, manufacturers or industry at large, may not always be opposed to environmental policies (Wurzel, 200 2). Industry may actually be favorable to environmental legislation especially in green countries, and push for Europeanization of state environmental regulations to level the playing fi eld with competition in other states. In many ways this is the other side of the same coin in regards to industry clout. As argued earlier, when industry is against environmental regulation (because of the additi onal cost), they will use their considerable clout with state government to undermine implementation. Whereas, when industry is favorable to environmental re gulation, they will use their clout to increa se compliance. Even though the effect of business oppositi on has been measured both qua litatively and quantitatively in other studies, the effect of industry suppor t for environmental regulation on compliance has never been addressed before. Accordingl y, the following hypothesis is offered: 145

PAGE 146

H8: States with favorable business climates will comply more. Additional pressures toward non-compliance are li kely to emerge from within regulatory agencies. Although regulations may be written at the EU level, it s the state agencies that are responsible for implementing regulation, and each of them has its own staffing levels, budgets, and, frequently, unique outlooks on regulation (Hedge 1993). It is the street -level bureaucrat (in this case state agencies) who ca rries out the actual physical task of an organizations objectives, and many times (Lipsky, 1976), street-level bure aucrats enjoy considerable discretion and autonomy in carrying out their responsibilities (as the earlier example of inspector styles illustrates). Given this reality, there is some reason to beli eve that political actors may not always work to block legislation from being implemented or delay the process of implementation. Whether they are powerful actors, or non-governmental organizations, or simply the industry, they may use bottom-up mechanisms to indirectly in fluence agency decision making including such tools as budgets, personnel decisions, and agen cy structure (see Stewar t et al., 1982; Weingast and Moran, 1983; Moe 1985; McCubbins, Noll, and Weingast, 1989). Hence, in extension of the previous hypotheses, it can be argu ed that interested political fo rces will use their influence to decrease the number agency ability to introduce and implement new regulations. It can also be argued though, that states favorable to envi ronmental legislation will invest more in governmental agencies (again depending on the state climate toward the environment). The following hypothesis is evident: H9: Increases in the staffing levels of the State agency will be associated with higher levels of compliance. 146

PAGE 147

The above hypothesis, of course, has to do with state capacity to impl ement legislation. If agencies charged with this implementation su ffer from limited resources then non-compliance might actually be unavoidable and involuntary. The management schoo l in international relations and public administration liter ature emphasizes that non-comp liance is related to the administrative capacity and efficiency of the state (Chayes and Chayes, 1995). State agencies may be constrained in implementing EU law by a lack of governmental resources (Lampinen and Uusikyla, 1998), and by structural inefficiencies in the bure aucracy (Ciavarini Azzi, 2000; Dimitrakopoulos, 2001). Verification of this in the EU literature is somewhat mixed.52 Pridham (1994, p. 99) argues that 'the southern countries (Spain, Greece, and Ital y) do have particular problems of administrative procedure and compet ence' while Borzel (2000) finds no southern problem. Mbaye (2001) finds that bureaucratic efficiency decreases non-compliance (also Haverland and Romeijn, 2007), while Perkins and Ne umayer (2007) find that efficiency has an insignificant positive effect on non-co mpliance. Administrative effici ency is notoriously difficult to measure53 and the conflicting results in the litera ture might be due to poor indicators. However, there may be another explanation. As argued above, this study expects that when industry is favorable to environmental legisla tion then compliance will increase. Additionally, when industry has clout in domestic politics, comp liance will decrease. This, of course, points to the existence of a cozy relationship between industry and the state. The EU studies literature has addressed th is coziness under the heading of such mediating variables as veto points (Haverla nd, 2000), arguing that government must satisfy many coalition partners and other actors who shap e both the quality and speed of implementation 52 See also, Coyle (1994) who argues that Ireland's administrative capacity governs its ability to implement policy. Weale et al. (1996), argue that institutional design matters. 53 For a discussion of the difficulties and the inadequacy of most bureaucratic quality i ndicators, see Van De Walle (2005). 147

PAGE 148

(regardless of goodness of f it). After being almost universally regarded as invalid (Ferner and Hyman, 1998) in the early 1990s, corporatism is also back in the limelight (Lampinen and Uusikyla, 1998; Mbaye, 2001; Kaeding, 2006). As the argument goes, a high level of corporatism is expected to have an adverse eff ect on veto players and thus result in increased compliance (Lampinen and Uusikyla, 1998). A close and cooperative arrangement between the state and interest groups will in crease compliance while an incr eased interest group involvement (level of pluralism) will lead to non-compliance (Konig and Luetgert, 2008). Finally, the level of partisan conflict and polarization within domestic veto players (Treib, 2003; Kaeding, 2006; Konig and Luetgert, 2008) are al so hypothesized to affect comp liance at the domestic level. However, empirical investigation of the effect of veto points remains in its infancy, while also riddled with mixed results. Some find the ex istence of veto player s a good explanation of non-compliance (Kaeding, 2006; Perkins and Neumayer 2007) while others find an insignificant negative relationship between ve to players and non-compliance (i.e., the existence of veto players increases compliance, Mbaye, 2001). Corpor atism is found to be insignificant, while also shown to have a negative effect on complia nce (contrary to the hypothesized relationship, Mbaye, 2001) or with a positive effect on compliance (Kaeding, 2006).54 Pluralism and partisan conflict is shown to increase non-compliance (Konig and Luetgert, 2008), while Kaeding (2006) finds a positive yet insi gnificant relationship. All these conflicting resu lts should make it abundantly clea r that we are either measuring the right things with the wrong indicators or that simply we need a bett er explanation of actor involvement. As argued earlier, th is study expects to find a cozy relationship between industry and the state. When industry is powerful then we should expect to see increased levels of noncompliance. Bureaucratic efficiency should play a role in this, but not in the way hypothesized in 54 For a discussion on quantitative measures of corporatism refer to Kenworthy (2000). 148

PAGE 149

other studies. If a bureaucracy is focused on priv ate sector development then we should expect that it will be efficient in implementing regula tion that promotes and permits private sector development. If the bureaucracy is focused on en vironmental protection then it will be efficient in promoting environmental protection. It is, the n, necessary to establish who it is exactly that creates this focus on the bureaucr acy. We should expect to find agencies susceptible to both politician and industry involvement, but these act ors will be successful in pushing the agency only if they are able to change the focus of that agency to match their preferences. To be concise, bureaucratic efficiency does not matter as much as whether that efficiency is focused toward environmental legislation or private sector development, which invariably helps industry. In this way, public and private actors may find it beneficial to change the efficiency focus of the bureaucracy to match thei r needs but this will depend on the receptivity of the bureaucracy to behavioral m odification (Wood, 1988). It is true that some states (especially the poor ones) are more focused on private sector development rather than environmental protection (similarly to the above hypothesis on overall commitment ), hence their bureaucracies will be efficient in promoting private sector de velopment rather than environmental protection. Even if these two goals are not necessarily mutually ex clusive, the presence of a strong industry with a strong lobby, will keep th e focus of bureaucratic efficiency closer to its own preferences. This means that if the bureaucracy is efficient in making regulation that helps industry then we should expect non-compliance to increase. Related to this argument is, of course, the presen ce of corruption in the public sector and in the administration (Heidenheimer et al., 1989; Mauro, 1995). Corruption has been found to have a significant negative effect on compliance (M baye, 2001; Kaeding, 2006) in systems where side-payments and patronage positions are the norm, tasks are accomplished only when 149

PAGE 150

bureaucrats have a personal incentive to get th ings done (Mbaye, 2001). It follows, coupled with the above argument on the power of street leve l bureaucrats, that if implementation does not produce personal incentives for bureaucrats, then non-compliance is possible. Conversely, if non-compliance produces a side-payment to bureau crats from a much powerful and wealthy industry sector then we should expect non-complia nce to be increased in states with corrupt administrations. Two final hypotheses are thus warranted: H11: States with a cozy regulator y efficiency will comply less. H12: States with corrupt administrations will comply less. The literature of regulatory federalism presen ts an array of hypotheses to better understand the determinants of regulatory enforcement and stat e behavior. It follows from the above analysis though that we should expect EU Me mber State compliance to reflect domestic politics, more so than it does in the U.S. literature. 5.4 Conclusion This chapter offers an integrative approach that aims to capture both external and internal influences on bureaucratic behavior, which determine the level of compliance with EU environmental directives. The presence of active environmental groups utilizing the monitoring mechanisms of the Commission, and membership in the European Parliaments Environment Committee should play a significant top-down role to affect comp liance. Yet, this study argues that several bottom-up influences (salience, commitment, business climate, industry clout, severity of the problem) should offer a better ex planation of non-compliance due to the role of domestic forces. While, also, internal agency characteristics and the influence of domestic political forces on these characteristics should prove powerful explanations of non-compliance. However, this is not to say we should treat th e top-down versus bo ttom-up dichotomy in a disjunctive way. To disregard the top-down infl uences completely when arguing for the power 150

PAGE 151

151 of bottom-up influences would be a fallacy as they measure different aspects of non-compliant behavior, and therefore the complexity and dup licity of an issue might be overlooked. Hence, these approaches should be seen as potentially co mplimentary, and even if an approach proves to be statistically insignificant in a quantitative anal ysis, this only adds to the explanatory power of a different approach th at proved significant. As Hedge and Johnson (2002) suggest, c ongressional-bureaucrat ic relations are conditioned by presidential politic s and the struggle between them Others have suggested that the ability of elected principals to shape outputs depends not only on the effective application of the tools of control, but also on the resources and receptivity of the bureaucracy to behavioral modification (Wood, 1988, citing the case of the EPA in the Reagan era). Still others seem to suggest that the relationship betw een principals and agents is not as one-sided as we thought and the behavior of political principa ls can be influenced by administ rative agencies (Krause, 1999). All of the claims and counter-claims seem to paint a discouraging picture for the understanding of the issue. When we focus on one tool of control, or on one agency, or on one political principal it is obvious that we are missing the d rama of national politics as Hedge and Johnson (2002) call it, instead we should en tertain approaches that look at the effects of the political context on both political control and its tools, making sure that we look at both mi cro (internal) and macro (external) influences to bureaucratic behavior.

PAGE 152

CHAPTER 6 DATA, METHODS, AND ANALYSIS 6.1 Introduction The most influential simplification of the polic y process is the stages model as conceived by Harold Lasswell (1956). It divides the polic y process into a series of stages problem identification, policy formulation, adoption, implementation, evaluation, and termination. However, three critical stages can be identified through which all polic ies must pass sooner or later: policy formulation, po licy adoption, and policy implem entation (Hayes, 1992). While the EU has created a dense mass of institutions a nd decision-making procedures for the first two stages, policy implementation (especially in the la ter years), remains largely in the hands of the states. This is the longest and most arduous part of the policy-ma king process, especially in the EU because of its institutional character. The Commission is in charge of implementation, but this is accomplished through secondary implementi ng legislation (directives), but also through constant and vigilant monitoring. All EU policie s must be implemented at the national level, generally through national implementing legislation (occasionally at the sub-national level). As a result the Commission can only assist and monitor national actors to try and insure that they follow through on their responsibilities (in a timely manner). This delegation of broad implementation powers to the states is one of the most important characteristics of contemporary EU government However, delegation through directives only defines goals to be achieved, and leaves it up to the states discretion to choose their own mechanisms and measures to achieve those goals w ithin specified periods of time. This is called the transposition stage and its completion is cruc ial to the process as incorrect transposition may lead to bad application and non-conformity with EU law. Discretion, however, is not limited to transposing legislation but also applies to application and en forcement. This constitutes a 152

PAGE 153

problem as bureaucratic discretion in implementing policy threatens the idea of the rule of law, that governmental actions be clear and specific and applied by officials in a non-discretionary manner (Bryner, 1987). Critical then, to policy succe ss is the commitment of the states to the objectives of a particular policy, and success is defined as correspondence between goals defined at the top and actions taken in the field. In Chapters 4 and 5 theories of international relations and regulatory federalism were introduced, respectively, to explain the i ssue of member-state non-compliance with environmental directives of the European Union. As a result, several hypo theses were developed to explain non-compliance of EU member states. The purpose of this chapte r is to parameterize these hypotheses by developing quantitative measures consistent with each hypothesis and, subsequently, to empirically evaluate the eff ect of these measures on non-compliance utilizing annual data for 15 member states during 19982007. I start with the operationalization of compliance as the dependent variable. 6.2 Operationalization of Compliance It is important to clarify th e concept of compliance as it is often misconstrued with policy implementation effectiveness. This misint erpretation leads to different variables and different data becoming important (like open in fringements), which this study feels miss the importance of the transposition st age of EU directives (for an overview, see Treib, 2006). Compliance refers to a state of conformity or identity between an actors behavior and a specified rule (Raustiala and Slaughter, 2002, p. 539). Compliance starts from a given norm and asks whether Member States conform. It is possible for someone to comply with the demands of a directive (transpose), without applying it (Treib, 2006), as well as the opposite. As such, to comply means something more than mere correctness in application enforcement. To comply entails conformity, which includes both the politics of complia nce and the processes 153

PAGE 154

leading to norm-conformity. Thus, to comply m eans an actual change in behavior (Haas, 1998), in-line with Commission mandates. It is the ar gument of this study that, non-communication of measures implementing directives, best exemplif ies such lack of norm conformity because the ease of detection is not enough to di scourage non-compliance (non-communication). Furthermore, this study is concerned with the processes and politics involved in transposing a directive rather than the ex post facto attempts at application and enforcement. Application and enforcement relates to what happens after a bill becomes a law (Treib, 2006) and the question of correctness entails whether the prescribed goals were achieved, which is conceptually different from compliance. In the EU case, as soon as a directive is transposed the burden of application and enforcement falls on the individual Member States administrative apparatuses and Courts (as we ll as the Commission through comp laints). By focusing on the politics of application and enforcement, one misses the crucia l importance of the compliance (transposition) stage pertaining to EU directives. This is critical as that is the stage at where most of the push and pull happens between governments, domestic groups and European principals (Commission). The transposition stag e is when states are more likely to shirk on their job, given the inability of the prin cipal to monitor compliance without someone hitting the fire alarm and it is the stage where policy drift is more likely since detection of this drift will happen some years after the policys transposition in the nonconformity and bad-app lication types of EU law infringement. More importantly, the transpos ition stage is the one during which states pick instruments to implement the directive, through a process closely resembling policy formulation (even if the goal is handed down by EU directives), which entails participation and politics. The amount of political participation states should allow in the policy process has always been a fundamental problem. 154

PAGE 155

Participation is a key component of the polic y formulation process as it may increase the likelihood of application or sustainability of a new policy. To the extent that groups feel allowed and can gain capacity for managing resour ces or new tasks and processes, there is an increased likelihood and sustainability of the inte nded policy. Furthermore, greater participation assures greater responsiveness to th e needs of the proposed beneficiaries, resulting in a better fit between needs and policy solutions, leading to in creased service-user satisfaction (Brinkerhoff and Crosby 2002). Transposition then, is one of the most important stages of policy implementation as it may very well determine th e success or failure of the policy. Figure 6-1, below, shows the stages of the policy process, th e actors involved, and the levels of participation as they exist in the EU policy process. EU Member States Policy formulation and adoption TranspositionPolicy formulation Application Enforcement -Commission -European Parliament -Council of Ministers -Administration -Interest groups -Administration -Administration -Courts -Commission (Citizen Complaints) Figure 6-1. The policy process in the EU: the actors and openness. Openness of Policy Debate However, transposition is not only important because it is the locus of politics in the implementation stage. As revealed in Chapter 3, while transposition (non -communication) is the easiest infringement to detect (by the Comm ission), oddly enough it is the largest source of infringement proceedings in the EU. Non-commu nication claims the lion s share with 61% (in the 1998-2007 period), as the legal base for cases in which infringement proceedings have been commenced. Additionally, 66% of non-communication cases reach the letter of formal notice 155

PAGE 156

stage in infringement proceedings, as compared to 60% for bad application, while 8% of these cases reach the referral to the Eu ropean Court of Justice stage, wh ich is less of a percentage than the other sources of non-complianc e, yet the total number of cases (920 cases) compared to the other sources of non compliance is accounts for 52% of all referrals. More specifically, the environment comes first in opened infringements (in 25% of all cases), and its share of noncommunication cases in infringement proceedings reaches again the first place with 61%. Figure 6-2 below summarizes these statistics. 020004000600080001000012000 Non-Communication Non-Conformity Bad Application Treaties Regulations Decisions Environment NonCommunication Non-ConformityBad Application Treaties Regulations Decisions Environment LFN 7228 775 1924 1333 3608 RO 2855 499 946 708 2091 ECJR 920 213 354 274 924 Figure 6-2. Total infringement proceedings by source of detection and the Environment sector, EU15, 1998-2007. More to the point, if we accept transposition data as valid indicators of member state noncompliance with EU environmental law, one of th e most striking features in transposition is the difference between the leaders and laggards on opened infringements and transpositions rates. The United Kingdom, a state considered a laggard under transpos ition rates, is actually 156

PAGE 157

part of the leader group in opened infringement proceedings with the fi fth best place (fewer opened infringements). As argued earlier, this oddity could be attributed to the Commission being wary of antagonizing bigger states, but that would not explain why France is second worse in opened infringement proceedings. Instead, a case could be made that the openness of the UK system to pressure groups accounts for the delays in transposition of direc tives. As the directive reaches the transposition stage it becomes part of the politics of tr ansposition that delay transposition but assist in bette r application since the issue was settled earlier (accounting for the relatively small number of opened infringements). Politics, is a feature of transposition and not of application, after law has been transposed into national legisla tion, political actors can do very little to resist its implementation and enforcemen t by a state dedicated to apply and pushed by the Commission with various enforcement and management mechanisms. Additionally, in reference to the leader group of states with an average transposition rate of over to 97%, most studies dealing with envi ronmental non-compliance either use data from before the 1995 accession or exclude Austria, Fi nland, and Sweden from the analysis (for a review of the litera ture, see Mastenbroek, 2005). This lim its the possible inferences and introduces a bias in the data as it eliminates interstate variation. The exclusion of the environmental leaders also creates a gap in the possible hypotheses as there is much to be said about the presence of these member states in European political institutions (lik e the European Parliament Environment Comm ittee). Any study wishing to include these countries, though, must take into account the newcomer effect, whereby the Commission grants a period of grace to newly accessed countries for approximately 2 years (Svedrup, 2004; Perkins and Neumayer, 2007). Thus, any study should begin from 1998 onwards. But what are other biases posed in the literature. 157

PAGE 158

6.3 Data Selection: Infringements as Biased Indicators of Compliance A first type of data used in non-compliance re search is infringement proceedings typically found in the Annual Reports on Monitori ng the Application of Community Law.55 Normally, these reports contain information on the source of non-compliance and the resulting steps in enforcement (formal notices, reasoned opinions and article 226,228 ECJ referrals56). As Hartlapp and Falkner (2009) suggest the stro ngest point of these data is the depiction of the interplay between the Commission and the states as the infringement proceedings progress. However, there are several documented limitations of infringements as indicators of compliance. First, there is simply no way to know wh ether the Commission, for whatever reason (resources, strategic-ness), responds to all infringements with th e same fervor. Put differently, the Commissions unwillingness or inability to mon itor and enforce all infringements of EU law introduces a bias in the dependent variable, and this type of resear ch looks only at the 'tip of the iceberg' of non-compliance (Hartlapp a nd Falkner, 2009, p. 292). When comparing the Commissions reaction to non comm unication and incorrect transpos ition (bad application, nonconformity), Hartlapp and Fal kner (2009, pp. 295-96) find that infr ingement procedures were initiated in 95% of their cases (40 out of 42) of non-communication, while incorrect transposition resulted in infri ngement proceedings in 51% of cases (22 out of 43). These findings are similar to the ones offered by this study, and highlight the fact that the Commission enforces non-communication more vigorously th an other sources of non-compliance. More importantly though, Hartlapp and Falkner (2009) find that infringement procedures were initiated in only 60% of the cases found to be in brea ch of EU law in their study, re-affirming the argument that infringement proceedings do not capture the actual level of non-compliance. 55 Available online at: H http://ec.europa.eu/community_law/infringeme nts/infringements_annual_report_en.htm 56 See Mendrinou, 1996; Mbaye, 2001; Bursens, 2002, Koutalakis, 2002; Trubek, 2002; Borzel, 2003; Sverdrup, 2004; Beach, 2005; Pank e, 2006; Perkins and Neumayer, 2007. 158

PAGE 159

This phenomenon may of course be attributed to the shortcomings of the Commission and the enforcement mechanisms of the EU. It is a fact that the Commi ssion has limited resources and even more limited ability to monitor a nd enforce non-compliance in the member states during the application phase. As such it depends largely on whatever the states report back on transposition through established reporting mechanisms which should account for the disproportionate follow-up on non-communications (since detection is easier). Infringement can be established with a simple yes or no, while assessing the correctness of notified measures requires more resources from the Commission and, is thus, more difficult. Still, states fail to report measures, knowing that the Commission will easily detect and disproportionately followup with enforcement. This exemplifies the element of compliance that is most important for this study, the lack of conform ity with EU norms. The Commission also depends on citizens, in terest groups, and companies to use the complaint procedure to ring the f ire-alarm in cases of bad application and non-conformity. In fact, out of 29,045 total detected cases of non-compliance for 1996-2007, 47% of them came from complaints, while 38% came from non-comm unication, and only 14% were a result of the Commissions own initiative. Complaints, the bigge st source for detecting infringements, are highly biased and therefore lead to highly unba lanced opening of infringement proceedings. Also, they are more likely to come in the posttransposition stages (after non-communication), and as such, they likely skew the true infringements numbers in member states. The distribution of tota l complaints, as shown in Figure 6-3, seems to verify theories pertaining to degrees of societ al activism (Eder and Kousis 2001), and that southern societies display a certain degree of distrust of their stat e institutions and therefore resort to the EU for assistance (Pridham and Cini, 1994). However, if one refers to the following Figure 6-4, the 159

PAGE 160

situation changes radically. Northe rn states (Ireland, Luxembourg) seem to complain more, and in fact three of the best performers of the EU (Finland, Denmark, Austria have higher complaints per capita than the worst performers (Italy, France Spain). The assumption is only verified in the case of Greece, which seems to be the only exception. 0 200 400 600 800 1000 1200 1400 1600 1800 LUXFINDENSWENLPORAUSBELUKIREGREFRAITAGERSPA Figure 6-3. Total number of complaints by Member State, 1998-2007. 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 LUXFINDENSWENLPORAUSBELUKIREGREFRAITAGERSPA Figure 6-4. Total number of complaints per capita, 1998-2007. Nevertheless, complaints, the major way th e Commission detects bad application and nonconformity does not seem to coincide with neither transposition nor infringement records. More to the point, complaints unus ually influence Commission follow-up on infringements, as such infringements are a biased toll with which to measure non-compliance. 160

PAGE 161

Further complicating the reliability of infri ngements as a measure of compliance is the odd practice by the Commission to chan ge counting rules, eliminate categories altogether for years and then bring them back in its reports. For instance, complaints were reported from 1982-1991, then lumped into a category that doesnt signi fy whether it is a complaint or not from 1992-1997, and then in 1998 complaints reappe ar as a category with the addition of own investigations and non-communication (Borzel, 2001). This makes the use of pre-1998 data highly unreliable, especially for infringement da ta. For example infringements by state and sector were only reported for 1988-1992, while since then infringement s have been reported per member state or per sector, making cross country comparisons in the same sector of policy practically impossible.57 6.4 The Dependent Variable: Transposition These issues with incomplete and inconsiste nt data are not unique to infringements though, transpositions suffer the same problems, though not to the same extent. The Annual Reports on Monitoring the Application of Community Law pr ovide a separate type of data reporting on progress in notification of measures im plementing directives (transposition).58 The data is consistent when it comes to years 2002-2009 with only a few changes needed as the Commission switched the columns to rows from 2002-2004. Th e situation worsens when it comes to 2001 and earlier as the Commissi on added more details in the environmental sub-sectors and introduced new sectors in 2001 such as Fisheries and Trade (which do not exist in 2002-2009, at least not in the way reported in 2001). Fisheries and Trad e dare also absent prior to 2000, while the remaining data and subsectors remain the same as in 2001. 57 Requests for original datasets to extract per state an d per sector data to both the Commission Secretariat and Environment DG remain unanswered for 6 months now and counting. 58 Since 2000 the same data are reported six ti mes a year on the Secretariat-Generals website: Hhttp://ec.europa.eu/community_law/directives/archmme_en.htm 161

PAGE 162

More importantly, as Hartlapp and Falkner ( 2009) argue, there are several issues with using transposition data as an indi cator of compliance. First, the date of notification may differ from the date of actual transposition. This s hould not be an alarming issue as states may complement their early transposition with addi tional measures. Some researchers have even attempted to complement notification data with national legislative so urces (see Mastenbroek, 2003; Kaeding, 2006), however utilizing an understand ing of compliance as conformity as in by this study makes this unnecessary, as a states notification of implementing measures is enough to designate compliance. A more important but equally less problematic issue with transpositi ons as indicators of compliance is that the data do not contain an o fficial statement different iating between timely and correct transposition (Hartlapp and Falkne r, 2009). The issue here is that data do not provide any information on whether the timel y notification of measures implementing a directive, are the correct measures. To worry about correct ness though is to introduce the same selection bias to the data as occurs when relying on the Commissions data on infringement proceedings, as explained earlier. Verification of the correctness of measures lays in the hands of the Commission, and given its limited expertise and resources the Commission may not pursue this goal to the fullest, same as with infringements. In fact, Hartlapp and Falkner (2009, p. 295) find that the Commission answered incorrect transposition by initiating an infringement proceeding just half of the time (51% in the cases of their study). Keeping in mind these limitations, this study do es not attempt to answer questions of correctness. Yet, as argued earlier and in ot her parts of this study, transposition timeliness depends heavily on the individual characteristics of each member state and the politics involved in transposing EU legislation to national law. The transposition stage is where the states are more 162

PAGE 163

likely to shirk and policy drifts occur, while it is also the easiest type of infringement to detect by the Commission and the type of infringement the Commission is almost certain to react to with infringement proceedings. Oddly enough, this rea lity should be enough to deter states from nonnotification and in fact it should lead states to always notify m easures implementing directives even if they are the incorrect ones.59 However, this moral hazard does not explain why states choose to disregard expected de tection and why they do so at an average of 61% for the 19982007 time period (as per the percent of infr ingement proceedings initiated for noncommunication). It seems more likely that states are wary of lying in their reports, even if the probability of being caught is less than the probability of faci ng the sharp end of the Commissions stick for non-communication. This alludes to the influe nce of voluntary and involuntary mechanisms behind the decision to not transpose a directiv e on time (in defiance of certain detection); something in line with compliance as conformity and should indicate th at transposition rates are an adequate indicator of noncompliance regardless of the corre ctness of the measures. More to the point, it is the argument of this study that since correct ness comes after the bill becomes law (or not), focusing on correctness as compliance misses the importance of politics, which largely take place in the transposition stage. Hence, taking a positivist legalistic approach (Konig, 2008), is appropriate to answer whether th e state completed the task assigned to it by the Commission or not. The Annual Reports on Monitoring the Applic ation of Community Law where used to count the number of directives for which memb er states did not notify the Commission with implementing measures. Cases were counted for the environment sector for 15 Member States 59 To notify measures that do not even exist in national la w is impossible as member states are required to provide standardized sheets with a paragraph of enacted national legislation corresponding to a directives provisions. 163

PAGE 164

(MS) from 1998 until 2007. Even though the Commission has changed the way they report transposition data throughout the years (especially for year s 1998-2000), it is possible to extract the data in a manner that makes all tables cons istent. The issues and changes made can be found in appendix A, while the following section provide s a discussion of the measures used in this study. 6.5 Operationalization of Independent Variables: International Relations Bargaining power (H1) in the Council of Ministers is simply the votes of the member states in the Council of Minist ers for the 1998-2007 period. This varies across states but is almost time invariant with a common change acro ss member states in year 2001 that coincides with the change in the allocation of the votes in the Council of Ministers in preparation for the 2004 Central and Eastern Eu ropean enlargement. The indicator for the EUs ability to ra ise governmental concern (H2) by providing information and assistance to domestic actors (w hich will in-turn press for action), will be measured using the Environment Directorate Ge neral program for providing operating grants to European environmental Non-Gove rnmental Organizations (NGOs).60 The list of funded NGOs is available for 1998-200761 and includes only non-profit a nd independent environmental nongovernmental organizations. These NGOs must be active at a Eur opean level to receive funding (have activities and members in at least three EU member states). This also makes the data an ideal candidate for establishing not only NGOs pres ence in the member states, but also whether they are active in coordinating and channeling the views of national organizat ions and citizens as input to the decision making pro cess, and raising awareness while educating the citizens and 60 The legal base for the program is the LIFE+ Regulation which provides for funding of operational activities of NGOs that are primarily active in protecting and enhancing the environment at European level and involved in the development and implementation of Community policy and legislation. 61 Available online at: http://ec.europa.eu/environment/ngos/list_ngos97_07.htm. 164

PAGE 165

governments about the environment. The list of funded NGOs includes, of course, several umbrella NGOs based in Brussels and active in most EU member stat es. Counting these NGOs as only active in Brussels woul d miss their governmental con cern effect on the respective governments of their members. Additionally, Be lgium would represent almost all the funded NGOs from the EU (as most are based there), sk ewing Belgiums governme ntal concern effect and under-representing the other Member States As such, extracting the funding data per Member State and year, required extensive archival research to make sure that all national suborganizations are accounted for. This included visiting funded NGO member lists for each of the years in this study and manually identifying each of the national members of these supranational NGOs. For instance, in 2007 there were 30 such NGOs funded by this program, comprising of 763 national NGO members. The indicator for whether the EU system of enforcement can be considered (by the member states) to provide for a reasonable contractual environment (H3, where non-compliance will be detected and followed-up by EU institutio ns) will be measured by European Court of Justice case filing at t-1. The assu mption being that Member States with a Court case filed in the previous year will be less inclined to non-comply with directives in the following year. In this case, I count the number of cases filed with Court against a member state only on environmental directives. Cases were counted by country and by year of filing, rather than by year of judgment. Because this study is concerned with the effect of an opened case to next years compliance the filing date is important as it is closer (time-wise) to the conditions that created the nontransposition.62 However, there is no such list of cas es compiled or available for 1998-2007, so the data was extracted by searching the online case-law da tabase of the ECJ.63 Specific attention 62 Mbaye (2001) uses the filling date as well. 63 Searchable database available online at H http://curia.europa.eu/jurisp/cgi-bin/form.pl?lang=en 165

PAGE 166

was paid to both the filing and judgment date (a nd data was extracted and compiled for both), as well as the number of environmental directives included in each judgment.64 To measure the EUs abi lity increase governmental capacity (H4, by transferring information/skills and aiding monetarily in the process), annual LIFE+ project funding is used for 1998-2007. The LIFE program is the EUs fi nancial instrument supporting environmental and nature conservation projects throughout the EU and is comprised of three components (only two of which are of specific interest for this study). The LIFE+ Environment Policy and Governance component co-finances technological projects th at offer significant environmental benefits, for example process or efficiency improve ments. This part of LIFE+ also helps projects that improve the implementation of EU environm ental legislation, that build the environmental policy knowledge base, and that develop enviro nmental information sources through monitoring. The LIFE+ Information and Communication progr am co-finances projects that spread information about environmental is sues, such as climate change and conservation. This strand of LIFE+ can also support environmental awarene ss and training campaigns. As such, the LIFE+ project is a direct measure of th e EUs ability (or lack thereof) to increase governmental capacity by transferring information and skills, while aidi ng monetarily in the implementation of EU legislation. Unfortunately, the data is not readily available at th e LIFE+ project database for all Member States per year in a tabular format. Hence, the data was extracted per member state in the predefined period (1998-2007)65 for both LIFE+ components and represents the count of LIFE+ projects in the member states. 64 It frequently happened that the ECJ delivered a judgment on several directives with one case against a member state. 65 Available online at: H http://ec.europa.eu/environment/life/project/Projects/index.cfm Under headings Information and Communication, Environment, and Environment Preparatory Action from the drop down menu per MS. 166

PAGE 167

Finally, the indicator for whether member state publics are favorable toward EU environmental policy-making (H5, as opposed to just by their own government) is measured using the European Commissions Standard Eurobarometer survey66 for the 15 member states examined by this study, from 1998 until 2007. The survey instrument used to get at favorable to EU environmental policy publics is QA20a for 2 007 (see Appendix B), which reads out a list of 21 policy areas (from crime to housing to prot ecting the environment) to the 1000 per state respondents and asks (for each of the areas), if people think that decisions should be made by their own government, or made jointly within th e European Union? The procedures and issues with data extraction are also discussed in Appendix B. All the raw datasets and survey instruments are available through the Inter-University Consortium for Polit ical Science Research (ICPSR).67 6.6 Operationalization of Independent Variables: Regulatory Federalism The existence of active environmental groups in the member states (H6), must not only include the number of groups in each state for the 1998-2007 peri od, but it must also indicate whether those groups are truly active in the domes tic arena. One could use environmental groups registered with the Commission as a source but th at does not mean they ar e necessarily active in national politics. Another source could be the funded NGOs under the Commissions LIFE+ program but except for the issues of multi-colinea rity with H2 above, it still does not ensure those groups are active in national politics as per definition they must have activities and members in at least three member states. Thus, the best way to measure this variable is the use of complaints to the Commission for the environm ent sector (by state) for the 1998-2007 period. 66 Each survey consists in approxima tely 1000 face-to-face interviews per Member State (except Germany: 1500, Luxembourg: 600, United Kingdom 1300 including 300 in Northern Ireland). Conducted between 2 and 5 times per year, with reports published twice yearly. Available at: H http://ec.europa.eu/public_opinion/standard_en.htm 67 Available online at: H http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/23368.xml 167

PAGE 168

Even though the indicator is somewhat crude an d there is no way to control for complaints coming from other sources, it is reasonable to assume that active Environmental NGOs will use this venue more than other pressure gr oups (businesses and private citizens). Membership in the European Parliaments E nvironment, Public Health, and Food Safety committee (H7) is measured using the E NVI committee members register for 1997-2008. However, there is no one register or members list for the ENVI committee for these years. As such multiple sources and methods were used. The members for the 6th parliamentary term (2004-2009) are readily available thr ough the European Parliament (EP).68 The members for the 5th parliamentary terms where located in the ENVIs activity report for 1999-2004.69 Locating the ENVI committee members for the 4th parliamentary term proved more difficult due to the unavailability of such a list. To this end, extensive rese arch was conducted in the archives of the European Parliament. This included looking up all 72970 Members of the European Parliament (MEP) listed for the 4th parliamentary term, and verify ing ENVI committee membership.71 Of course, specific attention was paid to dates of membership as this study requires only those MEPs that served on the ENVI committee fo r 1998-2007. However, the European Parliament does not have the same number of MEPs for the three parliamentary terms in this study due to consecutive new member state accessions and Treat y revisions. Additionally, the available seats on the ENVI committee are limited. As such any new member states will affect not only the number of MEPs, but also the composition of the ENVI committee (as w ith the 2004 Central and Eastern European enlargement). For these reasons the number of ENVI committee members per 68 Available online at: H http://www.europarl.europa.eu/members/expert/c ommittees/search.do?committee=2353&language=EN 69 Available online at: H http://www.europarl.europa.eu/comparl/envi/pdf /activity_report/activity_report_5th_leg.pdf 70 Even though the EP had officially 626 MEPs for 1994-1999. 71 Available online at H http://www.europarl.europa.eu/memb ers/archive/term4.do?language=EN 168

PAGE 169

member state was expressed as the ratio of E NVI members over the number of total MEPs for each Member State, per year. The indicator for salience (H8) is measured using the European Commissions Standard Eurobarometer survey72 for the 15 member states, examined by this study, from 1998 until 2007. The survey instrument used to measur e salience is QA6a for 2007 (see Appendix B), which reads out a list of 14 policy areas (from crime to housing to the environment) to the 1000 per state respondents and asks them to choose tw o of the most important issues facing their country. The procedures and issues with data ex traction are also discussed in Appendix B. All the raw datasets and survey instruments are av ailable through the Inte r-University Consortium for Political Science Research (ICPSR).73 The indicator for existence of special interest po litical parties (H9) in Member states is the number of seats won by G reen parties in Member State parliaments for 1998-2007.74 However, several of the states in this study had elections in 20 07 (while the aforementioned data are up to 2006). As such, an additional source was consulted for green party seats in national parliaments75 for 2007 (the last year of this study). Th e seats won were expressed as the number of seats over the total number of seats available in each state parliament because of the difference in the number of seats between member state par liaments (and in some cases the difference in total seats per national electi on). A final issue occurred with some of the Scandinavian leftist/green coalition parties (especially fo r Denmark). Denmark has had one MEP associate with the Greens of the Eu ropean Parliament for 1998-2007, a nd Danish left/green national 72 Each survey consists in approxima tely 1000 face-to-face interviews per Member State (except Germany: 1500, Luxembourg: 600, United Kingdom 1300 including 300 in Northern Ireland). Conducted between 2 and 5 times per year, with reports published twice yearly. Available at: H http://ec.europa.eu/public_opinion/standard_en.htm 73 Available online at: H http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/23368.xml 74 Available online at: H http://www.europeangreens.org/cms /default/dok/167/167260.national_el ection_results_since_1979@en.htm 75 Available online at: http://psephos.adam-carr.net/ 169

PAGE 170

parties share the same ideals as European Greens.76 Nonetheless, the greens of Denmark (De Gronne) failed to gain representation in the Danish Folketing for the years of this study; as such no green party was calculated for Denmark (as individual influence an d seats could not be determined). The indicator for member state overall commitment to the environment (H10) will be measured with Organization for Economic Co -Operation and Development (OECD) data on government expenditure by function.77 The amount of government expenditures for environmental protection is available for 1998 to 2007, and the data available is in millions of Euro for all countries but Denmark, Sweden, and the UK, for which the amounts reported are in national currencies (of course these were standardized into Euros). The indicator for industry clout (H11) is m easured with Organization for Economic CoOperation and Development (OECD) data on i ndustry (including energy) employment as a percent of total employment for each state for 1998-2007.78 The indicator for a member states need for regulation (H12) is measured by the Greenhouse gas emissions for each state for the 1998-2007 period.79 These are data on greenhouse gas emissi ons, sent by countries to the United Nations Framework Convention on Climate Ch ange (UNFCCC) and the EU Greenhouse Gas Monitoring Mechanism. To gauge whether the industry in each member state is environmentally conscious and collaborates with government au thorities in environmental pr otection (H13), EMAS (EcoManagement and Audit Scheme) data is used, which measures the number of organizations with 76For 1994-1999 one member in the Greens. For 1999-2004 one member in European United LeftNordic Green Left. For 2004-2009 one member from SF to European GreensEuropean Free Alliance. 77 Available online at H http://webnet.oecd.org/wbos/ under General Government Accounts. 78 Data available online at: H http://stats.oecd.org/Index.aspx?DataSetCode=SNA_TABLE3 The data was standardized for U.S. $, constant prices, constant PPPs, OECD base year. 79 Date available online at: H http://dataservice.eea.europa.eu/data service/metadetails.asp?id=1089 170

PAGE 171

a registered environmental management system. The data is readily available on the Eurostat website for the 1998-2007 period.80 The EMAS is a voluntarily environmental management system implemented by companies and other orga nizations from all sectors of economic activity including local authorities, to evaluate, report on, and improve their environmental performance. However, powerful economic interests may press fo r less legislation and enforcement especially in countries where industry is advanced and co mprises of a big portion of that states GDP. Staffing levels (H14) are an important indicat or of a governments capacity to transpose legislation. The indicator for st affing levels (H14) is measured with OECD data on total compensation of environmental employees paid by the government81 and these amounts are available for 1998 to 2007. To measure the existence of a cozy regulator y efficiency (H15) in policy making the World Banks World Governance Indicators database is used.82 This indicator is defined as measuring perceptions of the ability of the government to formulate and implement sound policies and regulations that permit and promote private sector development. The data is available from 1996-2008 but not available for 1999, so a linear interpolation of the values for that year was necessary. For a robustness chec k, data from the International Country Risk Guide83 on Bureaucratic Quality where also extr acted and compiled for the 1998-2007 period. Finally, the existence of corruption (H16) shoul d have a negative effect on compliance. In some cases, corruption is the ugly face of bureaucratic discretion as discretion leaves bureaucrats with ample time for bribes. The EU system of implementation is considerably decentralized and discretion (as to implementi ng measures) is the modus operandi for national 80Available online at: H http://epp.eurostat.ec.europa.eu/tgm/table.do?tab=t able&init=1&language=en&pcode=tsdpc410&plugin=1 81 ibid 82 Available online at http://info.worldbank.org/governance/wgi/index.asp 83 Available online at: H http://www.prsgroup.com/ through LexisNexis 171

PAGE 172

bureaucracies. The indicator for corruption will be the Transparency Internationals Corruption Perception Index for the 1998-2007 period84 (a higher level indicates a less corrupt state). For a robustness check, data was also extracted and co mpiled on the control of corruption from the World Banks World Governance Indicators. This indicator is defined as measuring perceptions of the extent to which public power is exercised for private gain, including both petty and grand forms of corruption, as well as "captu re" of the state by elites and private interests.85 6.7 Econometric Models The choice of estimation method is dictated by the nature of the measure used for noncompliance ( the dependent variable in the empirical model), which is the number of directives for which a state has not notified the EU Commissi on with measures implementing them during each year in the study period. That is non-compliance is a count variable that takes on non-negative and discrete values and with possi ble zero outcomes for at least some members of the population (if the state ha s complied with all di rectives). This pr operty of the noncompliance measure implies that a linear regression model (y(|) Ey xx, where denotes non-compliance in the population conditional on the vector of the independent variables, and (|) Ey xx is the vector of regression coefficients ) using the ordinary least squares (OLS) estimator is not well-suited, since it is possi ble to obtain negative predicted values for noncompliance. That is, since non-co mpliance is strictly positive ( ) the predicted value of noncompliance in the population, should be nonnegative for all However, if 0 y x(|) Ey x is the OLS estimated vector of regression coeffici ents, there will usually be values of x such that implying that the predicted value of is negative. x 0 y 84 Available online at: http://www.transparency.org/policy_research/surveys_indices/cpi 85 Available online at http://info.worldbank.org/governance/wgi/index.asp 172

PAGE 173

Hence with count data, should be modeled by choosing functional forms that ensure its positivity for any value of (Wooldridge, 2001). The most popular functional form is obtained by assuming that the dependent variable follows a Poisson distribution. However, the Poisson distributional assumption imposes the restriction that th e conditional variance and mean of the dependent variable are equal, which is more than often violated in empirical applications. As Cameron and Trivedi (2005) demonstrate, violation of this assumption leads to overestimating the statistical significance of the explanatory variables in the Poisson regression. To handle this problem several techniques have b een developed to obtain correct standard errors under Poisson regression for overdis persed (variance of dependent is higher than its mean) or underdispersed data (Cameron a nd Trivedi, 1998, pp. 70-76) but negative binomial regression has become a standard method to deal with such ki nd of data (Hilbe, 2008). In this chapter, both Poisson and negative binomial models will be employed to model non-compliance in the EU, while testing for the Poisson distributional assumption. (|) Ey xxAnother distinguishable characte ristic of the employed dataset is its time-series and crosssectional dimension since this study obser ves 15 countries annually from 1998 to 2007. Therefore, panel data techniques should be employed in addition to accounting for the count nature of the dependent variable. These techni ques allow taking advant age the cross-sectional information reflected in the differences betw een countries and the time-series information reflected in the changes within countries over time Further, it is possible to control for omitted variables that differ between countries but are constant over time and for omitted variables that vary over time but are constant between countries. However, one should keep in mind a critical distinction between two t ypes of panel data, which has both theoretical and practical implication for research. The first type of these is Time 173

PAGE 174

Series Cross Sectional (TSCS), otherwise known as macro-panels (Baltagi, 2008). This type of data usually involves a moderate size of countries or units, which is fixed and not a sample of the population of countries or un its, usually observed annually for a long period of time. In this sense, when someone wants to update the panel, countries are kept fi xed and only years are resampled, as inferences of interest are conditi onal on the observed countries and one has to deal with cross-country dependence (F reedman and Peters, 1984). In contrast, Panel data (otherwise known as micro-panels), involve re peated cross-section data collected over a short period of time for a large number of units or countries sampled out of the total population (in some random sampling scheme), hence there is likely no issue of cross-units dependence. This is because the units or individuals in micro-panels ar e of no real interest as all inferences are made about the underlying sampled popul ation of units (Beck, 2001). NTNThe theoretical difference between TSCS and micro-panels emanates from the different econometric care they require. Asym ptotics for TSCS data are in T (T and is fixed), which means that the number of units or countries is fixed and an asymptotic argument should be based on the time dimension. However, in some cases asymptotics can be also for large and (Baltagi, 2008). In contrast, all as ymptotics in micro-panels are in ( and T is fixed) as sample sizes can be t hought of as getting larger and larger (Beck, 2001). Also, with a long time-series for TSCS there is a danger of encountering issu es of nonstationarity, like unit roots, structural breaks and coin tegration, whereas, given that T is short for each country, micropanels are safe from thes e issues (Baltagi, 2008). NNN NTAs a general warning, hereafter, when this st udy uses panel it will refer to TSCS data as micro-panels and TSCS are both ty pes of panel data. The remainder of this chapter is organized as follows: Section 2 reviews and operationalizes the hypotheses developed in Chapters 4 and 5. 174

PAGE 175

In addition descriptive statistics of the developed measures are presented in this section. A review and empirical estimation of panel Poisson and negative binomial estimators is provided in Sections 3 and 4, respectively, while in Section 5 the best mode l that explains non-compliance is presented. Conclusions are dr awn in the last section. 6.7.1 Panel Poisson Regression Model The Poisson distribution is ofte n used in estimating the number of occurrences of an event over a specified interval of time or space. For th e number of occurrences of an event to be described as a Poisson-distributed random variable two propertie s must be satisfied: (1) the probability of an occurrence is the same for any two intervals of equal length and (2) the occurrence or nonoccurrence in any interval is independent of the occu rrence or nonoccurrence in any other interval of time or space. The basic Poisson probability specification is (Hausman et al., 1984) (|) !itit y it itit ite fy y x. (1) In this study, denotes the number of directives a country (ityi1,...,iN ) did not comply with (non-compliance) in time t ( ), is a -dimensional vector of linearly independent variables, 1,...,t ...,]kitTitx k1[it,it x x x, that are thought to determine is the probability of ity (|itfy )itxit y occurrences of non-compliance for country at time t conditional on the vector of independent variables, and iit denotes the intensity or rate parameter for country at time t. The first two moments of the Poisson distribution are given by (Hausman et al., 1984): iMean: (|)itititEy x (2) Variance: (|)itititVy x. (3) 175

PAGE 176

This shows the equidispersion property of th e Poisson distribution, namely the conditional mean is equal to the conditional variance of the distribution. As previously mentioned, in many empirical applications it is very common to find that the variance is greater than the mean, implying overdispersion in the data (see Cameron and Trivedi, 1998, pp. 97-106 for other sources of overdispersion). Hence, a test of overdispersion must be employed whenever the Poisson regression model is used and if overd ispersed cluster-robust standard errors or alternative estimators such as the negative bi nomial regression model should be employed. Further, by definition of the Po isson distribution it is assumed that all observations occur randomly and independently over time. The time independence property constitutes both an advantage and a disadvantage in em pirical applications of micro-pa nels. As an advantage, if the counting process is Poisson over time 1,...,tT with parameter it then the aggregate data over the specified period are also Poi sson distributed with parameter 1 T i t i t (Hausman et al., 1984), which permits development of the condi tional maximum likelihood estimation method. On the other hand, the time dependence assumption is often violated in empirical applications due to serial correlation of residua ls in econometric specifications. This creates the need to test for and correct if necessary for serial co rrelation or obtain robus t standard errors. For the parameterization of the intensity parameter it and to ensure that (|)0itititEy x ) (since the dependent variable in our model is non-negative), the standard assumption in the literature is to us e the exponential mean parameterization (|)exp(ititit itEy xx (4) where as above, is a (1-dimensional vector of linear ly independent variables and itx)k is a parameter vector. Equations (k 1)(1) and (4) jointly define the Poisson regression 176

PAGE 177

model. Note that since (|)(|)exp()itit itit itEyVy xxxityexp()it the Poisson regression is intrinsically heteroscedastic, implying the need again to obtain robust standard errors in the estimation procedure. In what follows, several panel data estimators along with their properties for the panel Poisson regression model are presented. 6.7.2 Pooled or Population-Avera ged Poisson Regression Model The pooled Poisson estimator assumes that is Poisson distributed with a mean given by equation (4) (|)ititEy xx and that the observations are independently distributed. Consequently, crosssection observations can be stacked over time and the model can be estimated using cross-sectional estimation methods. Substituting equation |itity x(4) in equation (1) the pooled Poisson regression model becomes exp( exp())exp(exp()) !it itit ity y(|)ititfy xx011)]ln(e) ...itit it x, (5) where it is implied that lnln[(|it it kit E yx x xx! ]itititeyx. (6) According to Hausman et al. (1984), the log likelihood for this Poisson specification for a sample of countries over T time periods is given by N11()[NT it itLy x, (7) Maximizing this likelihood function with respect to we obtain the grad ient (first-order conditions) and the Hessian (secondorder conditions), respectively 2[( )]0 [())]it itit itit itye He x xxxit itL L x, (8) 177

PAGE 178

The second order conditions for a maximum at (vector of estimated regression coefficients) are satisfied since the Hessian matrix H is negative definite implying that the loglikelihood is globally concave. To obta in the optimal parameter estimates ( ) from the firstorder condition in equation (8) an iterative algorithm must be used since the first-order condition is nonlinear in Hence, for the estimation of the pooled Poisson estimator the NewtonRaphson method is selected. This iteration algorithm starts with an arbitrar y starting value for the vector (e.g., values from an ordi nary least squares estimation) and new estimates are obtained in each iteration 1,..., j M from 1 )]1 [(() j jjHjg where () g denotes the gradient in equation (8) This iteration proce dure stops and the optimal parameter estimates are obtained when a convergence criterion is met, e.g., Trivedi and Munkin (2009) suggest obtaining a panel-robust or cluster-robust with cl ustering on the country i estimator of the covariance matrix.1 00010.jj86 This corrects standard errors fo r any dependence over time within each country and for overdispersion. A variation of this pooled Poisson estimator is the population-averaged Poisson estimator or Generalized Estimating E quations (GEE) estimator, whic h relaxes the assumption of independence of to allow explicitly for correlation among the observations and for different models of th e underlying correlation su ch as autoregressive processes. This estimator is based on the Ge neralized Method of Mo ments (GMM) procedure (see Cameron and Trivedi, 2005, Chapter 23.2 for a more general treatment) and consistency of this estimator (as the pooled estimator) re quires that the specification in equation ity,[()()]it ists it isrCoryeyexx (4) is correct. As the pooled estimator, it does not re quire for the data to be Poiss on distributed but in that case 86 A detailed description is provided in Cameron and Trivedi (2005), Chapter 24.5. 178

PAGE 179

robust standard errors should be obtained. Both the pooled Poisson model based on the maximum likelihood method and the populationaveraged model (GMM method) will be estimated and comparison of the estimates will be presented. The pooled or population-averaged Poi sson model presented in equations (5) and (6) makes the implicit assumption that the effects on non-compliance are homogeneous across countries and over time since all countries have the same intercept and slopes over the entire period. That is, withinand between-countries effects are equal. Wh ile this is a very restrictive assumption it can be tested empirically and is th e most commonly-used approach for TSCS data in the political scien ce literature because of its simp licity (Wilson and Bu tler, 2007). Although unobserved heterogeneity cannot be accounted for e xplicitly in the model and it might induce an omitted variable bias (Hsiao, 2003) the error terms of the regre ssion can be corrected for this unaccounted heterogeneity using panel-corrected standard errors (Beck and Katz, 1995) or cluster-robust standard errors with clustering on the countries as mentioned above. One way to explicitly account for heterogeneity across coun tries or through time is to use a variable intercepts model that is presented in the next section. 6.7.3 Fixed Effects Poisson Regression Model Fixed effects regression model allows controll ing for omitted variables that differ between countries but are constant over time or variable s that cannot be observe d across countries (e.g., culture) by including country-specific dummies in the regression model. As a side effect in this case, fixed-effects models cannot be used to in vestigate time-invariant causes of the dependent variable. That is, independent va riables that are time-invariant cannot be included in the model.87 87 An exception is the fixed-effects negative binomial model developed by Hausman et al., (1984) in which the model can contain an overall constant and time-invariant independent variables (Greene, 2007). 179

PAGE 180

It also allows controlling for unobservable va riables that change ove r time but not across countries (e.g., EU regulations, treaties etc.) by including tim e-specific dummies.88 To introduce heterogeneity across countries and time in the Po isson model, I consider the following parameterization of the conditional mean presented in equation (6) which is an extension of the parameterization used by Haus man et al. (1984) and Cameron and Trivedi (1998, p. 279) to allow for time fixed-effects (t wo-way fixed effects e rror component model): ()[|,,]it itititititit itEy e e x xx 1,..., iN 1,..., tT (9) or equivalently, ln()ln([|,,])it itititititEy xxi, (10) where lni denotes all the unobserved, omitted variables from equation (5) which are specific to each country and are time invariant, ln()tt denotes all unobserved omitted variables from equation (5) that are common to all countries but change over time (countryinvariant variables). That is, c ountry-invariant variables are the sa me for all cross-sectional units at a given point in time but vary through time. It is this ability to control for all time-invariant variables or country-invariant va riables whose omission could bias the estimates in a typical cross-section or time-series study th at reveals the advantages of a panel. That is, the fixed-effects model removes the effect of those time-invariant characteristics from the predictor variables (if using only country-specific dummi es) or of those country-invari ant characteristic s (when using only time-specific dummies) so that the net effect of the independent vari ables can be assessed. 88 It should be noted that a fixed-effects model in the literature is usually meant to include only country-specific dummy variables. If in addition, time-specific dummies are included then the fixed-effects model is termed as a twoway error component model. 180

PAGE 181

Given the way this study treats the unobservable country -specific effect i and the unobservable time-specific effect t it then differentiates betw een fixed-effects and randomeffects models. Focusing on the unobservable country-specific effect, if we treat i as fixed parameters to be estimated, e.g., as coefficients of country-specific dummies in the sample, then the fixed-effects approach shoul d be followed. In this case an overall intercept from is excluded. Instead, if we assume that itxi are random variables that ar e drawn from a distribution we have a random-effects model a nd an intercept is included in One important assumption of the fixed-effects model with country-specific dummies is that the data must be independent over countries for a given year. If that is the case then random-eff ects models which are considered in the next section might be more appropriate. The same arguments are true for the time fixed effects itxt Focusing on the fixed-effects model and given the small and T in this study, the fixedeffects Poisson model is easily estimated (C ameron and Trivedi, 1998, pg. 280-286 and 291) by maximum likelihood or the method of moments. Th at is, the exponential mean specification can be re-written N(10) as: 1 11 11ln() ...NT it jjitsistit kkit jsdDx x, (11) Where is a country-specific dummy va riable equal to one if the itth observation is for country jitd j and zero otherwise, and ist D is a time-specific dummy variable equal to one if the it observation is for time and zero otherwise. Therefore, I can regress on and where I have dropped one ti me-dummy to avoid the dummy variable trap. The model in sxity12,,...,ititNitddd121),iti itDD( TD,...,tit(11) can be simplified by considering only country-specific effects or 181

PAGE 182

time-specific effects, which resu lts in increased degrees of fr eedom. Whether country-specific dummies or time-specific dummies should be used is not only an empirical question (can be addressed by a joint test of the coefficients of each set of dummies) but also depends on the nature of independent variables included in the model and the th eoretical questions addressed in each study. Considering equation (11) with only country-specific effects, the country-specific dummies capture all of the between-countries vari ation in the data and so the effects of the independent variables capture solely within -country effects on the dependent variable (noncompliance).itx89 This implies that any time-invariant inde pendent variables (var iables that differ only across countries and not acros s time) are not identified a nd therefore between-countries hypotheses cannot be tested. This problem is cited by Mbaye (2001) in choosing a model without country-specific effects. Specifically it is argued that cross-sectiona l variables test theory and if proper theory is examined, then cross-country variance is taken into account by these crosssectional variables rath er than the country-s pecific dummies. Though, a lagged dependent is appended in the model that captures cross-coun try heterogeneity. Further, including countryspecific dummies has the effect of slowly-cha nging over time indepe ndent variables being insignificant (Beck, 2001). This study has several variables that are slowly-changing over tim e (e.g., measure for environmental committee membership) or are pu rely cross-sectional (e.g., measure for centralization). As Greene (2007) demonstrates, an estimator that allows including time-invariant independent variables is the ne gative binomial fixed-effects deve loped by Hausman et al. (1984) 89 In this case and for TSCS data in a linear specification (n ot the Poisson model considered here), as Bartels (2008) mentions, the effects of the independent variables are interp reted as: for a given country as the independent variable varies across time by one unit, then the dependent variable changes by units (increases by if is positive and decreases by if negative). 182

PAGE 183

and is presented in a subsequent section. In this estimator, the fixed effects enter the model through the variance of the random variable (non-compliance) rath er than the conditional mean function as in equation (11) This has the implication that time invariant variables can coexist with the country-specific fixed effects. Howe ver, Greene (2007) argues that the negative binomial fixed-effects of Hausman et al. (1984) is cons istent in the contex t of a model with a heterogeneous variance but cannot conclude whether it is a consiste nt estimator of a model that contains a heterogeneous mean. In the context of the present st udy is of particular importance to control for the effect of omitted and/or unobservable variables that capture general developments common to all member states but changing over time. Such developments include annual increases in the number of environment directives and other regulations, both of which might plausibly impact member state compliance (Borzel, 2001; and Neyer, 2004). They also incl ude changes in the commissions willingness to pursu e infringement proceedings agai nst member states (Hattan, 2003), developments in the European legal regi me for enforcing and sanctioning noncompliance (Alter, 2000), and institutional de velopments such as treaty revisions and enlargement. As long as these developments affect all member states a pproximately equally, they can be captured with year-specific time dummies, without the need to model each factor independently. Therefore, in all empirical models that will be considered in this study a set of time-specific dummies will always be included. 6.7.4 Random Effects Poisson Regression Model Random effects models allow controlling for unobserved omitted variables that are (1) constant over time but vary betw een countries, and/or (2) fixed between count ries but vary over time. Unlike the fixed effects model, the variatio n across countries is assumed to be random and uncorrelated with the independent variables included in the m odel. An advantage of random 183

PAGE 184

effects is that time-invariant variables can be incl uded as predictors, allowing to test the effects of between-countries variables, whereas in the fixed effects m odel are absorbed by the countryspecific dummies. The parameterization of th e conditional mean presented in equation (6) for the random effects model is similar to that of the fixed-effects model: ()[|,,]it iititititit iEy e e x xx 1,..., iN 1,..., tT (12) or equivalently, ln()ln([|,])it ititiiitEy xx, (13) where are identically and indepe ndently distributed (iid) and Poisson distributed; ityi are iid Gamma distributed with scale and shape parameters (i.e., (,)i ) so that ()1iE and ()1/iV The joint probability density fo r the Poisson random effects model with gamma-distributed random effects is then given by (Cameron and Trivedi, 1998, p. 288): 1(,...,|,) !(it it ty y it t it iiTiit it t t it it ty fyy y) x, (14) Where it is assumed that the random effects term i is not correlated w ith the independent variables. For this distribution, the conditional mean is ()ititEy and the variance is 2()(/)itititVy which accounts for potential overdispers ion in the data of the NB2 form since the variance is quadratic in the mean (Cameron and Trivedi, 1998, p. 63). This formulation also provides a testable hypothesis of the ra ndom effects versus the pooled Poisson model. Specifically, if the parameter (1/ ) is equal to zero, then the conditional mean ()ititEy is equal to the conditional variance ()ititVy and so the pooled Poiss on model is appropriate. 184

PAGE 185

In another formulation the random effects i (iie ) can follow a normal distribution 2(0,) N The joint probability density is then given by: 22/2 1(,...,|) (,) 2iiiTi ititii te fy y F y d xx (15) Both equations (14) and (15) can be estimated using the method of maximum likelihood in any standard software package. Baltagi (2008, p. 17) and Hsiao (2003, pp. 41-43) argue that it is more appropriate to use a fixed-effects model instead of a random effects model when infe rences are to be drawn upon the observed units (countries in this study) that ar e fixed and we are not interested in extending inference to a larger pop ulation of similar countries. In contra st, the random eff ects model is an appropriate specification if units or countries are randomly drawn from a large population of units or countries ( is large relative to T) and inference pertains to the population from which sample was randomly drawn. Although, in this st udy, the member states represent the population for the period of study, inferences can be drawn for a larger popul ation. To assess the appropriateness of the random effects model in co mparison to the fixed effects model a Hausman test should be performed (Cam eron and Trivedi, 1998, p. 293). NN6.7.5 Negative Binomial Regression Models The negative binomial distribution is used to handle the overdispersion problem apparent in many empirical applications under the Poisso n regression model. Fo llowing Hausman et al. (1984) the assumption is that non-compliance follows a Poisson distri bution with an intensity parameter ityit where the Poisson parameter it follows a gamma distribution with parameters 185

PAGE 186

(,)it and specify itite x with common both across firms and over time.90 Then, the negative binomial distribution of with parameters ityit and is given by () 1 1 (|) 1it ity ititfy x ()(itit itity y )itit1). (16) The first two moments of the negative binomial distribut ion are given by Mean: (|itEy x (|)itit (17) Variance: ( 1 )itVy x. (18) Since 0 the variance of the negative binomial always exceeds its means, thus allowing for overdispersion. The overdispersion vanishes if 0 and the negative binomial distribution converges to the Poisson distribut ion (in the limit) with parameter it (Winkelmann, 1997) allowing to test for overdispersion. The model in equation (16) presents the pooled or populatio n-averaged negative binomial regression models and standard maximum likelihood methods (Cameron and Trivedi, 1998) can be used to estimate them. To allow for fixe d and random effects, the negative binomial regression in equation (16) can be modified as follows: () 1 1 (|) 1it ity t i itit i ify x ()(iti itity y 1), (19) where the dispersion parameter i accounts for country-specific effects in the fixed-effects model and for both country-speci fic and time-specific effects in the random effects model. Looking at within-countries effects on ly, the specification in equation (19) yields a negative binomial for the ith country with dispersion (variance divided by the mean) equal to 1i That 90 Note, that our specification differs from that of Hausman et al. (1984) since our parameter is the inverse of their parameter 186

PAGE 187

is, we have constant dispersion within each country. For a random effects model, i is allowed to vary randomly across c ountries by assuming that 1/ (1)i follows a beta distribution with parameters (Hausman et al., 1984) resulting in the following joint probability of the counts for the th country (,) rsi 11 1) )T it t ity y 11()()( () ()(1) ()()(TT it t t itit it TT itit it ttrsrs y fy y rsrs (|)it x (20) For the fixed-effects with country-specific effects overdispersion model, the joint probability of the counts for each country is conditioned on the sum of the counts from the country, that is the observed This yields: 1 T it ty 11 1 11) ()T it t it 11()( () ) (1) ()TT T it it t t it it it TT t it it it tty y fyy y y (|,it xity. (21) To consider a fixed-effects with time-specific effects overdispersion model, the joint probability in equation (21) should be conditioned on the sum of counts across countries for each year, while one can estimate a two-way fixe d effects model by including time dummies in the vector of independent variables and then estimate the model in equation 1 N iitx(21) as shown in Cameron and Trivedi (1998, p. 291). The attractiveness of the above fixed-effects negative binom ial model, where the variance of the distribution is a multiple of its mean (NB1 form), relies on the fact that time invariant variables can be estimated along with the country-specific fixed effects. Although the most common estimator is of the NB2 form, where th e variance is quadratic in the mean (Cameron and Trivedi, 1998, p. 63), given th at a considerable number of pred ictors in this study are timeinvariant or slowly changing over time, the estimator in equation (21) might provide a feasible 187

PAGE 188

alternative to NB2 in which time-invariant variab les are not identified. However, consistency of the NB1 estimator is ambiguous when heterogeneity in the mean of the distribution (and not in the variance) is present (G reene, 2007, section 4.1). 6.8 Data Description For the period 1998 to 2007 data were obtained for the fifteen member states (EU15) resulting in 150 observations. Tabl es 6-1 and 6-2 present the desc riptive summary statistics for the collected dataset for each member state that conform to the operationalization of hypotheses provided in Sections 6.5 and 6.6 of this chapter. The dependent variable (non-compliance) is moderately overdispersed becau se the sample variance of (over time and across countries) is over three times the sample mean of 5.63. This makes it very likely that default standard errors for both cross-section and pa nel Poisson estimators will understate the true standard errors. Therefore, it is imperative that robust st andard errors are obtained. 24.4820.07 Analyzing the results in Tabl e 6-2 most predictors of noncompliance exhibit significant overall variance across countries and over time. This is evident by the large standard deviation and range (defined as the maximum minus the mini mum value) of each vari able. In terms of the shape of the distribution of each variable, several variables can be approximated by a normal distribution that has a skewness of zero and a kurtosis of three.91 However, most variables appear to be mildly positively skewed since their measure of skewness is greater than zero. Of particular importance though is each predictors variance between-countries (e.g., across countries) and within countries (over time). Specifically, depend ing on whether a predictor is time-invariant or not will affect the choice of the panel estimators. 91 Usually a value of three is subtracted from the fourth moment of the normal distribution to result in a kurtosis of zero. This convention is not followed in this study. 188

PAGE 189

As expected the measure for bargaining power ( ) is almost time-invariant (Table 62) during the period of study with a common change across member s states in year 2001 that coincides with the change in the allocation of the votes in the Counc il of Ministers. Similarly, the measure for ENVI membership (CoMV E NVI) is slowly changing over ti me with a common change across member states in year 2004 that coincides with the accession of 10 more countries in the EU. An exception is Luxembourg where membership in ENVI did not change for the period of study. In addition, the measures for governmental capacity ( L IFE) and favorable industry climate ( E MAS) were time-invariant for Luxembourg, whic h is justified by the small size of the country. Further, several other variables are slow ly changing over time for most or all of the member states, such as the measures of contractual environment (1 tECJ ), governmental capacity ( L IFEIndEmpl), Green party membership (G), overall commitment ( ), industry clout ( ), need for regulation (reenEnvExp E mis sion), corruption ( ), and cozy regulatory efficiency (). Such enlargement and voting-system events that have common repercussions across member states, as well as changes in the legal re gime for enforcing and penalizing non-compliance, justify the use of time-specific fixed e ffects models in the empirical specification. CorrRegQ ual 189

PAGE 190

Table 6-1. Variab le definitions Symbol Hypothesis Definition CoMV H1 Council of Minister Votes (annual count) NGOfun H2 EU funding to Environm ental NGOs per state 1 tECJ H3 European Court of Justice case fillings at t-1 L IFE H4 EU project funding through LIFE+ E Uact H5 Standard Eurobarometer survey: should decisions be made by government, or jointly within the European Union Complaints H6 Complaints to the Commission for the environment sector E NVI H7 Ratio of ENVI members over the numb er of total MEPs for each state, Salience H8 Standard Eurobarometer survey: tw o of the most important issues Green H9 Number of seats over the total number of seats available in each state EnvExp H10 Government expenditure by function (Environment) E MAS H11 Number of EMAS organizations per population (in 1,000 persons) IndEmpl H12 Industry (including energy) employment as a percent of total employment E mission H13 Greenhouse gas emissions (UNFCCC) Compen H14 Total compensation of environment employees paid by the government Corr H15 Transparency Internationals Corruption Perception Index RegQual H16 World Governance Indicators: regulatory quality 190

PAGE 191

Table 6-2. Descriptive summary statisti cs of the predictors of non-compliance Symbol Mean Standard Deviation Min Max Skewness Kyrtosis TimeInvariant CountryInvariant CoMV 12.80 9.23 2.00 29.00 0.92 2.31 *1 NGOfun 45.93 30.52 7.00 157.00 1.19 4.11 1 tECJ 1.34 1.91 0.00 10.00 1.89 6.73 *2 L IFE 5.86 5.79 0.00 28.00 1.43 5.17 **3 **4 E Uact 60.72 10.63 37.00 87.70 0.07 2.68 Complaints75.98 58.13 4.00 289.00 1.19 3.97 E NVI 10.84 6.57 0.00 27.27 0.61 2.99 *5 Salience 6.12 5.24 0.00 29.71 1.71 6.89 Green 3.89 3.47 0.00 13.60 0.72 2.65 *6 EnvExp 1.47 0.59 0.26 3.27 0.53 3.35 E MAS 0.008 0.01 0.00 0.045 1.41 4.00 *7 IndEmpl 17.21 3.10 11.17 24.10 -0.03 2.18 *8 E mission 8.07 8.33 0.26 30.74 1.21 3.41 Compen 1.57 1.49 0.27 6.97 2.78 10.09 Corr 7.63 1.56 4.20 10.00 -0.54 2.41 RegQual 1.45 0.35 0.71 2.01 -0.35 1.90 Time-Invariant: almost time-invariant within each country, ** time-invariant. Country-Invariant: ** country-invariant. 1 Time-invariant: same number of votes from 1998 to 2000 and from 2001 to 2007. 2 Slowly changing for Denmark, Finland, Netherlands, and Sweden. 3 Time-invariant for Luxembourg. 4 Country-invariant for 2001. 5 Slowly changing for all countries; time-invariant for Luxembourg. 6 Time-invariant for Greece an d United Kingdom and slowly-changin g for all countries over time. 7 Time-invariant for Luxembourg and Ireland. 8 Slowly-changing over time for all countries and decreasing monotonically over time. 191

PAGE 192

To further analyze the time-invariance prope rties of the independe nt predictors, the variation of each predictor was decomposed into between-country and within-country variation and the results are presented in Table 6-3. Specifically, the be tween-country standard deviation for a specific variable was calculated as the st andard deviation of the means for each country. That is, the mean of a variable X for country i was first calculated ( iX) and then the standard deviation of the means was obtained. Within-country variation for a specific variable X was measured as the standard deviation of the devia tion from respective country s average. In other words, the deviation of each observation for country i from its over-time mean was first calculated ( iti X X ) and then the standard deviation of a ll resulting observations was obtained. For most of the predictors, presented in Ta ble 6-3, most of the variation is betweenvariation rather than within-variation. For instan ce, the within variation of the measure for need for regulation ( ) is 0.37 while its between-variation is 8.59. Therefore, it is expected that country-specific fixed-effects estimators will not be very efficien t because they rely on within variation and they might exacerbate problems of measurem ent error, especially for the predictors with small within variation (close to being time-invariant). Al so, the country-specific fixed effects parameter estimates may differ cons iderably from the other estimators (e.g., pooled, time-specific, etc.) depending on the differen ces of withinand between-variation. EmissionOne predictor that deserves special attention is the measure for industry clout ( ). In theory, we should expect th at as industry employment increases the effect on non-compliance should increase at an increasing ra te signifying the importance of th e industry sector in the total economy and therefore on policy formulation (trans position). However, this effect should reach a plateau wherein after industry employment reac hes a certain point, any additional industry employment (clout) and its influence on non-co mpliance should level off due to elevated IndEmpl 192

PAGE 193

sanctioning by the Commission. Co nversely, non-compliance reaches its maximum level at a certain corresponding level of industry clout. To test this theory a quadratic term of industry employment ( ) will be added in the em pirical specification. If the theory is valid then the coefficient of is expected to be positive, while the effect of the quadratic is expected to be negative. Ancillary evidence of this effect is provide d in Figure 6-5, which presents the industry employment along with nonc ompliance during the period of study. In particular, industry employment in almost every country falls during the period of this study and this event coincides w ith a decrease in non-compliance, especially in the period 2003-2007. IndEmplSq IndEmpl IndEmpl 0 10 20 30 0 10 20 30 0 10 20 30 0 10 20 30 1995 2000 2005 Austria Belgium Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal 2010 Spain Sweden United Kingdom 2000 2005 2010 1995 1995 2000 2005 2010 1995 2000 2005 2010 noncompliance Industry Employment year Figure 6-5. Trends in non-compliance and Industry Employment by country during 1998-2007. 193

PAGE 194

Table 6-3. Overall, between, and within va riation of the predictors of non-compliance Variable Variation Mean Standard Deviation Min Max Observations CoMV overall 12.80 9.23 2.00 29.00 150 NT between 7.66 3.40 23.30 15 N within 5.49 -0.50 18.50 10 T NGOfun overall 45.93 30.52 7.00 157.00 150 NT between 27.20 9.40 108.60 15 N within 15.36 4.63 99.63 10 T 1 tECJ overall 1.34 1.91 0.00 10.00 150 NT between 0.89 0.10 2.60 15 N within 1.71 -1.26 8.84 10 T L IFE overall 5.86 5.79 0.00 28.00 150 NT between 4.28 0.60 15.30 15 N within 4.04 -9.44 18.96 10 T E Uact overall 60.72 10.63 37.00 87.75 150 NT between 6.46 48.25 70.79 15 N within 8.59 40.81 86.91 10 T Complaintsoverall 75.98 58.13 4.00 289.00 150 NT between 53.70 8.00 172.90 15 N within 25.89 -26.92 197.58 10 T E NVI overall 10.85 6.57 0.00 27.27 150 NT between 6.19 0.00 24.09 15 N within 2.68 -5.22 16.21 10 T Salience overall 6.13 5.24 0.00 29.71 150 NT between 3.26 2.33 13.10 15 N within 4.18 -3.78 22.73 10 T Green overall 3.90 3.48 0.00 13.66 150 NT between 3.32 0.00 9.67 15 N within 1.32 -0.42 8.32 10 T EnvExp overall 1.47 0.59 0.26 3.28 150 NT between 0.59 0.54 2.82 15 N within 0.16 1.02 2.04 10 T 194

PAGE 195

Table 6-3. Continued Variable Variation Mean Standard Deviation Min Max Observations E MAS overall 0.01 0.01 0.00 0.05 150 NT between 0.01 0.00 0.03 15 N within 0.00 -0.01 0.02 10 T IndEmpl overall 17.21 3.10 11.18 24.11 150 NT between 2.96 12.52 22.22 15 N within 1.19 14.17 20.68 10 T E mission overall 8.08 8.34 0.27 30.74 150 NT between 8.59 0.34 29.32 15 N within 0.37 6.35 9.49 10 T Compen overall 1.57 1.49 0.28 6.97 150 NT between 1.53 0.31 6.77 15 N within 0.16 1.04 2.26 10 T Corr overall 7.63 1.56 4.20 10.00 150 NT between 1.58 4.50 9.70 15 N within 0.32 6.19 8.49 10 T RegQual overall 1.45 0.35 0.72 2.01 150 NT between 0.34 0.87 1.85 15 N within 0.10 1.11 1.65 10 T 6.9 Model Selection In this section, a comparison between the diffe rent estimators developed in the previous sections is presented to determine the estimator that best fits the data and explains noncompliance across member states of the EU. In particular, th e empirical specifications and results are first presented for the pooled and popul ation-averaged Poisson models which serve as the benchmark in the comparison across estimators. Consequently, results fr om the application of panel Poisson estimators, such as time-specific and country-specific fi xed effects and random effects models, are presented. Several statistical tests are then employed to select the best model for this study. Starting with the empirical specification of th e pooled and population-averaged 195

PAGE 196

Poisson model and following the analysis in the previous section, the conditional mean of the Poisson distribution in equation (6) is specified as: 1231 456 7891 01 11 2 13 14 15 16 17ln()it it it it it it it it it it it it it it it it itCoMVNGO f unECJLIFEEUactComplaints E NVISalienceGreenEnvExpEMASIndEmpl IndEmplSqEmissionCompenCorrRegQu ital (22) Maximum likelihood estimation of the pooled Poisson model in equation (22) is performed in STATA. To account for potentia l overdispersion and autocorrela tion, cluster-robust standard errors are obtained with clustering on the countri es. Further, to allow for a particular form of autocorrelation, the population-averaged model is estimated using the Generalized Estimating Equations (GEE) estimator. Specifically, it is assumed that the error terms in equation (22) follow an autoregressive process of order one, AR(1) since has a lag of one period. ECJTable 6-4 presents the results for both the pooled and population-averaged Poisson models. Both estimators provide similar results for the predictors with the magnitude and statistical significance of the estimated coefficients being similar between the two models. One difference is that the coefficient of I ndEmplSq becomes significant in the popul ation-averaged model. To assess the Poisson probability distribution assumption for the dataset a goodness of fit 2 test was performed for the pooled model. The hypothesis that the data are Poisson distributed is rejected at any significance level ( -value = 0.00). However, the Poisson estimator requires only that equation 2224.5 p(22) is correctly specified to be consistent, yet it does not require that the data are actually Poisson dist ributed. It is essential though that robust standard errors are obtained if the data are not Poisson distributed (Cameron and Trivedi, 2009, p. 620). 196

PAGE 197

Table 6-4. Estimates of Pooled and Panel Poisson regression models Variables Pooled PA w ith Fixed Effects Random Model AR(1) Time Country Effects Bargaining power 0.036*** 0.039*** 0.023** 0.061*** 0.042*** (Co) MV(0.01) (0.01) (0.01) (0.01) (0.01) Governmental concern -0.028*** -0.027*** -0.012*** -0.029*** -0.031*** ( ) NGOfun (0.00) (0.00) (0.00) (0.00) (0.00) Contractual environment 0.063** 0.057** 0.039** 0.047*** 0.065*** ( ) 1 tECJ(0.03) (0.02) (0.02) (0.01) (0.02) Governmental capacity 0.004 0.003 0.003 0.011** 0.001 ( L IFE) (0.01) (0.01) (0.00) (0.01) (0.01) Approval of EU action -0.025*** -0.022*** -0.018*** -0.020*** -0.025*** ( E Uact) (0.00) (0.00) (0.01) (0.00) (0.01) Citizen groups 0.0002 0.0002 -0.001 0.001 0.001 (Co) mplaints(0.00) (0.00) (0.00) (0.00) (0.00) ENVI membership -0.023* -0.018 -0.005 0.003 -0.006 ( E NVI) (0.01) (0.01) (0.01) (0.02) (0.02) Salience -0.053*** -0.045*** -0.039** -0.062*** -0.055*** ( ) Salience(0.02) (0.01) (0.02) (0.01) (0.02) Green party membership 0.010 0.006 -0.001 0.003 0.013 (Gr) een(0.03) (0.03) (0.01) (0.02) (0.06) Overall commitment -0.541*** -0.527*** -0.219* -0.034 -0.413 ( ) EnvExp(0.17) (0.17) (0.13) (0.25) (0.32) Favorable industry climate 23.17*** -23.72*** -10.24** -38.59** -30.55 ( E MAS) (7.56) (8.06) (4.74) (16.49) (19.65) Industry clout 0.191 0.275 0.215* 0.288 0.179 ( ) IndEmpl(0.20) (0.22) (0.13) (0.27) (0.44) Industry clout squared -0.009 -0.011* -0.009** -0.006 -0.007 ( ) IndEmplSq(0.01) (0.01) (0.00) (0.01) (0.01) Need for regulation 0.073*** 0.071*** 0.046*** 0.138 0.066 ( E mission) (0.01) (0.02) (0.01) (0.13) (0.05) Staffing levels 0.015 0.020 0.052 0.119 -0.011 (Com) pen(0.05) (0.06) (0.03) (0.40) (0.26) Corruption -0.182** -0.198** -0.204*** -0.247** -0.152 (Corr) (0.09) (0.08) (0.05) (0.12) (0.15) Cozy regulatory efficiency 1.340*** 1.319*** 0.745*** 1.021** 1.283* ( ) RegQual(0.40) (0.37) (0.29) (0.52) (0.73) 197

PAGE 198

Table 6-4. Continued Variables Pooled PA w ith Fixed Effects Random Model AR(1) Time Country Effects Year1998 -0.281 (0.29) Year1999 -0.538*** (0.15) Year2000 -0.462*** (0.16) Year2001 -0.236** (0.10) Year2003 -0.728*** (0.19) Year2004 -0.294* (0.16) Year2005 -0.665*** (0.15) Year2006 -1.813*** (0.26) Year2007 -1.726*** (0.24) Constant ( ) 3.692*** 2.695* 3.129*** 2.973 (1.44) (1.57) (1.02) (3.57) Log likelihood -360.9 -303.5 -284.8 -357.2 Information Criteria: AIC BIC 749.8 791.9 634.9 677.1 603.6 654.8 752.3 809.5 Goodness-of-fit 2 (p-value) 224.5 (0.00) 131.1 (0.60) 0.00 (0.99) 7.49 (0.003) PseudoR2 0.28 0.39 Notes: significant at 0.10 level, ** significant at 0.05 level, *** significant at 0.01 level. The coefficients presented are the pooled, populatio n-averaged with AR(1) er ror terms, time-specific fixed effects, country-specific fixe d effects, and random effects Po isson estimators, respectively. Standard errors for all models are in parentheses. For the random effects model, standard errors were obtained through bootstrapping with 100 replications. For the population-averaged model, errors were assumed to follow an AR(1) process and robust standard errors were obtained. For the rest of the models cluster-robust standard errors were obtained with clustering on the countries. Note that the coefficients are not interpreted as in a multiple regression model. 198

PAGE 199

To account for common developments across memb er states and for ch anges in the number of applicable environmental directives that a state needs to comply w ith in each year, year dummies are incorporated in equation (22) This results in a time-specific fixed effects model as presented in equation (11) (without the country-sp ecific dummies) and the results are presented in Table 6-4, where, as before, cluster-robust st andard errors with clus tering in the countries were obtained to account for potential overdispersi on and autocorrelation in the data. Comparing the time-specific fixed effects model with th e pooled and population-averaged models, it is observed that the magnitude of the estimated coe fficients in the time-specific fixed effects model is lower than the other two models. In terms of statistical signifi cance of the estimated coefficients, all three models obtain similar resu lts with the exception of the coefficients of I ndEmpl and I ndEmplSq that are both statistically signifi cant in the time-specific fixed effects model. This supports the theory of a quadrati c effect of industry clout on noncompliance. Contrary to the pooled Poisson model, the null hypothesis that th e data are Poisson distributed cannot be rejected at a 0.01 level of significance ( -value = 0.60) in the time-specific fixed effects model (Table 6-4). Furthe r, to test the joint significance of the year dummies in the time-specific fixed effects model a Wald 2131.1p2 test was performed. The null hypothesis in this test is that all coefficients of the year dummies are simultaneously equal to zero, thus they have no effect on noncompliance, and the alternative hypothes is is that at least one coefficient is statistically significantly different than zero. The results support the alternative hypothesis at any leve l of significance ( -value = 0.00). Therefore, the fixed effects model with time-specific effects provides a bett er fit than the pooled or the populationaveraged Poisson models. This is evident also from the Akaike Information Criterion (AIC) and 2(9)170.5p 199

PAGE 200

the Bayesian Information Criterion (BIC) that are defined as follows (Cameron and Trivedi, 1998, pg. 183) 2ln 2lnln() AICLk B ICLNk (23) where ln L denotes the fitted log-likelihood function in each model, the number of independent variables, and the number of countries. Since the log-likelihood will increase as independent variables are added to a model, these criteria penalize models with larger number of independent variables. From the results in Tabl e 6-4, the time-specific fixed effects model has a lower AIC even if it has a larg er number of independent variables than the pooled model and thus provides a better fit for the data. k NTo account for heterogeneity across countries th at is not captured by the predictors in equation (22) a country-specific fixed effects model was considered. Specifically, country dummies were incorporated in equation (22) as presented in equation (11) but the overall constant term was excluded since it is not identified. One could also append time dummies in the country-specific fixed effects model to estimate the two-way error component model of equation (11) but given the small number of observ ations across countries/over time (small and T) considered in this study, it would result in a large degrees of freedom loss. NResults for the country-specific fixed effects model, estimated using the conditional maximum likelihood approach (Cameron and Trivedi, 1998, pp. 282-284) with cluster-robust standard errors, are also presente d in Table 6-4. Estimates for the regression coefficients differ substantially from the estimates of the pooled an d especially of the time-specific fixed effects model. This difference is accentuated for the pred ictors that are time-invariant for a subset of member states in the sample. For instance, the measure of favorable industry climate ( ) is time-invariant for Luxembourg and Ireland (Table s 6-2 and 6-3). This time-invariance, for a EMAS 200

PAGE 201

subset of countries, results in the estimated coefficient of under the country-specific fixed effects model to substantially differ than the other models. Similar results are obtained for other predictors that are almost timeinvariant such as bargaining power (CoM) and need for regulation ( ). EMAS V Emission L IFEFurther, with the exception of the measure of governmental capacity ( ) that becomes statistically significant since its w ithin variation is almost the same as its between variation (Table 6-3), several variables that were statistically significant in the pooled and time-specific fixed effects models have become insignificant (e.g., ENVI E nv Exp, and the measures for industry employment). This result is explained by the fact that these variables are slowly changing over time (Table 6-2) and so their wit hin variation is small (Table 6-3). As such their effect is absorbed by the countyspecific effects (Beck, 2001). As Mbaye (2001) noted, since the cross-sectional va riables test theory, and if proper theory is examined, then cross-country variance is taken in to account by the cross-sectional variables. This is evident in this study by the Wald Emission2 for the country-specific fixed e ffects model that tests the null hypothesis that all regression coefficients are simultaneously equal to zero. From Table 6-4, it is observed that the 2 test statistic is approximately equa l to zero suggesting that the countryspecific fixed effects model does not provide a good fit for the data ( -value = 0.99) and thus most cross-sectional varia tion has been accounted for by the independent variables included in the model. 2(1 7)0 pFinally, in the last column of Table 6-4 the results for the random effects model as specified in equation (14) are presented, where standard errors for the estimated coefficients were obtained through bootstrapping with 100 replic ations. It should be me ntioned that a random effects model might not be appropriate for our da ta since no random sample was drawn to select 201

PAGE 202

the countries in this study, instead the fifteen member states represent the population for the period of study. The estimated coefficients from the random effects model are similar to the ones obtained from the pooled model but several coefficients have become insignificant. To test the random effects specification versus the pooled model a likelihood ratio test is employed, noting that the random effects specification in equation (14) becomes the pooled estimator if the parameter (1/ ) is equal to zero. Specifically, th e null and alternative hypothesis are: 01 : H0 pooled model does not diffe r than the random effects model :aH 0 random effects model is a better specification To perform this hypothesis test, the log likeli hood of each model is obtained to form the likelihood ratio: 2(lnln)randompooledLRLL (24) where L R stands for likelihood ratio and follows a 2 distribution with one degree of freedom since one parameter (1/ ) is examined, is the log-likelihood of the pooled model, and ln is the log-likelihood of the random effects model. From equation lnpooleL360.9) 2 0.05d7.(1) rL2andom(1) (24) and Table 6-4, we obtain that Therefore, the null hypothesis is rejected and thus the random ef fects model is a better specification than th e pooled model, since the observed is greater than the critical at a 0.05 level of significance (obtained from a chi-square table). This result was expected since the random effects model accounts for both country-specific and time-specifi c effects, and as sh own earlier the timespecific fixed effects model was a better specification than the pooled model. 2(1)2(357.2 4 LR 3.87.44 To compare the random effects model with the time-specific fixed effects model that provides the best fit for the data in this study (according to AIC), the ra ndom effects model was 202

PAGE 203

estimated with year dummies (results are presente d in Appendix D) and was then compared with the time-specific fixed effects model using a likelihood ratio test. Spec ifically, the likelihood ratio test was where from Appendix D and from Table 6-4 resulting in 2(lnln)randomtimeFELRLL ln303.47778timeFEL ln303.47777randomL 2(1)0 LR Therefore, the null hypothesis that the time-specific fi xed effects model is a better specification than the random effects cannot be rejected. In conclusion, sufficient evidence was obtaine d to support that the time-specific fixed effects Poisson estimator provides the best fit fo r the dataset of this study. To explicitly account for potential overdispersion in our dataset the pooled and panel negative binomial estimators, developed previously, were also obtained. The resu lts presented in Table 6-5 are similar to those obtained from the Poisson estimators. In partic ular, it is observed that all negative binomial estimators have similar estimated coefficients and significance with the corresponding Poisson estimators except for the estimators with country -specific fixed effects. For the latter, the estimated coefficients of the measures for favorable industry climate ( ), cozy regulatory efficiency (EMAS R e g qual), and corruption (Co) have become insignificant under the negative binomial specification even if, as mentioned in a previous section, the negative binomial estimator with country-specific fixed effects (as a pplied in this study) allo ws for the inclusion of slowly changing over time or ti me-invariant variables. This result might suggest a possible misspecification problem of the negative binomial estimator. According to Cameron and Trivedi (2005, Chapters 20.2 and 20.4) the negative binomi al estimator is less robust to distributional misspecification than the Poisson. The Poisson estimator is consistent under the weak assumption that the conditional mean is correctly specified, even if the data are not Poisson distributed. Instead, the negative binomial estimator will be inconsistent even if the conditional rr 203

PAGE 204

mean is correctly specified, except for the case of the NB2 model wher e the variance of the model is considered to be quadratic in the mean However, in this case, time-invariant variables cannot be included in the mode l and slowly-changing over time independent variables might become insignificant. Further, obtaining cluster-robus t standard errors for the Po isson estimators corrects for potential problems of overdi spersion and autocorrelati on. To examine whether the overdispersion problem still exists in the time-sp ecific fixed effects Poiss on model that exhibited the best performance among the Poisson estimators a likelihood ratio is app lied. Specifically, the following null and alternativ e hypotheses are examined: 0:()()it it H EyVary which means the negative binomial model reduces to the Poisson :()()ait it H EyVary which implies overdispersion To perform this hypothesis test, the log-lik elihood of the time-specific fixed effects specification under the Poisson and negative binomi al models is obtained to form the likelihood ratio: 2(lnln) P oissonnegbinLRLL (25) where L R stands for likelihood ratio and follows a 2 distribution with one degree of freedom since there is only one constraint, is the log-likelihood of the negative binomial time-specific fixed-effects model, and is the corresponding l og-likelihood of the Poisson model. From Tables 6-4 and 6-5, we obtain that Since the observed is lower than the critical at a 0.05 level of significance, the null hypothesis cannot be rejected; thereby the Poisson estimator is more suitable for the dataset cons idered in this study. lnnegbinLPoissonln L2(1)2(303.5304.7)2.4(1)3.84 LR2 0.052(1)2.4 204

PAGE 205

Table 6-5. Estimates of Pooled and Pa nel Negative Binomial regression models Variables Pooled PA w ith Fixed Effects Random Model AR(1) Time Country Effects Bargaining power 0.032*** 0.031*** 0.024** 0.056*** 0.036*** (Co) MV(0.01) (0.01) (0.01) (0.01) (0.01) Governmental concern -0.027*** -0.027*** -0.014*** -0.031*** -0.029*** ( ) NGOfun(0.00) (0.00) (0.00) (0.01) (0.00) Contractual environment 0.075*** 0.087*** 0.043** 0.055*** 0.065*** ( ) 1 tECJ(0.02) (0.02) (0.02) (0.02) (0.03) Governmental capacity 0.002 -0.001 0.006 0.009 0.003 ( L IFE) (0.01) (0.01) (0.00) (0.01) (0.01) Approval of EU action -0.027*** -0.027*** -0.018*** -0.021*** -0.025*** ( E Uact) (0.00) (0.01) (0.01) (0.01) (0.01) Citizen groups 0.001 0.001 -0.001 0.001 0.001 (Co) mplaints(0.00) (0.00) (0.00) (0.00) (0.00) ENVI membership -0.021 -0.014 -0.006 0.000 -0.017 ( E NVI) (0.01) (0.01) (0.01) (0.03) (0.02) Salience -0.051*** -0.045*** -0.028** -0.056*** -0.055*** ( ) Salience(0.02) (0.02) (0.01) (0.02) (0.02) Green party membership 0.012 0.008 0.002 0.005 0.012 (Gr) een(0.03) (0.03) (0.01) (0.04) (0.06) Overall commitment -0.552*** -0.556*** -0.240* 0.054 -0.499 ( ) EnvExp(0.20) (0.20) (0.13) (0.65) (0.38) Favorable industry climate 22.05*** -21.71*** -11.69** -37.16 -25.65 ( E MAS) (7.84) (8.22) (4.64) (40.64) (16.52) Industry clout 0.161 0.146 0.216* 0.401 0.148 ( ) IndEmpl(0.21) (0.23) (0.12) (0.40) (0.34) Industry clout squared -0.008 -0.008 -0.009** -0.010 -0.008 ( ) IndEmplSq(0.01) (0.01) (0.00) (0.01) (0.01) Need for regulation 0.071*** 0.068*** 0.050*** -0.066 0.070*** ( E mission) (0.01) (0.02) (0.01) (0.23) (0.03) Staffing levels 0.031 0.059 0.046 0.010 0.009 (Com) pen(0.05) (0.06) (0.03) (1.07) (0.23) Corruption -0.174** -0.165* -0.216*** -0.217 -0.135 (Corr) (0.09) (0.09) (0.06) (0.17) (0.14) Cozy regulatory efficiency 1.272*** 1.161*** 0.842*** 1.113 1.213* ( ) RegQual(0.44) (0.42) (0.31) (0.91) (0.71) 205

PAGE 206

Table 6-5. Continued Variables Pooled PA w ith Fixed Effects Random Model AR(1) Time Country Effects Year1998 -0.269 (0.30) Year1999 -0.484*** (0.16) Year2000 -0.364** (0.17) Year2003 -0.599*** (0.16) Year2004 -0.188 (0.15) Year2005 -0.536*** (0.15) Year2006 -1.672*** (0.26) Year2007 -1.639*** (0.25) Constant ( ) 3.947*** 3.910*** 3.095*** 2.563 4.530 (1.48) (1.47) (1.00) (4.27) (2.99) Log likelihood -353.9 -304.7 -283.4 -353.3 Information Criteria: AIC BIC 735.9 778.2 639.4 684.6 602.8 656.9 746.6 806.9 Notes: significant at 0.10 level, ** significant at 0.05 level, *** significant at 0.01 level. The coefficients presented are the pooled, populatio n-averaged with AR(1) er ror terms, time-specific fixed effects, country-specific fixed effects, and random effects negative binomial estimators, respectively. Standard errors for all models are in parentheses. For the random effects and fixed effects with countryspecific effects models, standard errors were obtain ed through bootstrapping with 100 replications. For the population-averaged model, errors were assume d to follow an AR(1) process and robust standard errors were obtained. For the rest of the models cl uster-robust standard errors were obtained with clustering on the countries. Note that the coefficients are not interpreted as in a multiple regression model. 6.10 Discussion and Conclusions As per the above discussion, the time-specific fixed effects Poisson estimator provides the best fit for the dataset of this study (the results are reported in Table 6-4 above). With regard to 206

PAGE 207

the hypotheses offered in this study, the results ar e mostly consistent with expectations. Starting off with variables under the rubric of international relations, th e model begins with bargaining power. States whose bargaining pow er in the Council of Ministers is high actually commit more non-compliance. This result is consistent with previous findings (Mbaye, 2001), yet the hypothesized relationship under previ ous studies and the goodness of fit hypothesis, that states comply better with European demands because they use their influence to craft policies closer to their preferences, is empirically disconfirmed in this study. It s eems more likely, that given the qualified majority rules in the Council of Ministers, power is used to shirk compliance. Under qualified majority voting, member states will not be able to veto a decision that goes against their preferences. For this reason, me mber states outvoted in the Counc il of Ministers will have an incentive to delay implementation, which of cour se is in line with the argument that actors behave strategically. This suggestion goes a long way in settling the goodness of fit debate by pointing out that fitness does not matter in determining compliance as much as the bargaining power needed to attain it or the power to avoid it (as is show n in this study). The second variable of the model pertains to the ability of established supranational institutions (the Commission) to affect governmental concer n by providing information and resources to specific domestic actors that will in turn press for action. This study finds a statistically significant nega tive relationship between Commi ssion funding for Environmental Non-Governmental Organizations (ENGOs) and compliance, indica ting that increased funding to domestic NGOs is more likely to be associated with increased levels of compliance. This finding is in line with Neoliberal institutionalisms (management approach) suggestions on the ability of institutions to enhance governmental concern by enhancing the power of sub-national actors. As expected, this finding gives suppor t to the ability of management mechanisms in influencing 207

PAGE 208

compliance. Of course, this finding also lends suppor t to the EU studies lite rature on the effect of national actors in obstructing or facilitating compliance (Duina, 1997; Haverland, 2000; Mbaye, 2001; Giuliani, 2003; Kaeding, 2006; Perkins an d Neumayer, 2007; Konig and Luetgert, 2008). However, the effect of the principal (EU through the Commission) in enhancing compliance by tending to the needs of th ese domestic forces or recognizing thei r effect and using it to ensure compliance is largely neglected in the literature. As this finding suggests, regimes can increase governmental concern, and thus compliance, by doing exactly that (using domestic players). The ability of the EU to offer a reason able contractual environment by monitoring compliance and assuring states that cheating will be detected and punished, is found to be statistically significant, but stands contrary to the hypothesized relationship. In short, the effect of European Court of Justice (ECJ) cases agai nst member states is found to actually increase non-compliance in the next year. This odd impact may very well be due to the ope rationalization of this variable, as the filling date rather than the judgment date of a case was selected. This was done or two reasons. First, the fil ling date is closer to the conditions that created non-compliance, and owing to the backlogs of the ECJ, it would not have been prude nt to select judgment dates as those may come several years down the road and would thus influence the time-specific character of this study. Second, f illing dates are expected to have a relatively bigger impact that judgment dates on non-compliance because of the reput ation effect and the de finitive threat of sanctions for directives states did not transpose re cently, rather than the eff ect of a judgment that pertains to non-complied directives from severa l years earlier. For th ese two reasons it seems more likely that the indicator was operationalized correctly, rath er than that the hypothesized relationship is correct. It seems more sensible to argue that, the filling of a case with the ECJ will affect next years transposition record negatively because states are not af raid that the case will 208

PAGE 209

be judged any time soon (owing to ECJ backlogs), or because they hope to resolve the case before it truly reaches the Court, or finally, becau se they are simply not afraid of the Commission sticks. It may also be that (some) states ar e not averted to being named and shamed by the Commission. Whatever the reason, this resu lt shows that the Commissions enforcement mechanisms are ineffective in decreasing future non-transposition. To be precise, states do comply more when named and shamed (as s hown by the decrease of non-compliance as the Commission moves from Reasoned Op inions to ECJ Referrals), but this does not mean that states learned their lesson and will play nice next year. Thus, two conclusions should be kept in mind when analyzing the EU c ontractual environment. First, that it is not reasonable enough to avert future non-compliance, and second that sticks dont really work as a compliance instrument for next years compliance (in fact th ey increase non-compliance in the next year). With regard to the ability of the EU to enhance governmental capacity and thus compliance, this study finds an insignifi cant positive relationship between EU-funded environment projects and non-compliance. It could be said that the measure of the EUs ability to enhance governmental capacity is crude, but as argued elsewhere in this study the less environmental problems a state has to deal with the more resources it can dedicate to other problems (if of course this is what it wants). Nevertheless, this finding fails to provide support for the ability of the EU to enhance governmental capacity. It may very well be that this management mechanism is inadequate, but it ma y also be that EU f unding is enough to enhance the ability of environmental expenditure by the state (addressed with H10 on overall commitment below). The final hypothesis emanating fr om international relations theory, on public approval of EU action on environmental policy is found to have a statistically significant negative 209

PAGE 210

relationship with non-compliance. Constructivist explanations ha ve largely been ignored in previous quantitative studies, this study finds however, that construc tivism can add to our understanding of what drives non-c ompliance. States with public s more favorable to EU action on environmental policy are found to have an eas ier task in implementing EU policy since the public normatively accepts EU policy as legitimate. However, it is the publics approval of EU action on environmental policy that this study measures and not the sentiments of the public toward the EU in general (Lampinen and Uusikyla, 1998 ; Mbaye, 2001; Kaeding, 2006; Perkins and Neumayer, 2007). Hence, this study offers a novel and considerable c ontribution to theory and empirical investigation as it is not wise to assume that states with favorable EU sentiments will invariably comply with all EU activity. The normative acceptance of EU action as legitimate is mediated by the existence of policy entrepreneurs, which ma y not always be effective in creating and disseminating normative understandings in other sectors of policy. Thus, even though, normative acceptance may work for the envi ronment to increase compliance, it may not work in other sectors of policy, providing us with a possible explanat ion of non-compliance in other sectors. Moving to regulatory federalism explanations of non-compliance, this study finds that states with more active envir onmental groups will have more compliance, but the result is statistically insignificant. This may be due to the fact that complaints (the indicator used) are an inadequate measure of the existence of these groups but a different explanation is more plausible. As argued earlier in this chapter, complaints co me later in the implementation process (when the directive is transposed and bei ng applied or enforced). Complaints are also a heavily biased indicator of non-compliance when the dependent variable is infringement proceedings. This is 210

PAGE 211

because of the way complaints are reported by the Commission (as a source of detection).92 Given that they are the major way the Commission detects bad application and non-conformity they unusually influence Commi ssion follow-up on infringements, and a such infringements are an unbalanced way to measure non -compliance. This result may provide support to the use of transposition rates to de signate non-compliance rather than in fringements, but it fails to support top-down (enforcement) theories of non-compliance. The second top-down infl uence on non-compliance coming from the regulatory federalism literature, also fails to attain statisti cal significance. This study expected, in line with the EU studies literature on the goodness of fit, that membership in the European Parliaments Environment Committee (ENVI) would enhance compliance through stricter scrutiny of Commission action (using petitions and questions as a source of non-compliance detection), but also through the use of their clout in ENVI to attain more favora ble results from the legislative process. This did not prove to be the case in this study. It is true, however that the existence of differential decision making proced ures (like qualified majority voting), might serve to decrease the usefulness of membership in the EP Environment Committ ee as a tool to control the Commissions product (directives), or as a mechanisms to control how close those products will be to the desired ones. This finding is in line with the above finding (Counc il of Ministers) on the inability of states to affect the regulatory out comes of EU decision-making using its institutions and complying due the goodness of fit they pursued and attained. As such, it fails to provide support for the goodness of fit assumption, as well as for the ability of top-down mechanisms to induce compliance (a result consistent with the arguments of this study). Moving, then, to the bottom-up influen ces on compliance, this study finds a highly significant negative relationship be tween the salience of the envir onmental issue and the level of 92 For more on this, please refer to section 6.3 of this chapter. 211

PAGE 212

non-compliance. As hypothesized, salience does ha ve an adverse effect on non-compliance, more so than the EU studies literature would suggest. There, the context of the policy becomes important only when there is moderate pressu re on administrative styles from European requirements (Knill and Lenschow, 1998). This study finds salience to be more than a mediating factor or a secondary casual factor to the cons traints created by the in stitutional framework in which the policy will take place. According to my argument, in cases where the environment is elevated to a crisis status by interested political forces (public and private) then agencies will be pressured to treat the environmental issue at ha nd as a crisis and thus with an added degree of professionalism, hence increasing compliance. Beyond the added benefit of being the first quantitative study to statistically verify the im portance of salience in determining compliance, this finding lends support to th e bottom-up camp of regulatory federalism, while it also provides evidence against the goodness of fit ar gument and the purporte d negative effect of powerful actors (industry) on environmenta l compliance (more on this below). The existence of green parties in national legislatures has a statistically insignificant effect on non-compliance (albeit a negative one). This finding, also, confirms the power of salience as a standalone variable (among the bottom-up ones) in explaining compliance under the mechanism offered in this study. Salience, does seem to need change agents (Versluis, 2003) using their powers or their institutionalized relationships with the political principals of bureaucracies to pressure for change and compliance. It seems that the sali ence of an issue with the public is enough to attr act the attention of politicians with out the mediation of other actors, who in turn listen and pressu re themselves and the admini strative agencies involved in transposition to comply. 212

PAGE 213

As hypothesized, the overall commitment of states to environmental regulation has a negative effect on non-compliance. A states commitment to environmental regulation (presumably) has a positive effect on compliance as investment in implementation and pollution abatement rises and reflects this commitment. Additionally, environmental expenditures not only show the level of commitment, but also the capacity of the sate to implement such legislation. As such this finding could be shown as provi ding support to the EU studies literature on administrative constraints to compliance, and to international relations theories on involuntary non-compliance (Chayes and Chayes, 1993). However, a states conscious e ffort to decrease the impact of environmental problem s by increasing environmental i nvestment, and thus compliance with EU requirements, provides more direct support to the rationali st (Fearon, 1998) and bottom-up camps. But any increase in compliance with environmental directives from states will have a negative effect on industry. It is, thus, also expected that the industry will push for less legislation or for more non-compliance with EU directives. As hypothesi zed, industry clout does have a negative effect on compliance. In theory, it is expected as indus try employment increases the effect on non-compliance should increase at an increasing rate signifying the importance of the industry sector in the total economy and therefore on policy formulation (transposition). However, this purported effect should reach a plateau wherein after industry employment reaches a certain point, any a dditional industry employment (c lout) and its influence on noncompliance should level off due to elevated sanctioning by the Commi ssion. Conversely, noncompliance reaches its maximum level at a certa in corresponding level of industry clout. The findings in the statistical model are consistent with this effect I find that industry clout has a 213

PAGE 214

positive effect on non-compliance (as per the hypothesi zed relationship), but this effect levels off and becomes negative with high er levels of industry clout. On a related front, states with a greater need for regulation actually comply less with environmental directives. It is true that if the environmental problem is big, then the resources need to alleviate it will also be considerable (Z ito, 2000). This makes it necessary for states with a big pollution issue to invest more resources than other states, but it does not necessarily mean that states with a greater need for regulation (more pollution) will automatically invest more resources. This is because heavily polluted st ates will have a corresponding large industrial sector that will mobilize against environmental regulation and investment. Alternatively, given the responsiveness of regulators to various political forces, it is not surprising that the rate of compliance does not reflect the objective need for regulation. Taken togeth er, the significance of these findings on industry clout a nd need for regulation, suggest that the presen ce of powerful interests will affect compliance with environmental directives. Th is lends support to both the EU studies literature on domestic adjustment cost s and the roles of domestic actors, and the voluntary camp of international relations. Mo re importantly, they provide support to the bottom-up influences on compliance as suggested by the literature on regulatory federalism. However, even if the negative role of domestic actors on compliance is statistically verified here and in other studies, the assumption that this role will always be negative must be problematized. Manufacturers, or in dustry at large, may not always be opposed to environmental policies (Wurzel, 2002). Industry may actually be favorable to environmental legislation especially in green countries, and push for Eu ropeanization of state en vironmental regulations to level the playing field with competition in ot her states. This study finds that when the business climate is favorable to environmental legislati on then compliance with environmental directives 214

PAGE 215

increases. This statistically significant rela tionship has never before been empirically or theoretically addressed in the literature, as such the implications to theory are considerable. Most veto points (institutional structures that afford players with the ability to modify and block legislation) or domestic politics understandings of compliance seem to suggest an invariably negative relationship between dom estic actors and compliance as the level and amount of actor involvement will make the process slower. Especially on business opposition (Underdal, 1998; Grant et al., 2000; Weinthal a nd Parag, 2003; Beach, 2005; Perkins and Neumayer, 2007) the effect is almost always expected to be negative. This study suggests that we s hould not always assume a powerful (clout) actors stance as objectively true. In this case, even the industry is shown to have a positive effect on compliance with environmental directives. Even if their stance is assumed to hurt their own interests, though, industry favorable to e nvironmental legislation will push for Europeanization of directives and more enforcement by the Commission. Another route actors may use to block legislation is by directly influencing the states capacity to implement legislation using administ rative agencies. Understaffed agencies will have a harder time participating in the process of tran sposition, and be even mo re susceptible to other players in the process. However, this study does not find a statistically significant relationship between the staffing levels of ad ministrative agencies charged on the environment and levels of compliance. Connected to administrative capacity this finding suggests that the capacity of administrations is not an important factor in determining compliance. When coupled with the above finding on the significance of government al commitment, this finding goes against management theories of non-compliance, as it shows that compliance has little to do with administrative capacity and more to do with whet her states choose to ch annel that capacity to environmental legislation or not. 215

PAGE 216

This conclusion ties perfectly with another hypothesized relationshi p in this study. The existence of a cozy regulatory efficiency focu s is highly correlated wi th non-compliance. The more states are focused on pr ivate sector development the less is they comply with environmental directives. The capacity of a state to comply with environmental directives, and bureaucratic efficiency in general, has little to do with actual compliance. It is in fact the ability of powerful interests, in focusing the efficiency of the government toward their own interest that hurts compliance. The EU studies literature has addressed this coziness under such mediating variables as veto points (Haver land, 2000), corporatism (Lampine n and Uusikyla, 1998; Mbaye, 2001; Kaeding, 2006), and the level of partisan conflict and polarization within domestic veto players (Treib, 2003; Kaeding, 2006; Konig and Luetgert, 2008). The mere presence of actors with political clout or th e other mediating variables aforementioned will not be sufficient in blocki ng or allowing compliance. It is actually the entrance of these actors into a cozy relationshi p with the government that affects compliance, while efficiency has little to do with it. Administ rations are efficient in doing what they do, and states do have the capacity to implement legislation. Regardless of the existence of veto points, powerful actors, and the degree of fitness betwee n European requirements, it can be argued that some states are simply focused on doing different things (that se rve their preferences), and are, indeed, very efficient in doing t hose. In extension to this, corruption is also found to play a statistically significant positive role on non-complianc e. It is true that w ithout the existence of corruption, interests could never enter the cozy relationship men tioned above and focus bureaucratic efficiency to serve their goals. Finally, what can be said about the relative e xplanatory power of the model? Given that the time-specific fixed effects Poisson estimation does not lend itself to a measure of fit such as 216

PAGE 217

217 adjusted R-squared,93 statistical information criteria such as the Akaike information criterion (AIC) and the Bayesian informa tion criterion (BIC) can be empl oyed. These criteria assess the goodness of fit by assessing the explanatory power of non-linear models with reference to their log-likelihood94(Cameron and Trivedi, 2005). Given the fact that lower AIC and BIC values are preferred, the full model offered here presents a c onsiderably better fit than previous studies that employ the same techniques,95 however other studies have used infringements as the dependent variable, as such the comparabil ity of the scores does not provide for a direct comparison of full model strength. One can instead try to assess w ithin model explanatory power, and in this respect regulatory fede ralism offers a stronger model than international relations, while topdown explanations are weaker than bottom-up ones, and vol untary explanations are stronger than involuntary ones. However, even if it wa s theoretically and met hodologically possible to differentiate between these models the full model offered here offers the best explanation of noncompliance.96 It is an exercise in theoretical futi lity to separate the effect of top-down mechanisms from voluntary and international rela tions ones, or from bottom-up mechanisms for that matter. The full picture of co mpliance would be lost in theoreti cal differentiation. It is more likely that when images are taken together they provide the full image. As such, all theories add to the model and should not be c onsidered in disjunctive terms. 93 Even though the technique does offer a PseudoR2 of 0.39 for the full model. Compared to the .16 of Mbaye (2001). 94 For more information refers to section 6.9 of this chapter. 95 Perkins and Neumayers (2007) full model has an AIC of 1176.0 and a BIC of 1304.4, as compared to the AIC of 634.9 and a BIC of 677.1 in this study. 96 The international relations model has an AIC of 710.8 and BIC of 752.9 while the Regulatory federalism one has 658.2 and 700.3. The top-down model has an AIC of 707.6 and BIC of 749.8 while the bottom-up has 647.9 and 690.0. The voluntary model has an AIC of 647.8 and BIC of 690.1 while the involuntary has 716.8 and 759.02. In all cases, the full model has a lower AIC value that the individual ones, as such it is preferred.

PAGE 218

CHAPTER 7 CONCLUSION 7.1 Introduction The past three decades have witnessed a rapid proliferation of studies seeking to explain the implementation deficits of European Union (E U) policy. The main contributions have come from the disciplines of public administration, in ternational relations, and comparative politics. However, the disciplinary lines we have drawn have not proved helpful in understanding and explaining the sources of non-compliance with a high degree of empirical confirmation. Ridden with problems of ambiguous and co nflicting results theorists have tried to complement, fix, or expand on previously empirically disconfirmed hypotheses about the source of non-compliance. To be fair, there is a degree of overlap between disciplines, and a consid erable diversity within schools of thought (Checkel, 2001; Raustalia and Slaughter, 2002), but the theoretical murkiness created by several attempts to attain empirical reli ability has caused theorists to still be muddling through conflicting and divergen t explanations of non-complia nce that add little to our understanding and to theore tical cumulativeness. 7.2 The Unnecessary Simplificati on of the Goodness of Fit The most notorious such expl anation has been the goodness of fit hypothesis. Formulated by Heritier (1995), it basically ar gued that EU Member States attempt to export their policymaking attributes to the EU. This of course impl ied that Member States will try to protect their administrative and legal traditions by shirking th e implementation of EU legislation. Member States are seen as guardians of the status quo, as the shield protecting national legaladministrative traditions (Duina, 1997, p. 157). As a consequence governments who failed to upload their own policies to th e EU level would try to resist during the downloading process, when the agreed-upon measures were to be implemented (Borzel, 2002). 218

PAGE 219

However, empirical investigation of the hypot hesis has been less en couraging than its promise for theoretical un-ambiguity and parsim ony. Various empirically oriented researchers, focusing primarily on environmental policy, inve stigated the hypothesi s empirical grounds; unfortunately the results were ra ther dispiriting. Knill and Lenchow (1998), in an analysis of compliance focusing exclusively on the goodness of fit on four environmental directives in Germany and the UK found that th eir hypothesis was validated in three of the eight cases at hand. Similarly, Haverland ( 2000) analyzing the implementa tion of the Packaging Waste directive in Germany, the Netherlands, and the UK found that the country with the greatest misfit (the UK) adopted more successfully than the co untry that needed only incremental adjustments (Germany), while the latters record was even worse than the Dutch, despite the higher adaptation pressure for them (more incompa tibility). The final blow on the hypothesis was delivered by Falkner et al. (2005), which reports a failure of the misf it hypothesis as only 22% of their cases were in li ne with the hypothesis. In this case, this study makes an important cont ribution. I find that states with increased levels of bargaining power in the Council of Mini sters, actually comply less with environmental directives. As the argument goes, when states ha ve the power to bargain an agreement close to their interests and situations, then those states will have no problem implementing that agreement as it is closer to their self-i nterest (Fearon, 1998). So, supposedly, bargaining power allows states to preempt their goodness of fit by European izing already fitting policies. Conversely, powerless states may have a hard time bargaini ng agreements arrangement close to their own interests and, as such, will not comply as well or as often. The results of this study strike a blow to this relationship. Not only is goodness of fit disconfir med by the fact that more powerful states actually have a worse implementation record (contr ary to the expectation that their power would 219

PAGE 220

bring goodness of fit), but also fitness does not seem to matter in determining compliance as much as the bargaining power needed to attain it or the power to avoid it. This result if further strengthened by another finding of this study. In lin e with the EU studies literature on the goodne ss of fit, membership in the European Parliaments Environment Committee (ENVI) would enhance compliance through stricter scrutiny of Commission action (using petitions and questions as a source of non-compliance detection), but also through the use of state clout in the Comm ittee to attain more favorable results from the legislative process. This also di d not prove to be the case in this study. It seems more likely that, given the qualified majority rule s in the Council of Ministers a nd European Parliament, member states will not be able to veto decisions th at go against their prefer ences. For this reason, powerful member states outvoted in the Council of Ministers and European Parliament are better positioned to defy legislation given their relative political and economic weight in the EU system (Sverdrup, 2004). Both findings on the inability of states to affect the regulatory outcomes of EU decision-making by using its institutions and complying due the goodness of fit they pursued and supposedly attained, challenge th e goodness of fit argument. In th e end, it seems non-compliance has more to do with the power of th e state to defy legislation rather than the ability to affect the fit of EU policies with domestic traditions. To make matters theoretically worse, seve ral researchers decided to complement the original hypothesis with auxiliary ones instead of scrapping it all-together. For instance, Knill and Lenschow (1998) complement the original fra mework by adding three more variables. They include the degree of embeddedness that national in stitutions experience, while the explanatory value of the policy context become s important only in situations where the level of institutional embeddedness suggests the more ambiguous picture of moderate adaptation pressure, and it is 220

PAGE 221

with this picture that such notions as polic y salience become important (Knill and Lenschow, 1998: 611). Heritier et al. (2001) also revise the original hypothesis, this time adding to the impact of pre-existing policies the dynamics of the political process. Such elements as, the stage of liberalization, the national reform capacity, a nd the dominant belief system or problem-solving approach, are added as the determ inant of fit. However, they k eep goodness of fit as a necessary (downgraded from sufficient) condition, which nonetheless does not help explain why some states chose to comply and some didnt withou t the existence of those auxiliary variables. A similar strategy if followed by Borzel (2003), with her push-pull model incorporates policy misfit (as an initial reason for non-compliance), mobilization of domestic actors pressuring for implementation, and pressure from above where the Commission may initiate infringement proceedings, as the venues for effective implementation. However, not all cases where goodness of fit was a problem led to implementation problems. This study makes another important contribution in this case. The sali ence of an issue in domestic politics has largely been neglected in the EU studies literature, with a few notable exceptions (Knill and Lenschow, 1998; Verslu is, 2003). Knill and Lenschow (1998), treat salience as a secondary casual f actor to the constraints created by the institutional framework in which the policy will take place. In short, it is argued that policy salience will be an insufficient explanation of what drives compliance, as th e institutional framework (goodness of fit between European requirements and regulatory styles) will define whether salience will be effective or not. This study finds that issue salience, is more than a media ting factor. In cases where the environment is elevated to a crisis status by inter ested political forces (public and private) then 221

PAGE 222

agencies will be pressured to tr eat the environmental issue at hand as a crisis and thus with an added degree of professionalism, hence increasing compliance. When taken together, two more findings in this study seem to paint the same picture about the importance of what aforementioned studies treat as mediating factors to the goodness of fit. The existence of green parties in national legisl atures has a statistically insignificant effect on non-compliance (albeit a negative one). Additionally, the existence of environmental groups using the Commissions complaint mechanisms to induce compliance is also found to be insignificant. These findings, confirm the power of salience, even more, as a standalone variable in explaining compliance under the mechanism offe red in this study. Sali ence does seem to need change agents (Versluis, 2003) using their power s or their institutionali zed relationships with the political principals of bureaucracies to pressure for change and compliance. It seems that the salience of an issue with the public is enough to attract the attention of politicians without the mediation of other players, who in turn listen and pressure themselves and the administrative agencies involved in tr ansposition to comply. In this sense it becomes evident that we need not incorporate goodness of fit as a necessary (let alone sufficient condition) for im plementation. If domestic preferences can be shown to overcome fitness then why keep the original hypothesis in theoretical models, even as a necessary condition? One reason for this may be that it is theoretica lly rigorous while also being focused on empirical confirmation, howev er, Knill and Lenchow ( 1998), found that their hypothesis was validated in three of the eight cas es at hand. Another reason may be that the effects of the goodness of fit cannot be adequate ly captured by other explanatory variables in quantitative studies. This study argues, similarly to Mastenbroe k and Kaeding (2006), that the purported casual relationship be tween fitness and implementati on is spurious, and offers a 222

PAGE 223

powerful explanation of non-compliance (salience) as a factor that does not need fitness and is not mediating to fitness. The causa l effect is then in the misfit (o r fit) with underlying beliefs or preferences in domestic politics rather than the fitness of directives with institutional characteristics at home. 7.3 Domestic Politics and the Return of Administrative Explanations Theorists in the third wave of EU implemen tation research have moved away from the goodness of fit and focused more on domestic polit ics variables. Starting with Haverland (2000) variables such as veto points have been o ffered, arguing that govern ment must satisfy many coalition partners and other actors who shape both the quality and speed of implementation (regardless of goodness of f it). After being almost universally regarded as invalid (Ferner and Hyman, 1998) in the early 1990s, corporatism is also back in the limelight (Lampinen and Uusikyla, 1998; Mbaye, 2001; Kaeding, 2006) As the argument goes, a high level of corporatism will have an adverse effect on the leve l of veto players and thus result in increased compliance (Lampinen and Uusikyla, 1998). A close and cooperative arrangement between the state and interest groups will increase complian ce while an increased interest group involvement (level of pluralism) will lead to non-compliance (Konig and Luetgert, 2008). Finally, the level of partisan conflict and polarization within domes tic veto players (Treib, 2003; Kaeding, 2006; Konig and Luetgert, 2008) are al so hypothesized to affect comp liance at the domestic level. However, empirical investigation of the effect of veto points remains in its infancy, while also riddled with empirical ambiguity and mixed results. Some find the existence of veto players a good explanation of non-compliance (Kaeding 2006; Perkins and Neumayer 2007) while others find an insignificant negative relationship be tween veto players and non-compliance (i.e. the 223

PAGE 224

existence of veto players incr eases compliance, Mbaye, 2001).97 Corporatism is found to be insignificant while also shown to have a negative effect on compliance (contrary to the hypothesized relationship, Mbaye, 2001), or with a positive effect on compliance (Kaeding, 2006).98 Pluralism and partisan conflict is show n to increase non-compliance (Konig and Luetgert 2008), while Kaeding (2006) finds a positive yet insignificant relationship. Regardless of their empirical rigor, the theoretical insights of the third wave quantitative studies have been rather inconclu sive (as was the case with most misfits). This back and forth in empirical confirmation might ha ve to do with comparing differe nt fields of policy or with indicator strength, but this study offers a different explanation. Even though the importance of domestic politics has been rec ognized, we have missed the importance of several presumably mediating variables by focusing on standalone hypot hesis as the goodness of fit. While, even if the focus is taken away from the goodness of f it we still seem to miss the exact way those domestic preferences, actor interests, and the in terplay between them and the political principals of bureaucracies and the supranati onal principals (EU) of the st ate implementation effort will affect compliance. In a way, in their attempt to gain statis tical significance and theo retical rigor, even domestic politics explanations seem to have forgotten the drama of politics and the intergovernmental setting in whic h regulation takes place in the European Union. In a federal system bureaucratic agents are pulled in multiple directions by a multitude of political principals, who in turn experience their own pressures and competing role e xpectations (Hedge et al., 1990, p. 1075). Similarly, success of political control depends on the degree of coincidence of national and sub-national preferences (Hedge and Sci cchitano, 1994). As such, even though this study 97 See also: Borzel et al. 2004; Borghetto et al. 2005 98 For a discussion on quantitative measures of corporatism refer to Kenworthy (2000). 224

PAGE 225

follows the third wave of EU implementation research and its focus on domestic politics, it affords two main criticisms pertaining to both methodological and theoretical shortcomings. On the one hand, both qualitative and quantit ative studies seem to suffer from selection bias, as they almost invariably exclude states from the analysis. Qualitative researchers tend to focus on countries with a bad record of complia nce (see UK), or countries with supposedly good records (see Germany), and disreg ard such countries as France (but see Falkner et al., 2005). While countries such as Austria, Finland, and Sweden, which, as already shown, possess the best compliance records on environmental policy in the EU are left out on both qualitative and quantitative studies. This could lead to inaccura te conclusions supporting the non-existence of compliance problems thesis (Borzel, 2001; 2003), as spatial variation is more evident if we include the aforementioned compliers in enviro nmental policy specifically. From their side, most quantitative studies either use data from before the 1995 accession, or use data up to 2004 (see Kaeding, 2006) that ex clude Austria, Finland, and Sweden from the analysis (f or a review of the literature, see Ma stenbroek, 2005). The exclusion of the environmental leaders in quantitative studies creates a gap in the possible hypotheses, as there is much to be said about the presence of these member states in European political institutions. On the other hand, although the third wave of research has added considerably to our understanding of non-compliance, mo st approaches are ra ther disjunctive than complementary to each other. The effect of the Eu ropean principals on inducing or forcing state compliance have largely been neglected by the l iterature, while the challenge is to theorize how and under what conditions domestic opposition plays a role, in a manner consistent to theory. As argued elsewhere in this study, the EU has both federa l and intergovernmental characteristics. When research seeks to understand a states choice to comply by focusing only on the effect of 225

PAGE 226

international/systemic pressures, it may miss the intricate interplay between systemic and domestic forces. In this respect this study has a series of additi onal interesting results to offer. Results that speak both to the international relations and regulatory federalism literatures, while not also losing the important focus on comp liance with EU directives and the EU studies literature. Studies have found that the structural properties of domestic politics, such as the number of veto players, have a significant impact on complia nce (Lampinen and Uusikyla, 1998; Giuliani, 2003; Linos, 2006; Kaeding, 2006; Perkins and Neumay er, 2007), while others do not (Mbaye, 2001; Borzel et al., 2004; Borghetto et al., 2005), in fact some find an insignif icant negative relationship between veto play ers and non-compliance (i.e. the existence of veto players increases compliance, Mbaye, 2001). Corporatism is found to be insignificant while also shown to have a negative effect on compliance (contrar y to the hypothesized relationship, Mbaye, 2001) or with a positive effect on compliance (Kaeding, 2006).99 Pluralism and partisan conflict is shown to increase non-compliance (Konig and Luetgert, 2008), while Kaeding (2006) finds a positive yet insignificant relationship. Focusing on the third wave of implementation research, this study finds that the existence of veto points (institutional structures that afford players with the ability to modify and block legislation) should not necessarily mean that the process of tr ansposition becomes slower (noncompliance). Five main insights come from the th eories on regulatory federalism. First, even the richest and most powerful (clout ) of actors (the industry), may not always be opposed to policy that goes against thei r presumed interests (Wurzel, 2002). Several theorists have falsely presumed the preferences of domestic actors (U nderdal, 1998; Grant et al., 2000; Weinthal and Parag, 2003; Beach, 2005; Perkins and Neumayer, 2007). This study argues that the preferences 99 For a discussion on quantitative measures of corporatism refer to Kenworthy (2000). 226

PAGE 227

of actors should always be problematized as, for instance, when industry is favorable to environmental legislation then compliance with environmental directives increases in the EU case. Hence, the business climate of a state should be taken into account as the industry represents one of the stronger players. The second insight, on the ability or mechanisms of powerful actors influence, is offered by the hypothesized relationships on overall commitment, industry clout and need for regulation in this study. A states commitment to environmental regulation is found to have a positive effect on compliance as investment in implementation and pollution abatement rises and reflects this commitment. But any increase in compliance with environmental directives from states will have a negative effect on industry. It is, thus, al so expected that the industry will push for less legislation or for more non-compliance with EU directives. However, this purported effect should reach a plateau wherein after industry em ployment reaches a certain point, any additional industry employment (clout) and its influen ce on non-compliance should level off due to elevated sanctioning by the Commissi on. The findings in the statistical model are consistent with this effect. I find that industry clout increases non-comp liance (as per the hypothesized relationship), but this effect leve ls off and becomes negative with higher levels of industry clout. On a related front, states with a greater need for regulation actually comply less with environmental directives. It is true that if the environmental problem is big, then the resources need to alleviate it will also be considerable (Z ito, 2000). This makes it necessary for states with a big pollution issue to invest more resources than other states, but it does not necessarily mean that states with a greater need for regulation (more pollution) will automatically invest more resources. This is because heavily polluted st ates will have a corresponding large industrial sector that will mobilize against environmental regulation and investment. Alternatively, given 227

PAGE 228

the responsiveness of regulators to various political forces, it is not surprising that the rate of compliance does not reflect the objective need for regulation. Taken togeth er, the significance of these findings on industry clout a nd need for regulation, suggest that the presen ce of powerful interests will affect compliance with environmental directives. These findings are consistent with previous research on the effect of domestic actors on implementation but offer a more theoretically rigorous picture of the push-and-pull between private interests and governments, which should he lp to enhance the cons istency of empirical results, a major problem with most studies fo cusing on the structural properties of domestic politics, as mentioned earlier. However, the ab ility of players to us e veto points to stall compliance, will also depend on the receptivity of the state and its bureaucracy to behavioral modification (Wood, 1988). Hence, an additional insight is offered by the findings of this study on the interplay between administrative capacity corruption, and the existence of cozy regulatory efficiency. The only factors that seem to find support in most quantitative analyses in the EU studies literature are the various aspects of administrative capabilities (M baye, 2001; Borzel et al., 2004; Linos, 2006; Borghetto et al., 2005; Berglund et al., 2005). Bureaucratic efficiency is thought to decrease non-compliance (also Haverland a nd Romeijn, 2007), but Perkins and Neumayer (2007) find that efficiency has an insignifican t positive effect on non-compliance. Additionally, Pridham (1994, p. 99) argues that 'the southern countries (Spain, Greece, and Italy) do have particular problems of administrative procedur e and competence' while Borzel (2000) finds no southern problem. Administra tive efficiency is notori ously difficult to measure100 and the conflicting results in the literatur e might be due to just bad indi cators. However, there may be 100 For a discussion of the difficulties and the inadequacy of most bureaucratic quality i ndicators, see Van De Walle (2005). 228

PAGE 229

another explanation. As shown earlier, this study finds that wh en industry is favorable to environmental legislation then compliance will in crease. Additionally, when industry has clout in domestic politics, compliance will decrease. This, of course, points to the existence of a cozy relationship between indu stry and the state. It may be that administrations are efficient in doing what they do, and states do have the capacity to implement legislation. Regardless of the existence of veto points, and the degree of fitness between European requirements, it can be argued that some states are simply focused on doing different things (that serve th eir preferences), and are, indeed, very efficient in doing those. For instance, it is true that some states (especially the poor ones) are more focused on private sector development rather than environmenta l protection; hence their bureaucracies will be efficient in promoting private sector development rather than environmental protection. Even if these two goals are not necessarily mutually exclusive, the efficiency focus of the administration and its capacity is pointed toward different goals and values It should be then more important, to argue on how administrations internalize new values than whether their structure precludes change (as with the goodness of f it argument). A possible route available to private actors may be to block legislation by directly influencing the states capacity to implement legislation using administrative agen cies. Understaffed agencies will have a harder time participating in the process of transposition, and be even more susceptible to other players in the process. However, this study does not find a statistically signifi cant relationship between the staffing levels of administrative agencies charged on the environment and levels of compliance. Connected to administrative capacity this finding suggests that the capacity of administrations is not an important factor in determining compliance. When coupled with the above finding on the significance of government al commitment, this finding shows that 229

PAGE 230

compliance has little to do with administrative capacity and mo re to do with whether states choose to channel that ca pacity to environmental legislation or not. To be concise, bureaucratic efficiency does not matter as much as whether that efficiency is focused toward environmental legislation or private sector development, which invariably helps industry. The existence of a cozy regulatory efficiency focus is highly correlated with noncompliance in the environmental sector. The more states are focused on private sector development the less is the compliance with envir onmental directives. The capacity of a state to comply with environmental directives and bureaucratic efficiency in general, have little to do with actual compliance, it is in fact the ability of powerful interests, in focusing the efficiency of the government toward their own in terest that hurts compliance. The EU studies literature has addressed this coziness under such mediating variables as veto poi nts (Haverland, 2000), corporatism (Lampinen and Uusikyla, 1998; Mb aye, 2001; Kaeding, 2006), and the level of partisan conflict and polarization within domes tic veto players (Treib, 2003; Kaeding, 2006; Konig and Luetgert, 2008). The mere presence of veto points and powerful actors uti lizing them, or the other mediating variables aforementioned will not be sufficient in blocking compliance. It is actually the entrance of these actors into a cozy relationship with the government that affects compliance, while efficiency has little to do with it. Additionally, it is true that without the existence of corruption, interests could never enter the cozy relations hip mentioned above and focus bureaucratic efficiency to serve their goa ls. This is because, in systems ridden with corruption, tasks are accomplished only when bureau crats have a personal incentive to get things done. As such, corruption and cozy efficiencies go hand in hand since non-compliance produces a side-payment to bureaucrats from a much po werful and wealthy indus try sector. Consistent 230

PAGE 231

with this assumption, this study finds corruption to play a statis tically significant positive role on non-compliance. As a modest advice to practitioners of implementation, it becomes clear that the elimination of corruption closes the door to private in terests that might ot herwise capture not only the political principals of bureaucracies but also the whole regulatory efficiency focus of a state. Finally, EU compliance is increasingly m odeled along the lines of sociological institutionalism. Such studies start from the assumption that a rule will be complied with if it is deemed appropriate by the stakeholders (Mastenb roek, 2005). It is argued that compliance will not be automatic and will dependent on a process whereby the rule becomes internalized through socialization, persuasion, or learning (Finnemore and Sikkink, 1998; Checkel, 2001; Risse, 2000; Sending, 2002). This approach argues for the innate notion of the culture of compliance, which holds that member states differ in their basic proclivity to comply with EU requirements (Tallberg, 2002, p. 619; Falkner et al., 2005). However, these propositions are not unlike mo st quantitative and qualitative studies. To group states according to the most dominant ch aracteristic they exem plify is a rather deterministic approach that neglects the value of both the top-down and bottom up influences on compliance. To say that Greece, France or Portugal will always behave the same, regardless of the existence of top-down mechanisms to induce compliance from the Commission (like monitoring, or enforcement), is to miss half of the explanatory power that these top-down mechanisms offer. A typology likes this seems to suggest that any such mechanisms will be inefficient in raising governmental concern, and thus compliance, through the use of nongovernmental actors (which are assumed to be ab sent in these countries). This study argues that both top-down and bottom up mechanisms will be effective in determining compliance, while 231

PAGE 232

the presence of domestic actors will be the biggest qualifying f actor of success in all member states, regardless of the prevailing legal culture or administrative style, or veto points. To be precise, the presence of domestic ac tors or their relative power to affect results can be influenced by supranational institutions as well, and not only by the domestic legal traditions and institutionalized structures. Along the same theoretical lines, it has been argued that support fo r European integration is an important factor that facilitates compliance (Mbaye 2001: public support; Linos, 2006: support by government parties), while others do not find such a relationship (Lampinen and Uusikyla, 1998). And some have even found statis tically significant negati ve correlation between these two variables (Borzel et al., 2004). However, it can be also argued that support for the EU might not be evident in all policy domains. After all, it would take a lot of entrepreneurship to create the required linkages be tween policy domains to achieve normative compliance in all sectors of policy. Consequently, decisions to comply in one area must not be confused as normative acceptance of compliance in all policy ar eas as there is a considerable amount of linkage required to disseminate normative complia nce to other policy sectors, which have actors with different interests, ideas and beliefs (Haas, 1998). Hence, this study offers a novel and considerable contribution to theory and empirical investigation as it is not wise to assume that states with favorable EU sentiments will invari ably comply with all EU activity. Thus, even though, normative acceptance may work for the envi ronment to increase compliance, it may not work in other sectors of policy, providing us with a possible explanat ion of non-compliance in other sectors. Theoretical and empirical inves tigation, it seems, is not with out a sense of irony. At the same time as qualitative studies in the thir d wave have increasingly accepted the domestic 232

PAGE 233

political dimension of compliance, the results of quantitative research seem to point back to the arguments of the early a-politic al research that stressed the importance of efficient and well coordinated administrations (Treib, 2006). However, we should not see this as a weakness. This study uses theories of regulatory federalism to illustrate a theore tically and empirically rigorous way to bring politics back into the picture of implementation studies wi thout forgetting that influences might come from above, below, and within administrations; something which is consistent with the intergovernmental setting in which regulation takes place in the European Union. 7.4 Alarms, Patrols, Sticks, and Carrots: Enforcement Versus Management The EU is a unique type of institution. On the one hand, the EU is not an intergovernmental organization as it has an asso rtment of supranational institutions, which have the power of co-deciding with the governments of the member st ates in an increasing number of competencies. On the other hand, it is not a federal state either, because it neither has submitted to constitutional law nor do those institutions constitute a strong center with supreme powers, as the federal authorities in a federation (Sbragia, 1993), as states play an important role in decision-making. Additi onally, the European Union uses an integrated system of governance characterized by a mi x of deterrence and cooperation (Gormley, 1998). Even though in later years the focus and rhetoric has switched to more cooperative models and decentralization in the policy making process, th e Commission still uses deterrence instruments by a large extent (Holzinger et al., 2006). The primary centra lized enforcement mechanisms (monitoring, sanctions) are complemented with management tools seeking to enhance the capacity of states to comply and prevent involuntary no-compliance. To address this unique type of institution, this study takes two impor tant steps. If we conceive the issue of non-compliance in the EU as an internal political co nflict between a central 233

PAGE 234

authority and a set of semi-autonomous sub-units; then models of state implementation of federal policies drawn from the American context (in particular theories of regulatory federalism and bureaucratic control), may be particularly useful in explaining patterns of non-compliance in the EU context. Under this context, when regulation is centralized the focus is on uniform measures, which must be satisfied in all locations, us ing a great deal of top-down influence and deterrence. While, when regulation is decen tralized the focus is on local choice and cooperation leading to non-uniform implementation and enforcement of regulations. The first image focuses on central controls over agency st ructure, tasks and emphasizes the importance of elected officials in determining bureaucratic behavior, while the second image focuses on the nature of the policy task and the organizational characteristics of the bureaucracy (Scholz and Wei, 1986, p. 1249). A similar adaptation exists in the international relations literature. This time, noncompliance can be voluntary (cost-avoidance) or involuntary (lacking capacity). In voluntary non-compliance, state willingness and strategic choice has been shown to be affected by such factors as the goodness of fit, veto points, dom estic actor preferences (with limited success and mixed results). The major mechanism to deal wi th compliance problems, under this approach, is by increasing the likelihood and costs of detection through monitoring and th e threat of sanctions (Tallberg, 2002). In in -voluntary non-compliance, the focu s is on the domestic administrative capacity to comply, or legal ambiguity and comp lexity. By consequence, non-compliance is best addressed through a problem-solvi ng strategy of capacity buildi ng, rule interpretation, and transparency, rather than through coercive en forcement from supranational institutions (Keohane, Haas, and Levy 1993; Janicke, 1990). 234

PAGE 235

The similarities between theories of regulatory federalism and international relations theories are obvious. These similarities allow for the examination of the EU implementation deficit under both rubrics without necessarily making a distinction be tween the two from previous studies. However, this study argues that while both theories cannot, and should not, be considered in disjunctive terms, the theoretical mu rkiness that previous research has created must be addressed. Theorists have offered simple administrative explanati ons, followed by strategic choice explanations, followed by ma nagement explanations comple mented with strategic choice and administrative ones. All the while, empirical results are mixed, confusing, and lend little support to the theories used to ex tract explanations, adding very litt le to theoretica l and empirical cumulativeness. To address these issues, this study, first, decomposes the theoretical underpinnings of previous approaches, second, suggests the use of regul atory federalism and international relations as the gui des to research, and th ird, it recomposes th ese two theories under a common empirical model focusing on the ability of multiple actors to affect compliance. The benefit is clear. To understand what affects compliance we need to focus on mechanisms forgotten by previous researchers along the way of empi rical and theoretical disconfirmation and patchwork theorizing. For instance, even though o ffered by international relations theories, the effects of the European principals (Com mission) in capacity building or enhancing of governmental concern have been largely neglected. Similarly, even though offered from regulatory federalism, the concept of po licy salience was masked under the overarching effect of adaptation pressure. Guided by international relations theories, this study evaluates for the first time the ability of the EU to condition the behavior of member states using its enforcement and management mechanisms. This study finds a statistically significant negative relationship between 235

PAGE 236

Commission funding for Environmental Non-Governmental Organizations (ENGOs) and noncompliance, indicating that increas ed funding to domestic NGOs is more likely to be associated with increased levels of compliance. This finding is in line with neoliberal institutionalisms (management approach) suggestions on the ability of institutions to enhance governmental concern by enhancing the power of sub-nati onal actors. Even though this finding can be associated with the voluntary camp of state/actor choice, it is actually the effect of the principal (EU through the Commission) that enhances compliance by tending to the needs of these domestic forces or recognizing their e ffect and using it to ensure compliance.101 The ability of the EU to offer a reason able contractual environment by monitoring compliance and assuring states that cheating will be detected and punished, is found to be statistically significant, but the effect of European Court of Justice (ECJ) cases against member states is found to actually increase non-compliance in the next year. It seems that filling of a case with the ECJ will affect next years transpositio n record negatively because states are not afraid that the case will be judged any time soon (owing to ECJ backlogs), or because they hope to resolve the case before it truly reaches the Court, or finally, because they are simply not afraid of the Commission sticks. It may al so be that (some) states are not averted to being named and shamed by the Commission. The ineffectiveness of enforcement mechanisms to ensure future compliance should be kept in mind when assessing tools of compliance. Pr evious literature has neglected the importance of this mechanism a nd used its output as an indicator of noncompliance. For reasons mentioned elsewhere 102 in this study, using an enforcement 101 The ability of the EU to enhance the cap acity of states is found to be insign ificant. This may very well be because this management mechanism is inadequate, but it may also be that EU funding is enough to enhance the ability of environmental expenditure by the state. 102 For more on this, please refer to Chapter 6 of this study, section 6.3. 236

PAGE 237

mechanisms proceedings in measuring compliance not only represents a statistical impossibility but it also confounds the importance of t hose mechanisms in ensuring compliance. Guided by theories of regulatory federalism, th e previous section illu minated the strengths and shortcomings of strategic ( voluntary) explanations of non-compliance. What remains here is where we go from here. Taken together, the findings of this study lead to three conclusions. First, the reasons for non-compliance (or compliance) ar e multiple and complex. It is clear that variations in compliance cannot be reduced to a single explanati on, even drawing from a single theory. Evidence for one or the other model hardly constitutes conclusive evidence that it is more important in explaining non-compliance. Adding several variables that represent different theoretical approaches to one full econometric m odel, allows for capturing the full effect of all variables when taken together rather than disj unctively. The benefit is both theoretical and empirical. Adding all approaches to a single model allows the researcher to determine whether variables capturing predictions might lose thei r explanatory power once variables from other theoretical approaches are included. This allows for an evaluation of whether a theoretical approach actually adds to explanatory power. If this is the case, then the approach should be thought as complementing existing th eoretical approaches (rather th an be thought in a disjunctive way), when attempting to gain a full understanding of the bigger picture on what determines noncompliance. Second, given the contextual nature of several of the variables offered in this study, the results cannot and should not be taken as easily generalizable. Even though the environment sector is the highest in noncompliance and expends most of the Commission and private actor resources in monitoring and enforcement, severa l variables in this study are contextual. For instance, membership in the European Parliame nts Environment Committee, the salience of the 237

PAGE 238

environment with state publics and the support of state publics for EU environmental policy, among others, are contextual to the environmen t. This should not be considered a weakness though. As argued earlier, it is a mistake to assu me that support for the EU means support for EU involvement in environmental po licy. Decomposing support for the EU, as encountered in other studies under the rubric of constructivism (mostly), into its contextual counterpart serves a great purpose. It may very well be that it is the suppo rt of state publics for EU involvement in a particular sector of policy that might be held ac countable for non-compliance. It could be that the salience of unemployment issues leads to complian ce with internal market directives. Different policy types might imply different conflicts and therefore lead to different types of implementation processes (Treib, 2006). Nevertheless, the caution shoul d be clear. Even though cont extual explanations are abundant, cross-sectoral ones are in short supply. Most quant itative implementation studies focus on one sector, while the same is true of qualitativ e studies (but see Siedentopf and Ziller, 1988; Duina, 1999). It is true that all of us cover cross-sectoral data, but none has sought to test whether there are cross-sectoral differences in co mpliance patterns. It may very well be that noncompliance in one sector is due to compliance in a nother sector or even mo re interestingly due to non-compliance in another sector, which may or may not verify theories of cultures of compliance (Falkner et al. 2005) for instance. Treib (2006) al so argues on broadening our perspectives to cover policies that have attrac ted little attention, such as agricultural policy, which might not have so many implementation pr oblems but is the sector where the biggest money games are played. Finally, even if all quantifiably possible vari ables are included in this study, it does not expend all the variance in the data. To be fair, this study does provide th e largest f it statistics 238

PAGE 239

239 ever encountered in quantitative studies, pointing to the benefits of the theoretical approach in this study, but a considerable amount of variation remains unexplained.103 To start adding variables without theoretical background would not be prudent. It may very well be that the EU case is as unique as the theories needed to explai n it. But it may also be that that we need to model the compliance deficit facing EU policy using a more theoretically integrative and methodologically rigorous design. Mastenbroek (2005) suggests we use using a mixed-method design, running a statistical model containing as many variables as possible, and then proceeding to analyze any remaining variance comparing we ll and poorly explained cas es using qualitative methods. But another route may also be prudent. Instead of supplementing quantitative studies with qualitative explanations of the remaining vari ance, we could conduct co llaborative research projects such as the one from Falkner et al. ( 2005) which demonstrates th at qualitative research is not always confined to small-N settings. It is true, however, that even quantitative studies can achieve high levels of understand ing, if they stay true to th e model of theoretical understanding offered in this study. Theoretica l clarity can be achieved by allo wing theories to speak without muddling them with supe rfluous and contradictory explanatio ns. Empirical reliability can be achieved if theories are allowe d to see different spaces of implementation troubles. Whether these are top or bottom, enforcement or manageme nt, they all are parts of the full picture as implementation in the EU is an intergovernmental enterprise and it should be understood in its own terms, always keeping in mind to look both ways. 103 This study attains a PseudoR2 of 0.39 for the full model. Compared to the 0.16 of Mbaye (2001), and 0.26 for Kaeding (2006). While also, th e AIC and BIC values are considerably lowe r in this study, indicating a better fit. Perkins and Neumayers (2007) full model has an AIC of 1176.0 and a BIC of 1304.4, as compared to the AIC of 634.9 and a BIC of 677.1 in this study.

PAGE 240

APPENDIX A TRANSPOSITION DATA, ISSUES AND CHANGES Data extraction is straightforward when it comes to years 2001-2007 with only a few changes needed as the Commission switched th e columns to rows from 2002-2004 (of course there is no way to download the data in a usab le format, and it all has to happen by hand). The situation worsens when it comes to 2001 downwar d as the commission added more details in the environmental sub-sectors and introduced new sectors in 2001 such as Fisheries and Trade (which do not exist in 2002-2007, at least not in the way reported in 2001). Fisheries and Trade disappear again in 2000 while the remaining data and subsectors remain the same with 2001. Following is a direct depicti on of the existing chapter/sect ors/subsectors in 1999 and their rearrangement to match the 2000 data (w hich matches the 2001 onwards data): Table A-1. Commission categories and sectors 1999 Original Table 1999 matching 2000 Chapter Sector Sub-sector Chapter Sector/Subsector Internal Market Free movement of persons and citizenship of the union Entry and residence Enterprise Motor vehicles Right to vote and eligibility Cosmetics Regulated professions Pharmaceuticals Commercial agents Construction products Free movement of goods Technical regulations Textiles and shoes Food Capital goods Pharmaceuticals Chemicals Chemicals Technical regulations Motor vehicles Competition Construction products Employment and Social Affairs Safety and hygiene at work + Employment legislation = Social Affairs Capital goods Nondiscrimination = Entry and residence 240

PAGE 241

Table A-1. Continued 1999 Original Table 1999 matching 2000 Chapter Sector Sub-sector Chapter Sector/Subsector Cosmetics Free movement of workers = Right to vote and eligibility + Commercial Agents Textiles and shoes Agriculture Liability for defective products Transport All Special freedom of movement arrangements Energy All Customs Union Environment Freedom of access to information Free movement of services Financial services Environmental impact assessment The Business Environment Company law Air Intellectual property Water Data protection Nature Public procurement Noise Direct taxation Chemicals and Biotechnology VAT Waste Other indirect taxes Environment and Industry Competition Radiation protection Social Affairs Information Society = Telecommunications Telecommunications Energy Internal market Free movement of goods = Special freedom of movement arrangements, Liability for defective products Transport Free movement of services = Financial Services Telecommunications Business environment = Company law, Intellectual Property, Data protection, Public procurement 241

PAGE 242

Table A-1. Continued 1999 Original Table 1999 matching 2000 Chapter Sector Sub-sector Chapter Sector/Subsector TV without Frontiers Professions regulated as regards qualifications = Regulated professions Environment Freedom of access to information Taxation and customs unions Direct taxation Environmental impact assessment VAT Air Other indirect taxes Water customs union Nature Education and culture = TV without Frontier television without frontier Noise Consumer and health protection veterinary matters Chemicals and Biotechnology plant health Waste seeds and plants Environment and Industry Food Radiation protection Animal feed = feeding stuffs Agriculture veterinary matters forestry plant health Consumer protection = Consumers seeds and plants Eurostat =Statistics feeding stuffs forestry Consumers Eurostat Starting then from the Enterprise Sector for 1999 (which does not exist), it was created to match its existence in 2000 by using the subsectors as they existed in the Internal Market chapter for 1999. The subsectors of Employment and social Affairs for 2000 where: Employment legislation, Non-discri mination, Free movement of worker s, and Safety and hygiene at work. These subsectors do not exist in 1999 so the data where extracte d from similar subsectors under different sectors in 1999 that do not exist in 2000 (i.e. Employment and So cial Affairs: Safety 242

PAGE 243

243 and hygiene at work + Employment legislation = Social Affair s in 1999, Non-discrimination = Entry and residence in 1999, Free movement of workers = Right to vote and eligibility + Commercial Agents in 1999). The 1998-99 Tel ecommunications chapter becomes Information Society for 2000-07 while the tv without frontie rs chapter of 1999 was changed to Education and culture for 2000-07. Table A-2. Employment and Soci al Affairs: 199 9 to 2000 changes Sector 2000 Subsector 2000 Subsectors in 1999 Employment and Social Affairs Employment legislation Social Affairs Safety and hygiene at work Non-discrimination Entry and residence Free movement of workers Commercial agents Right to vote and eligibility Finally, the internal market chapter for 2000 includes such subsectors as: Free movement of goods, Free movement of services, Business environment, Professions regulated as regards qualifications. Which include Special freedom of movement arrangements+ Liability for defective products; Financia l Services; Company law+In tellectual Property+Data protection+Public procurement; and Regulated professions respectively from 1999. The matching sectors for 1998-99 and 2000-07 where of course left unchanged. The aforementioned changes were also performed for the 1998 data which has identical chap ter/sectors/subsectors with 1999. Table A-3. Internal Market: 1999 to 2000 changes Sector 2000 Subsector 2000 Subsectors in 1999 Internal market Free movement of goods Special freedom of movement arrangements, Liability for defective products Free movement of services Financial Services Business environment Company law, Intellectual Property, Data protection, Public procurement Professions regulated as regards qualifications Regulated professions

PAGE 244

APPENDIX B FAVORABLE PUBLICS IN TH E EUROBAROMETER SURVEY The 2007 question reads: QA20a: For each of the following areas, do you think that decisions should be made by the (NATIONALITY) Government, or made jointly within the European Union? (READ OUT ROTATE): 1. Fighting Crime, 2. Taxation, 3. Fi ghting unemployment, 4. Fighting terrorism, 5. Defense/Foreign Affairs, 6. Immigration, 7. Education, 8. Pensions, 9. Protecting the environment, 10. Health and social welfare, 11. Agriculture and fi shery, 12. Consumer protection, 13. Scientific and technological research, 14. S upport for regions facing economic difficulties, 15. Energy, 16. Competition, 17. Tran sports, 18. Economy, 19. Fighting Inflation. Following is a table with all the re levant information for each year: Table B-1. Data sources for Favorable Publics Year Question Number Eurobarometer Number Weblink 2007 QA20a 68.1 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/23368.xml 2006 QA24 66.1 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/21281.xml 2005 QA31 64.2 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/04580.xml 2004 Q.35 62.0 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/04289.xml 2003 Q.28 60.1 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/03991.xml 2002 Q.25a 58.1 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/03543.xml 2001 Q.30 56.2 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/03480.xml 2000 Q.33 53.0 http://www.icpsr.umich.edu/icps rweb/ICPSR/studies/03064.xml 1999 Q.21 51.0 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/02865.xml 1998 Q.36 50.0 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/02830.xml 244

PAGE 245

APPENDIX C SALIENCE IN THE EU ROBAROMETER SURVEY The 2007 question reads: QA6a: What do you think are th e two most important issues facing (OUR COUNTRY) (SHOW CARD READ OUT MAX. 2 ANSWERS): 1. Crime, 2. Economic situation, 3. Rising prices\ inflation, 4. Taxation, 5. Unemployment, 6. Terrorism 7. Defence\ Foreign affairs, 8. Housing, 9. Immigration, 10. Healthcare syst em, 11. The educational system, 12. Pensions, 13. Protecting the environment, 14. Energy relate d issues, 15. Other (SPONTANEOUS), 16. DK. Following is a table with all the re levant information for each year: Table C-1. Data sources for Salience Year Question Number Eurobarometer Number Weblink 2007 QA6a 68.1 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/23368.xml 2006 QA23 66.1 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/21281.xml 2005 QA30 64.2 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/04580.xml 2004 Q.33 62.0 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/04289.xml 2003 Q.26 60.1 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/03991.xml 2002 Q.2 57.2 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/03543.xml 2001 Q.3 56.3 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/03480.xml 2000 http://ec.europa.eu/public_opinion/archives/eb/eb54/eb54_ann.pdf 1999 Q.2 51.1 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/02865.xml 1998 Q.21 50.0 http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/02830.xml Years 2002-2007 have the same question, while for 2001,104 1999,105 and 1998106 the most similar question changes. Unfortunately the most similar question for 2000 from Eurobarometer 104 Q.3. In your opinion, what are the two most positive aspects of the European Union today? 105 Q.2. Here is a list of things that some people say they are afraid of. Please tell me which one worries you the most? (SHOW CARD ONE ANSWER ONLY) b) And next most 106 Twenty five years ago, the member states of the Europ ean Community were trying to solve a certain number of common problems together. Here is a list of some of them. Could you please tell me which one of the problems, you think, is the most important at the present time? (SHOW CARD) b) And which is the next most important problem? (SHOW SAME CARD) 245

PAGE 246

246 53 reads Q.43. Here is a list. Can you tell me which policy area, or areas, the European Parliament should pay particular attention to to defend your interests? You can choose a maximum of three where the environment choi ce was lumped protection of the consumer, hence a linear interpolation of the values was necessary for that year.

PAGE 247

APPENDIX D RANDOM EFFECTS POISSON WITH TIME DUMMIES Table D-1. Random Eff ects Poisson regression m odel with time dummies Variables Random Effects Bargaining power 0.023* (Co) MV (0.01) Governmental concern -0.012** ( ) NGOfun (0.01) Contractual environment 0.038* ( ) 1 tECJ (0.02) Governmental capacity 0.003 ( L IFE) (0.01) Approval of EU action -0.018*** ( E Uact) (0.01) Citizen groups -0.001 (Co) mplaints (0.00) ENVI membership -0.005 ( E NVI) (0.02) Salience -0.039** ( ) Salience (0.02) Green party membership -0.001 (Gr) een (0.04) Overall commitment -0.218 ( ) EnvExp (0.34) Favorable industry climate -10.240 ( E MAS) (16.58) Industry clout 0.215 ( ) IndEmpl (0.27) Industry clout squared -0.009 ( ) IndEmplSq (0.01) Need for regulation 0.046 ( E mission) (0.03) Staffing levels 0.052 (Com) pen (0.41) Corruption -0.203* (Corr) (0.12) Cozy regulatory efficiency 0.745 ( ) RegQual (0.49) 247

PAGE 248

248 Table D-1. Continued Variables Random Effects Year1998 -0.281 (0.34) Year1999 -0.538*** (0.18) Year2000 -0.462** (0.19) Year2001 -0.236* (0.14) Year2003 -0.728*** (0.18) Year2004 -0.294* (0.18) Year2005 -0.665*** (0.20) Year2006 -1.813*** (0.30) Year2007 -1.726*** (0.32) Constant ( ) 3.129 (2.44) Log likelihood -303.4 Information Criteria: AIC BIC 662.9 747.3 Notes: significant at 0.10 level, ** significant at 0.05 level, *** significant at 0.01 level. Standard errors obtained through bootstrapping with 100 replications are in parentheses.

PAGE 249

LIST OF REFERENCES Aberbach, J. & Rochman, B. (1976). Changing Be liefs Within the Executive Branch: The Nixon Administration Bureaucracy. American Po litical Science Review, 70 (2), 456-468. Alter, K. (2000), The European Unions Lega l Syatem and Domestic Policy: Spillover or Backlash? International Organization, 54(3), 489-518. Axelrod, R. & Keohane, R. O. (1986). Achieving Cooperation U nder Anarchy: Strategies and Institutions. In Kenneth A. Oye (ed.), Cooperation Under Anarchy (226). Princeton, N.J.: Princeton University Press. Axelrod, R. A. (1984). The Evolution of Cooperation. New York: Basic Books. Baltagi, B. (2008). Econometric Analysis of Panel Data. 4th edition, John Wiley & Sons Ltd. UK. Banks, J. & Weingast, B. (1992). The Political Control of Bureaucraci es Under Assymetric Information, American Journal of Political Science, 36 (2), 509-525. Bardach, E. (1977). The Implementation Game: What Happens after a Bill becomes a Law. Cambridge, Mass.: MIT Press. Barkin, S. & Shambaugh, G. (1999). Hypotheses on the international Pol itics of Common Pool Resources. In Barkin, S. & Shambaugh, G. (eds), Anarchy and the Environment: The International Relations of Comm on Pool Resources, State University of New York Press. Barnes, P. & Barnes, I. (2000). Environmental Policy in the European Union, London: Edward Elgar Pub. Bartels, Brandon L. (2008). Beyond Fixed Versus Random Effects: A Fr amework for Improving Substantive and Statistical Anal ysis of Panel, Time-Series Cross-Sectional, and Multilevel Data. Political Methodology Conference, A nn Arbor, MI, July 9-12 2008. Retrieved from http://home.gwu.edu/~bartels/research.html. Beach, D. (2005). Why Governments Comply: An Integrative Compliance Model that Bridges the Gap between Instrumental and Normative M odels of Compliance, Journal of European Public Policy, 12(11), 113. Beck, N. & Katz, J. (1995). What to Do (and Not to Do) with Time-SeriesCross-Section Data, American Political Science Review, 89, 634-647. Beck, N. (2001). Time-Series Cross-Section Data : What Have We Learned in the Past Few Years? Annual Review of Po litical Science, 4, 271-93. Benda, P. & Levine, C. (1988). The Reagan Bur eaucracy: The Bequest, The Promise, and the Legacy. In Jones, C. (ed.), The Regan Leg acy: Promise and Performance. Chatham, NJ: Chatham House. 249

PAGE 250

Berglund, S., Gange, I. & Van Waarden, F. ( 2006). Mass Production of Law. Routinization in Transposition of European Directives. A Soci ologicalInstitutiona l Account, Journal of European Public Policy, 13(5), 692. Bernstein, S. (2002). Liberal Environmentalis m and Global Environmental Governance, Global Environmental Politi cs, 2:3. MIT. Borghetto, E., Franchino, F., & Giannetti, D. (2005), Complying w ith the Transposition Deadlines of EU Directives: Evidence from It aly. Paper presented at the 3rd ECPR General Conference, Budapest. Bertola, G., Boeri, T. & Cazes, S. (2000). Empl oyment Protection in Industrialized Countries: The Case for New Indicators, Intern ational Labour Review, 139(1), 57. Borzel, T. & Risse, T. (2003). Conceptualizing th e domestic impact of Europe. In Featherstone, K. & Radaelli, C. (eds.). The Politics of Europeanization (pp. 57). Oxford: Oxford University Press. Borzel, T. (2000). Why there is no Southern Problem. On Environmental Leader and Laggards in the EU. Journal of European Public Policy 7(1), 141-162. Borzel, T. (2001). Non-compliance in the Europe an Union: pathology or statistical artefact? Journal of European Public Policy, 8(5), 803. Borzel, T., Hofmann, T. & Panke, D. (2005). Who s Afraid of the ECJ? Member States, Court Referrals, and (Non-) Compliance. Paper pres ented at the ECPR Joint Sessions, 14 April, Granada. Brinkerhoff, D. & Crosby, B. (2002). Ma naging Policy Reform, Kumarian Press. Bursens, P. (2002). Why Denmark and Belgium Have Different Implementation Records: On Transposition Laggards and Leaders in the EU Scandinavian Politi cal Studies, 25(2), 173 95. Calvert, R., McCubbins, M. & Weingast, B. (198 9). A Theory of Political Control and Agency Discretion. American Journal of Political Science, 33(3), 588-611. Cameron, C. & Trivedi, P. (1998). Regression Analysis of Count Data. Econometric Society Monographs No. 30, Cambridge University Press, New York. Cameron, C. & Trivedi, P. (2005). Microeconom etrics: Methods and A pplications. Cambridge University Press, New York. Cameron, C. & Trivedi, P. (2009). Microeconom etrics Using Stata. Stata Press, Texas. Chayes, A. & Chayes, A. H. (1993). On complia nce. International Orga nization, 47(2), 175-206. 250

PAGE 251

Chayes, A. & Chayes, A. H. (1995). The New Sovereignty: Complian ce with International Agreements, Cambridge, Mass.: Harvard University Press. Checkel, J.T. (2001). Why comply? Social learning and European identity change. International Organization, 55(3), 473. Chubb, J. E. (1985). The Political Economy of Fe deralism. American Political Science Review, 79(4), 994-1015. Ciavarini Azzi, G. (1988). What Is This New Research into the Implementation of Community Legislation Bringing Us? In H. Siedentopf & J. Ziller (eds). Making European Policies Work: The Implementation of Community Legi slation in the Member States (pp. 190 201). London: Sage. Ciavarini Azzi, G. (2000). The slow march of European legisl ation: the implementation of directives. In K. Neunreither & A. Wiener (e ds), European Integration after Amsterdam: Institutional Dynamics and Prospects fo r Democracy (pp. 52). Oxford: Oxford University Press. Collins, K. & Earnshaw, D. (1992.) The implementation and enforcement of European Community legislation. Envir onmental Politics, 1(4), 213. Connolly, B. (1999). Asymmetrical Rivalry in Common Pool Res ources and European Responses to Acid Rain. In S. Barkin & G. Shambaugh(eds.). Anarchy and the Environment: The International Relations of Common Pool Res ources, State University of New York Press. Coyle, C. (1994). Administrative Capacity and the Implementati on of EU Environmental Policy in Ireland. Regional Politics and Policy, 4(1), 62-79. DeSombre, E. (2000). Domestic Sources of In ternational Environmental Policy: Industry, Environmentalism, and U.S. Power. MIT Press. DeSombre, E. (2002), The Global Environmen t and World Politics. Continuum, NY. Dimitrakopoulos, D.G. (2001a). The transpositio n of EU law: post-decisional politics and institutional economy. Europ ean Law Journal, 7(4), 442. Dimitrakopoulos, D.G. (2001b). Learning and steering: changing implementation patterns and the Greek central government. Journal of European Public Policy, 8(4), 604. Dimitrova, A. & Rhinard, M. (2005). The power of norms in the transposition of EU directives. Paper prepared for the ECPR Joint Sessions, Granada, Spain, 14 April 2005. Dimitrova, A. & Steunenberg, B. (2000). The Sear ch for Convergence of National Policies in the European Union: An Impossible Quest? European Union Politics, 1(2), 201. 251

PAGE 252

Dinan, D. (2005). Ever Closer Union An Introduction To European Integration. New York: Lynne Rienner. Downs, G. W., Rocke, D. & Barsoom, P. ( 1996). Is the Good News about Compliance Good News about Cooperation? Internatio nal Organization, 50 (3), 379 Duina, F. & Blithe, F. (1999). Nation-states a nd common markets: the institutional conditions for acceptance. Review of Internationa l Political Economy, 6(4), 494. Duina, F. (1997). Explaining legal implementation in the European Union. International Journal of the Sociology of Law, 25, 155. Duina, F. (1999). Harmonizing Europe: Nation-St ates within the Common Market. Albany, State University of New York Press. Eisner, M. & Meier, K. (1990). Presidential Control and Bureau cratic Power: Explaining the Regan Revolution in Antitrust. American J ournal of Political Science, 34(1), 269-287. Elmore, R. (1982). Backward Mapping: Implemen tation Research and Policy Decisions. In Williams, W. (ed.), Studying Implementation: Methodological and Administrative Issues (pp. 18). Chatham House Publishers, Chatham. Emmelin, L. (1972). The Stockholm Conferences. Ambio, 1(4), 135-140. European Communities (2001). Sixth Environm ental Action Program 2001-2010: Our Future, Our Choice. Luxembourg: Office for Official Publications of the European Communities. Evans, P., Jacobson, H, & Putnam, R. (eds .). (1993). Double-Edged Diplomacy. Berkeley: University of California Press. Falkner, G., Treib, O., Hartlapp. M. and Leiber S. (2005). Complying with Europe? The Impact of EU Minimum Harmonizat ion and Soft Law in the Member States. Cambridge: Cambridge University Press. Fearon, J. (1998). Bargaining, Enforcement, and International Coope ration. International Organization, 52(2), 269-305. Finnemore, M. (1996). National Interests in International Security. Cornell UP. Finnemore, M. and Sikkink, K. (1998). International norm dynamics and political change. International Organi zation, 52(4), 887. Fiorina, M. (1981). Congressional Control of the Bureaucracy: A Mismatch of Incentives and Capabilities. In Dodd, L. & Oppenheimer, B. (eds.), Congress Reconsidered. Washington: Congressional Quarterly. Freedman, D. & Peters, S. (1984) Bootstrapping a regression equa tion: some empirical results. Journal of American Statis tical Association, 79, 97 252

PAGE 253

From, J. & Stava, P. (1993). Implementation of Community law: the la st stronghold of national control. In S. Andersen & K. Eliass en (eds.), Making Policy in Europe: The Europeification of National Policy Making (pp. 55). London: Sage. Giuliani, M. (2003). Europeanization in perspec tive: institutional fit and adaptation. In K. Featherstone & C. Radaelli (eds.), The Polit ics of Europeanization ( pp. 57). Oxford: Oxford University Press. Golden, M. (2000). What Motivate Bureaucrats? Politics and Administra tion During the Reagan Years. New York: Columbia University Press. Gormley, W. (1989). Taming the Bureaucracy: Muscle s, Players, and Other Strategies. Princeton NJ: Princeton University Press. Gormley, W. (1998). Regulatory En forcement Styles. Political Re search Quarterly, 51(2), 363383. Gourevitch, P. (1986). Polities in Hard Ti mes. Ithaca: Cornell University Press Gourevitch, P. (1996). Domestic sources of intern ational cooperation. International Organization, 50(2), 349-73. Green Cowles, M., Caporaso, J. and Risse, T. (eds.). (2001). Transforming Europe: Europeanization and Domestic Change. Ithaca: Cornell University Press. Greene,W.H. (2007). Functional form and heteroge neity in models for count data. Foundations and Trends in Econometrics, 1 (2), 113-218. Haas, P. M., Keohane, R. & Levy, M. (eds.). (1993). Institutions for th e Earth: Sources of Effective International Environmental Pr otection. Cambridge, Mass.: MIT Press. Haas, P.M. (1998). Compliance with EU Directives: Insights from International Relations and Comparative Politics. Journal of Eu ropean Public Policy, 5(1), 17. Hartlapp, M. and Falkner, G. (2009). Problem s of Operationalization and Data in EU Compliance Research. European Union Politics, 10(2), 281. Hattan, E. (2003), The Implementation of EU En vironmental Law. Journal of Environmental Law, 15(3), 273-288. Haverland, M. (2000). National Adaptation to European Integration: The Importance of Institutional Veto Points. Journa l of Public Policy, 20(1), 83. Haverland, M. (2005). Does the EU cause dom estic developments? The problem of caseselection in Europeanization research. Eu ropean Integration online Papers 9(2). Haverland, M. and Romeijn, M. (2007). Do Member States Make European Policies Work? Analyzing the EU Transposition Deficit, Public Administration, 85(3), 757. 253

PAGE 254

Hayes, M.T. (1992). Incrementalism and Public Policy, New York: Longman. Hedge, D. M. (1993). A Spatial Model of Regulat ion. American Politics Quarterly, 22(4), 387409 Hedge, D. M. and Johnson, R. (2002). The Plot that Failed: The Re publican Revolution and Political Control of the Bureaucracy. Jour nal of Public Administration Research and Theory, 12(3), 333-351. Hedge, D. M. and Scicchitano, M. J. (1994). Regulating in time and Space: The Case of Regulatory Federalism. The Journal of Politics, 56(1), 134-153. Hedge, D. M., Jallow, S, (1990). The Federal Cont ext of Regulation: The Spatial Allocation of Federal Enforcement, Social Sc ience Quarterly, 71(4), 786-801. Hedge, D. M., Scicchitano, M.J. & Metz, P. (1 991). The Principal Agen t Model of Regulatory Federalism, The Western Political Quarterly, 44(4), 1055-1080. Heidenheimer, A., Johnston, M. & LeVine, V. (eds). (1989). Political Corruption: A Handbook. New Brunswick: Trans action Publishers. Helm, D. (2009). Climates of Change: Sustainabili ty Challenges for Enterprise. University of Oxford: Smith School of Enterprise and the Environment. Heritier, A. (1995). Leaders and laggards in European clean air policy. In F. van Waarden and B. Unger (eds.). Convergence or Diversit y? Internationalization and Economic Policy Response (pp. 278), Aldershot: Avebury. Heritier, A. (1996). The Acco mmodation of Diversity in Eu ropean Policy-Making and its Outcomes. In Jordan, A. (ed.). (2002). E nvironmental Policy in the European Union. Earthscan Publications. Heritier, A. (2001). Differential Europe: National Administrative Responses to Community Policy. In Cowles, M. Caporaso, J. & Risse, T. (eds.). Transforming Europe: Europeanization and Domestic Change (pp. 44). Cornell University Press, Ithaca. Heritier, A., Kerwer, D., Knill, C., Lehmkuhl D., Teutsch, M. & Douillet, A. (2001). Differential Europe: New Opportunities and Rest rictions for Policymaking in the Member States. Lanham, MD: Rowman & Littlefield. Hey, C. (2006). EU Environmental Policies: A short history of the policy strategies. In Scheuer, Stefan (ed.), EU environmental policy handbook a critical analysis of EU environmental legislation: making it accessible to environm entalists and decision makers. Brussels: European Environmental Bureau. Hilbe, J. (2008). Negative Binomial Regressi on. Cambridge University Press, New York. 254

PAGE 255

Hildebrand, P. (1993). The European Community s Environmental Policy, in Environmental Policy in the European Union. In Jordan, A. (ed.). (2002). Environmental Policy in the European Union. Earthscan Publications. Hix, S. (1999). The Political System of the European Un ion. St. Martins Press. Holzinger K., Knill, C. & Schafer, A. (2006). Rh etoric or Reality? New Governance in EU Environmental Policy. European Law Journal, 12(3), 403. Hox, J. (2002). Multilevel An alysis: Techniques and Applic ations, Mahwah, NJ: Lawrence Erlbaum. Hsiao, C. (2003). Analysis of Panel Data. 2nd edition, Cambridge University Press, New York. Huelshoff, M. (1992). Environmen tal Policy in the EC: Neo-Func tionalist Sovereignty Transfer or Neo-Realist Gate-Keeping. In ternational Journal, 136, 135-160. Ingraham, P. (1995). The Foundation of Merit: Public Service in American Democracy. Baltimore: Johns Hopkins University Press. Janicke, M. & Weidner, H. (eds.). (1997). Na tional Environmental Policies. A Comparative Study of Capacity-Bu ilding. Springer. Jeppesen, T. (2000). EU environmental policy: ever closer to th e citizen? European Environment, 10( 2), 96-105. Jeppesen, T. (2002). Environmental Regulation in a Federal System: Framing Environmental Policy in the EU. Edward Elgar Publishing. Jordan, A. & Jeppesen, T. (2000). EU environm ental policy: adapting to the principle of subsidiarity. European Environment, 10(2), 64-74. Jordan, A. (1999). The implementation of EU en vironmental policy: a policy problem without a political solution. Environment and Pla nning C: Government and Policy, 17, 69. Jordan, A. (2000), The politics of multilevel environmental governan ce: subsidiarity and environmental policy in the European Union. Environment and Planning, 32, 1307-1324. Judge, D. (1992). Predestined to Save the Eart h: The Environment Committee of the European Parliament. In Jordan, A. (ed.). (2002). E nvironmental Policy in the European Union. Earthscan Publications. Kaeding, M. (2006). Determinants of Transposition Delay in the European Union. Journal of Public Policy, 26(3), 229. Kaeding, M. (2008). Lost in Tr anslation or Full Steam Ahead? Eu ropean Union Politics, 9(1), 115. 255

PAGE 256

Kallestrup, M. (2002). European isation as a discourse: domesti c policy legitimization through the articulation of a need for adaptation. Public Policy and Admi nistration, 17(2), 110 24. Katzenstein, P.J., (ed.) (1978). Between Power and Plenty: Foreign Economic Policies of Advanced Industrial States. Madiso n: Wisconsin University Press. Kaufman, H (1967). The Forest Ranger. Washington DC, Resources for the Future. Keck, M. & Sikkink, K. (1998). Ac tivists Beyond Borders: Advocacy Networks in International Politics. Ithaca, NY: Cornell University Press Keefer, P. & Stasavage, D. (2003). The Limits of Delegation: Veto Players, Central Bank Independence, and the Credibility of Monetary Policy. American Political Science Review, 97(3), 407-423 Keohane, R. & Hoffmann, S. (eds.). (1991). The New European Comm unity. Boulder, CO: Westview Press. Keohane, R. & Martin, L. (1995). The promise of in stitutionalist theory. International Security, 20(1), 39-51. Keohane, R. & Levy, M. (eds.). ( 1996). Institutions for Environmental Aid: Pitfalls and Promise. Cambridge, Mass.: MIT Press. Keohane, R. & Milner, H. (eds.) (1996). Interna tionalization and Domestic Politics. Cambridge: Cambridge University Press. Keohane, R. & Nye, J. (1977). Power and Interdep endence: World Politics in Transition. Little, Brown. Boston. Keohane, R., Haas, P. & Levy, M. (eds.). (1993). In stitutions for the Earth: Sources of Effective International Environmental Protection. MIT press. Keohane, R.O. (1984). After Hegemony: Coopera tion and Discord in the World Political Economy. Princeton UP. Keohane, R.O. (1989), Internationa l Institutions and State Power: Essays in International Relations Theory. Boulder, Westview Press. Knill, C. and Lenschow, A. (1998). Coping with Europe: the impact of British and German administrations on the implementation of EU environmental policy. Journal of European Public Policy, 5(4), 595. Knill, C. and Lenschow, A. (2001), Adjusting to European Union Environmental Policy: Change and Persistence of Domestic Administrations In J. Caporaso, M. Cowles, & T. Risse (Eds.). Transforming Europe: Europeanizati on and Domestic Change. Ithaca: Cornell University press. 256

PAGE 257

Knill, C. and Lenschow, A. (2005). Compliance, Communication and Competition: Patterns of EU Environmental Policy Making and Their Impact on Policy Convergence. European Environment, 15, 114. Konig, T., Luetgert, B. & Dannwolf, T. (2006) Quantifying European Legislative Research: Using CELEX and PreLex in EU Legislative Studies. European Union Politics, 7(4), 553 574. Konig, T. & Luetgert, B. (2008). Troubles with Transposition: Explaini ng Trends in Member State Notification Failure and Timelines. British Journal of Political Science, 39(1), 163 94. Koutalakis, C. (2002). Making European Polic ies Work in the South. Explaining NonCompliance in Italy and Greece. Mimeo, Free University of Berlin. Krasner, S. (1982). Structural Causes of Regime Conseque nces: Regimes as intervening Variables. International Or ganization, 36(2), 185-205. Krause, G. (1999). A Two-Way Street: The Institutional Dynamics of the Modern Administrative State. Pittsburg, PA: University of Pittsburg Press. Krislov, S., Ehlermann, C. & Weiler, J. (1986) The political organs and the decision-making process in the United States and the Eur opean Community. In M. Cappelletti, M. Seccombe and J. Weiler (eds.). Integration through Law: Europe and the American Federal Experience (pp. 3). (Volume 1, book 2). Berlin: Walter de Gruyter. Lampinen, R. & Uusikyla, P. (1998). Implementati on deficit why member states do not comply with EU directives. Scandinavian Political Studies, 21(3), 231. Levy, R. (2000). Implementing European Uni on Public Policy. Cheltenham: Elgar. Linos, K. (2007). How Can International Organizations Shape Nati onal Welfare States? Evidence from Compliance with European Un ion Directives. Comparative Political Studies, 40(5), 547. Lipsky, M. (1980). Street-Level Bureaucracy: Th e Dilemmas of Indivi duals in the Public Service. Sage, New York. Lowi, T. J., (1972). Four Systems of Policy, Poli tics and Choice. Public Administration Review, 32(4), 298. Ludlow, Peter (1991). The Europ ean Commission. In R.O. Keohane & S. Hoffmann (eds.). The New European Community. D ecision making and Institut ional Change (pp. 85). Boulder, CO: Westview. Mair, P. (2004). The Europeanization dimension. Journal of European Public Policy, 11(2), 337 48. 257

PAGE 258

Majone, G. (2002). The European Commission: The Limits of Centralization and the Perils of Parliamentarization. Governance, 15(3), 375. March, J.G. & Olsen, J.P. (1989). Rediscovering In stitutions: The Orga nizational Basis of Politics. Free Press, New York. March, J.G. & Olsen, J.P. (1998). The Institutiona l Dynamics of International Political Orders. International Organiza tion, 52 (4), 943-969. Mastenbroek, E. (2003) Survivi ng the deadline: the transpositi on of EU directives in the Netherlands. European Union Politics, 4 (4), 371. Mastenbroek, E. (2005). EU Compliance: Still a Black Hole ? Journal of European Public Policy, 12(6), 1103. Mastenbroek, E. and Van Keulen, M. (2005). Beyond the goodness of fit. A preference based account of Europeanization. In Haverland M., & Holzhacker R. (eds). European Research Reloaded: Cooperation and Integration among Europeanized States. Kluwer: Deventer. Mastenbroek, E. & Kaeding, M. (2006). Europeanization Beyond the Goodness of Fit: Domestic Politics in the Forefront. Comparative European Politics, 4, 331-354. Mbaye, H. (2001). Why national states comp ly with supranational law. Explaining Implementation Infringements in the Eu ropean Union 1972-1993. European Union Politics, 2(3), 259. McCubbins, M. and Schwartz, T. (1984). Congressional Oversight Overlooked: Police Patrols Versus Fire Alarms. American Journal of Political Science, 28(1), 165-179. McCubbins, M., Noll, R. & Weingast, B. (1989). Structure and process, politics and policy: Administrative arrangements a nd the political control of agencies. Virginia Law Review, 75, 431-481. Mendrinou, M. (1996). Non-complia nce and the European Commissi ons role in integration. Journal of European Public Policy, 3(1), 1. Milner, H. (1988). Resisting Protectionism: Globa l Industries and the Politics of International Trade. Princeton: Princeton University Press Mitchell, R. (1996). Compliance Theory: An Over view. In J. Cameron, J. Werksman & P. Roderick (Eds.). Improving Compliance with International Environmental Law. London: Earthscan. Moravcsik, A. (1997), Taki ng Preferences Seriously: A Liberal Theory of International Politics. International Organiza tion, 51(4), 513-553. Nakamura, R. T. (1987), The Textbook Policy Process and Implementation Research. Policy Studies Review, 7(1), 142. 258

PAGE 259

Neyer, J. (2004). Explaining th e Unexpected: Efficiency and Effectiveness in European Decision-Making. Journal of Europe an Public Policy, 11(1), 19-38. Nugent, N. (2001), The European Co mmission, Palgrave Publishers. Olsen, J.P. (1996). Europeanization and nationstate dynamics. In S. Gustavsson & L. Lewin (Eds.). The Future of the Nation State: Es says on Cultural Pluralism and Political Integration (pp. 245). London: Routledge. Panke, D. (2006). Why Big States Cannot Do What They Want. International Courts and Compliance. Paper presented at the Annual Meeting of the La w and Society Association, 6 July, Baltimore, MD. Pappas, S.A. (1995). National Admi nistrative Procedures for the Preparation and Implementation of Community Decisions. European Institute of Public Administration, Maastricht. Perkins, R. & Neumayer, E. (2007). Implementi ng Multilateral Environmental Agreements: An Analysis of EU Directives. Global Environmental Politics, 7(3), 13-41. Pierson, P. (2000). Increasing Returns, Path Dependence, and the Study of Politics. American Political Science Review, 94(2), 251. Pinder, D. (1998), The New Europe: Economy, So ciety, and Environment. Chichester: John Wiley & Sons. Pressman, J. and Wildavsky A.. (1973). Implemen tation: How Great Expect ations in Washington are Dashed in Oakland: or, why it's amazing that federal programs work at all, this being the saga of the Economic Development Ad ministration as told by two sympathetic observers who seek to build morals on a founda tion of ruined hopes. Berkeley: University of California Press. Pridham, G. (1994). National Environmental Policy -making in the European Framework: Spain, Greece, and Italy in Comparison. Regi onal Politics and Policy, 4, 80-101. Puchala, D.J. (1975). Domestic politics a nd regional harmonization in the European Communities. World Politics, 27(4), 496. Putnam, R. & Bayne, N. (1987). Hanging Together Cambridge, MA: Harvard University Press. Putnam, R.D. (1988). Diplomacy and Domestic Politics. The Logic of Two-Level Games. International Organiza tion, 42 (2), 427-460. Putnam, R., Leonardi, R. & Nanetti, R. (1993). Making Democracy Work: Civic Traditions in Modern Italy. Princeton, NJ: Princeton University Press. Raustiala, K. & Slaughter, A. (2001). Internationa l Law, International Relations and Compliance. In W. Carlsnaes, T. Risse & B. Simmons (e ds.). Handbook of International Relations (pp. 538). London: Sage. 259

PAGE 260

Raustiala, K. (2000). Compliance & Effectiveness in International Regulatory Cooperation. Case Western Reserve Journal of In ternational Law, 32(3), 387. Richardson, J. (1996). Eroding EU Policies: Implementation Gaps, Cheating and Re-steering. In J. Richardson (ed.). European Union: Power and Policy-making (pp. 278). London: Routledge. Ripley, R. & Franklin, G. (1982). Bureaucrac y and Policy Implementation. Dorsey Press, Homewood, IL. Risse, T. & Sikkink, K. (1999). The Socialization of Internat ional Human Rights Norms into Domestic Practices: Introducti on. In T. Risse, S. C. R opp and K. Sikkink (Eds.). The Power of Human Rights. International Norms and Domestic Change. Cambridge: Cambridge University Press. Risse, T. (2000). Lets argue!: communicativ e action in world politics. International Organization, 54(1), 1. Risse, T. (ed.). (1995). Bringing Transnationa l Relations Back In. Cambridge: Cambridge University Press. Risse, T., Green Cowles, M. & Caporaso, J. ( 2001). Europeanization and Domestic Change: Introduction. In M. Cowles, J. Caporaso, & T. Risse (Eds.).Transforming Europe: Europeanization and Domestic Change (pp. 1). Cornell University Press, Ithaca. Sabatier, P.A. & Mazmanian, D.A. (1981). E ffective Policy Implementation. Lexington Books, Lexington. Sabatier, P. & Mazmanian, D. (1983). Can Regulation Work? The Implementation of the 1972 California Coastal Initiative. Plenum. Sbragia, A. (1993). The European Community : a balancing act. Publius, 23, 2338. Sbragia. A. (1996). The Push-Pull of Envir onmental Policy-Making. In H. Wallace & W. Wallace (eds.). Policy-Making in the European Union. 3rd edition, Oxford University Press. Scholz, J.T. & Wei, F.H. (1986). Regulatory En forcement in a Federalist System. American Political Science Review, 80(4), 1249-1270. Siedentopf, H. and Ziller, J. (eds.). (1988). Making European Policies Work: The Implementation of Community Legislati on in the Member States. London: Sage. Smith, M.P. (1997). The Commission made me do it: the European Commission as a strategic asset in domestic politics. In N. Nugent (ed.), At the Heart of the Union: Studies of the European Commission (pp. 167). Basingstoke: Macmillan. 260

PAGE 261

Steunenberg, B. (2004). EU policy, domestic inte rests, and the transpos ition of directives, Leiden: mimeo. Steunenberg, B. (2005). Turning swift policymaki ng into deadlock and delay: national policy coordination and the transposition of EU dir ectives. Paper prepared for the ECPR Joint Sessions, Granada, Spain, 14 April 2005. Steunenberg, B. (2006). Turning Swift Policy-Maki ng into Deadlock and Delay: National Policy Coordination and the Transposition of EU Direc tives. European Union Politics, 7(3), 293 319. Stewart, J., Anderson, J. & Taylor, Z. (1982) Presidential and c ongressional support for "Independent" regulatory commissions: Impli cations for the budgetary process. Western Political Quarterly, 85, 318-27. Sverdrup, U. (2004). Compliance and conflict ma nagement in the European Union: Nordic Exceptionalism. Scandinavian Po litical Studies, 27(1), 23. Tallberg, J. (2002). Paths to Compliance: Enforc ement, Management, and the European Union. International Organiza tion, 56(3), 609 Tallberg, J. (2003). European Governance and Supranational Institutions. Making States Comply. London: Routledge. Thelen, K. & Steinmo, S. (1992). Historical In stitutionalism in Comparative Politics. In S. Steinmo, K. Thelen, & F. Longstreth (E ds.). Structuring Po litics: Historical Institutionalism in Comparative Analysis (pp. 1). Cambridge University Press. Thompson, F. & Scicchitano, M.J. (1985). Stat e Implementation Effort and Federal Regulatory Policy: The case of Occupationa l Safety and Health. The Jour nal of Politics, 46(2), 686703. Thomson, R., Torenvlied, R. & Arregui, J. (2007 ). The Paradox of Compliance: Infringements and Delays in Transposing European Union Dire ctives. British Journal of Political Science, 37(4), 685. Treib, O. (2003). EU governance, misfit and the partisan logic of domestic adaptation: an actorcentered perspective on the transposition of EU directives. Paper presented at the EUSA 8th International Biennial Conference Nashville, Tennessee, 27 March 2003. Treib, Oliver (2008). Implemen ting and Complying with EU Governance Outputs. Living Reviews in European Govern ance 3(5), Retrieved from: http://www.livingreviews.org/lreg-2008. Trivedi, P. & Munkin, M. (2009). Recent Deve lopments in Cross Section and Panel Count Models, working paper, http://mypage.iu.edu/~trivedi/ 261

PAGE 262

Underdal, A. (1998). Explaining Compliance and De fection: Three Models. European Journal of International Relations, 4(1), 5. van den Bossche, P. (1996). In Search of Reme dies for Non-Compliance: The Experience of the European Community. Maastricht Journal of European and Comparative Law, 3(4), 371 398. van der Vleuten, A. (2005). Pincers and Presti ge: Explaining the Impl ementation of EU Gender Equality Legislation. Comparative European Politics, 3, 464488. van Meter, D. & van Horn, C. (1975). The Po licy Implementation Process: A Conceptual Framework. Administration & Society, 6(4), 445. Versluis, E. (2003) Explaining Variations in implementation of EU Directiv es. European Integration Online Papers, 8(19). Retrieved from http://eiop.or.at/ei op/texte/2004-019a.htm Versluis, E. (2007). Even Rules, Uneven Practic es: Opening the Black Box of EU Law in Action. West European Politics 30(1), 50. Victor, D., Raustiala, K. & Skolnikoff, E. (eds.). (1998). The Implementation and Effectiveness of International Environmental Commitments. Cambridge, MA: MIT Press Weale, A. (1999). European environmental policy by stealth. Environm ent and Planning C: Government and Policy, 17, 37-52. Weale, A., Pridham, G., Williams, A., & Porter, M. (1996). Environmental Administration in Six European States: Secular Convergence or Natio nal Distinctiveness? P ublic Administration, 74, 255-74. Weiler, J. (1991). The transformation of Europe. The Yale Law Journal, 100, 2403. Weingast, B. & Moran, M (1983). Bureaucratic discretion or congr essional contro l? Regulatory policymaking by the Federal Trade Commission. Journal of Political Economy, 91, 765800. Wilson, S. & Butler, D. (2007). A Lot More to Do: The Sensitivity of Time-Series Cross-Section Analyses to Simple Alternative Specif ications. Political Analysis, 15, 101-23 Winkelmann, R. (1997). Econometric Analysis of Count Data. Second edition. Springer-Verlag, New York. Wood, D.B. & Waterman, R. (1991). The Dynamics of Political Contro l of the Bureaucracy. American Political Science Review, 85(3), 801-28. Wood, D.B. (1988). Principals, Bureaucrats, and Responsiveness in Clean Air Enforcement. American Political Science Review, 82(1), 213-34. 262

PAGE 263

263 Wood, D.B. (1992), Modeling Federal Implemen tation as a System: The Clean Air Case. American Journal of Political Science, 36(1), 40-67. Wooldridge, J.M. (2001). Economet ric Analysis of Cross Secti on and Panel Data. MIT Press. Young, Oran (1992), The Effectiveness of International Institutions: Ha rd Cases and Critical Variables. In Rosenau & Erst-Otto Czempi el (Eds.). Governance without Government: Order and Change in World Politics (160 ) Cambridge: Cambridge University Press. Zurn, M. & Joerges, C. (eds.). (2005). Law a nd Governance in Postnati onal Europe: Compliance Beyond the Nation-State. Cambridge: Cambridge University Press.

PAGE 264

BIOGRAPHICAL SKETCH Ioannis Livanis was born in At hens, Hellas in 1976. After gr aduating from high school, his quest for knowledge made him take the national examinations for entrance into a Hellenic university. In 1996, he was admitted to the Department of Political Science and Public Administration at the University of Athens. After receiving his Bachelor of Arts with a double major in Political Science and Public Administration, with high honors in the fall of 2000, he entered the national examinations of the Hell enic Scholarship Foundati on. He tested first amongst his colleagues and was awarded a prestig ious Post Graduate Fellowship for pursuing MA/Ph.D. degrees at the United States (2001-2004). He was also offered an assistantship to pursue graduate studies in polic y and administration at the Un iversity of Florida (UF). In the fall of 2001, he enrolled in the Depart ment of Political Science at the University of Florida. He received his M.A. degree in Poli tical Science with a concentration in Public Affairs/Administration in May 2003. He continued on toward his doctoral degree in the Fall of 2003, in the same department, specializing in Policy and Administration, International Relations, and Comparative Politics. In the fall of 2005, he became a part of the Center for European Studies at the University of Florida, where he has taught severa l courses on European Politics. He fulfilled all the requirements and coursework, with a 3.8 overall G.P.A., and was awarded a Doctor of Philosophy degree in August 2010. 264