eu agencies : legal and political limits to the transformation of the eu administration

OX FOR D ST U DIE S I N EU ROPE A N L AWSeries Editors

PAUL CR AIGProfessor of English Law at St John’s College, Oxford

GR ÁINNE DE BÚRCAProfessor of Law at New York University School of Law

EU Agencies

Economic Governance in EuropeComparative Paradoxes, Constitutional

ChallengesFederico Fabbrini

Foreign Policy Objectives in European Constitutional Law

Joris LarikPrivate Regulation and the Internal MarketSports, Legal Services, and Standard Setting

in EU Economic LawMislav Mataija

The EU Deep Trade AgendaLaw and Policy

Billy A. Melo AraujoThe Human Rights of Migrants and Refugees

in European LawCathryn Costello

An Ever More Powerful Court?The Political Constraints of Legal Integration

in the European UnionDorte Sindbjerg Martinsen

The Concept of State Aid under EU LawFrom Internal Market to Competition

and BeyondJuan Jorge Piernas López

Justice in the EUThe Emergence of Transnational Solidarity

Floris de WitteThe Euro Area Crisis in Constitutional

PerspectiveAlicia Hinarejos

The European Fundamental FreedomsA Contextual Approach

Pedro Caro de Sousa

National Identity in EU LawElke Cloots

The Constitutional Foundations of European Contract Law

A Comparative AnalysisKathleen Gutman

The Criminalization of European Cartel Enforcement

Theoretical, Legal, and Practical Challenges

Peter WhelanFundamental Rights in Europe

Challenges and Transformations in Comparative Perspective

Federico FabbriniThe Principle of Loyalty in EU Law

Marcus KlamertConstitutional Pluralism in the EU

Klemen JaklicEU Consumer Law and

Human RightsIris Benöhr

The Principle of Mutual Recognition in EU Law

Christine JanssensThe Coherence of EU Free

Movement LawConstitutional Responsibility and

the Court of JusticeNiamh Nic Shuibhne

European Law and New Health Technologies

Edited by Mark Flear, Anne- Maree Farrell, Tamara Hervey, and Thérèse Murphy

OX FOR D ST U DIE S I N EU ROPE A N L AWSeries Editors

Paul Craig, Professor of English Law at St John’s College, Oxford andGráinne de Búrca, Professor of Law at New York University School of Law

The aim of this series is to publish important and original research on EU law. The focus is on scholarly monographs, with a particular emphasis on those which are interdisciplinary in nature. Edited collections of essays will also be included where they are appropriate. The series is wide in scope and aims to cover studies of particular areas of substantive and of institutional law, historical works, theoretical studies, and analyses of current debates, as well as questions of perennial interest such as the relationship between national and EU law and the novel forms of governance emerging in and beyond Europe. The fact that many of the works are interdisciplinary will make the series of interest to all those concerned with the governance and operation of the EU.

Ot h er t i t l e s in t h is ser i e s

1

EU AgenciesLegal and Political Limits to the Transformation of

the EU Administration

M ER IJ N CH A MON

1Great Clarendon Street, Oxford, OX2 6DP,

United Kingdom

Oxford University Press is a department of the University of Oxford.It furthers the University’s objective of excellence in research, scholarship,

and education by publishing worldwide. Oxford is a registered trade mark ofOxford University Press in the UK and in certain other countries

© M. Chamon 2016

The moral rights of the author have been asserted

First Edition published in 2016Impression: 1

All rights reserved. No part of this publication may be reproduced, stored ina retrieval system, or transmitted, in any form or by any means, without the

prior permission in writing of Oxford University Press, or as expressly permittedby law, by licence or under terms agreed with the appropriate reprographics

rights organization. Enquiries concerning reproduction outside the scope of theabove should be sent to the Rights Department, Oxford University Press, at the

address above

You must not circulate this work in any other formand you must impose this same condition on any acquirer

Crown copyright material is reproduced under Class LicenceNumber C01P0000148 with the permission of OPSI

and the Queen’s Printer for Scotland

Published in the United States of America by Oxford University Press198 Madison Avenue, New York, NY 10016, United States of America

British Library Cataloguing in Publication DataData available

Library of Congress Control Number: 2016934696

ISBN 978– 0– 19– 878448– 7

Printed and bound byCPI Group (UK) Ltd, Croydon, CR0 4YY

Links to third party websites are provided by Oxford in good faith andfor information only. Oxford disclaims any responsibility for the materials

contained in any third party website referenced in this work.

The present book is the abridged version of my doctoral thesis. As a result, I would first like to thank my supervisor, Professor Govaere, for her

guidance and trust in me. I would also like to thank the other members of my jury, Professors Everson, Lenaerts, Türk, Van Elsuwege, and

Lannon, for their observations and questions on the thesis, from which the present book has evidently benefited. I would also like to thank the people of OUP for the friendly and professional cooperation leading to the publication of this volume. Finally, for their support, I would like to

thank my parents.

Series Editors’ Preface

This books aims to explore the striking growth and spread of agencies within the EU, and to examine the institutional and constitutional implications of this phe-nomenon for the EU more generally.

Merijn Chamon sets out to provide a thoroughly comprehensive account of the field of EU agencies, and to define and classify the array of different types of agencies in existence, given the absence of a single formal model of EU agency and in view of their ad hoc development. He then proceeds to examine the driv-ers of this development as well as the political and legal limits to their spread. As the book notes, there is now almost no area of EU law and policy without an agency which is responsible for much of the governance and administration of that area.

The author argues that the political limits to the process of agency- creation and spread are related to the origins of the agencies, most of which emerged from some existing committee or bureaucratic structure already in existence at EU level, and were driven by the political interests of the relevant actors rather than by a purely functional logic. He suggests ultimately that the process of ‘agencification’ (to use the rather ugly term which is increasingly used to de-scribe the process of agency spread in the EU) is unlikely to develop in ways which the Commission, the Council, or the Parliament considers to be contrary to its interests.

On the other hand, the author argues that the legal limits to the establishment of EU agencies, and the legal principles which should inform their establishment, are to be found in the treaty provisions: in the principle of conferred powers, as well as in the principles of subsidiarity, proportionality, and— despite the diffi-culty in defining this as a legal principle— institutional balance. The book also contains an extensive discussion of the famous Meroni doctrine on delegation and institutional balance, and its relevance for EU agencies today.

A comparison with the legal and constitutional controls over the establishment and functioning of agencies in Germany and the USA not only sheds some light on the importance of the kind of federal system in question, but also highlights differences between the kind of agencies which have been created in the EU and elsewhere. Ultimately, the author concludes that while agencies in the EU rest still on a somewhat uncertain legal and treaty foundation, their legitimacy resides in the broad political support they enjoy. He proposes an amendment to the treaties that would more securely anchor existing and future EU agencies within the EU institutional and constitutional framework.

At a time when the EU’s legitimacy is under constant question and scrutiny, and when its political and legal structures are contested and under strain, a thor-ough and comprehensive inquiry into the phenomenon of agency growth in

Series Editors’ Prefaceviii

the EU is welcome and timely. Chamon’s thorough and careful analysis of this phenomenon and his proposals for a better anchoring of agencies under EU law should be of interest to legal scholars and political scientists alike, and to all those concerned with EU administration.

Paul Craig and Gráinne de Búrca

Table of Contents

List of Figures xiiiTable of Cases xvTable of Legislation xxiiiList of Abbreviations xxvii

Introduction 1

I. Setting the Scene: EU Agencies, Agencification, and the EU Administration 31 EU Agencies and the Agency Phenomenon 3

1.1 Definition 51.2 EU agencies 151.3 Agencification 451.4 The EU administration 46

2 Situating the EU Agencies in the EU Administration 472.1 Executive federalism in the EU 482.2 EU agencies under a functional and procedural perspective 502.3 Conclusion 51

II. In Search of an Agency Model: The Provisions in Agencies’ Establishing Acts 521 Introduction 522 The EU’s Sources of Legitimacy 53

2.1 The EU’s democratic deficit 532.2 Legitimacy of the EU administration 61

3 The EU Agencies’ Statutes 643.1 The Boards 653.2 The Directors 763.3 Work programmes and annual reports 813.4 External relations 863.5 Seat Agreement 933.6 The Common Approach’s contribution to an agency model 97

4 Practice Following the Common Approach 984.1 Review of agencies’ acts following the Common Approach 99

5 Conclusion 100

III. The Political Limits to Agencification 1021 The Origins of Agencies 1022 Explaining Agencification 106

2.1 The functional accounts of agencification 1072.2 Non- functional accounts of agencification 1092.3 Concluding remarks 113

Table of Contentsx

3 Agencification and Its Effects 1143.1 EU agencies: Trojan horses, part of executive centre formation,

or part of a multi- layered administration? 1153.2 Concluding remarks 117

4 Identifying the Political Limits 1184.1 The Council 1194.2 The Commission 1234.3 The European Parliament 1264.4 The role of the Court of Justice 1314.5 Concluding remarks 131

5 Conclusion on the Political Limits 132

IV. The Legal Limits to Agencification 1341 Limits Flowing from the Principle of Conferred Powers 136

1.1 The competence to establish a new body within the EU’s institutional architecture 136

1.2 The choice of instrument 1531.3 Conclusion 157

2 Limits Flowing from the Principle of Subsidiarity 1572.1 The agency instrument under the subsidiarity principle 1592.2 Individual tasks and powers of EU agencies under

the subsidiarity principle 1622.3 Conclusion 164

3 Limits Flowing from the Principle of Proportionality 1643.1 EU agencies’ tasks and powers under the principle of proportionality 1663.2 EU agencies’ establishment under the principle of proportionality 1673.3 The proportionality of agencification in legislative practice 1713.4 Conclusion 173

4 Concluding Remarks: Limits Flowing from the General Principles of EU Law 174

5 The Meroni Jurisprudence and the Notion of Delegation 1755.1 The original Meroni ruling 1765.2 Contemporary reception of the Meroni ruling 1835.3 Present- day reception and valuation of the Meroni ruling 1855.4 EU agencies under Meroni: Does it make sense? 1925.5 Meroni in the subsequent case law of the Court 2125.6 Assessment of Meroni and deducing a doctrine 2255.7 Delegating powers in the EU legal order 2305.8 EU agencies and an EU Meroni doctrine 2425.9 Conclusion 248

6 A Prelude to the Institutional Balance: Romano 2496.1 The Romano case 2506.2 Conclusion 258

7 Limits Flowing from the Institutional Balance 2587.1 Separation of powers 2597.2 The EU’s institutional balance 2687.3 Conclusion 296

8 Conclusion on the Legal Limits 297

Table of Contents xi

V. Controlling Agencification 2991 Agencification in the United States of America 299

1.1 Designing the executive sphere 3001.2 Designing independent agencies vs. the President’s control

over the executive 3001.3 Control over which powers? 3011.4 Interpretation of agencies’ statutory mandates 3021.5 US agencies’ constitutionality 304

2 Agencification in the Federal Republic of Germany 3062.1 German agencies 3072.2 The relation between Bund and Länder 3072.3 German agencies and the separation of powers 3102.4 Deference to agencies’ decisions 313

3 Agencification in the US and Germany 3154 Controlling EU Agencies 316

4.1 Controlling agencification 3164.2 Controlling individual agencies 317

5 Concluding Remarks on the Political Control of Agencification 3266 Judicial Scrutiny of EU Agencies’ Acts 327

6.1 Passive locus standi of EU agencies 3286.2 Active locus standi of EU agencies 3626.3 Conclusion 367

VI. Conclusions and Proposal 3691 The EU Agency Instrument— Conclusions 3692 The EU Agency Instrument— Proposal 372

2.1 Proposals in the IGCs and the Convention 3732.2 A legal basis in the Treaties 376

Selected Bibliography 383Index 387

List of Figures

1.1 Functional classifications of EU agencies 211.2 A functional typology of EU agencies 241.3 Hierarchy of agencies’ operational powers 261.4 Hierarchy of agencies’ (non- )decision- making powers 271.5 Number of EU agencies and number of staff 461.6 Number of EU agencies and budget size 46

Table of Cases

DECISIONS OF THE EU COURTS A ND OPINIONS OF THE A DVOC ATE S GENER A L

Opinion of AG Roemer in Cases 9/ 56 & 10/ 56, Meroni & Co. v. High Authority, [1957–1958] ECR 177 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175, 177, 179, 199, 230, 252

Case 9/ 56, Meroni & Co. v. High Authority, [1957–1958] ECR 133 . . . . . . . . . . . . . . . . . . 8, 9, 39, 68, 127, 175– 249, 250, 252– 6, 257– 8,

279– 81, 283, 293, 297, 301, 327, 335, 336, 350, 353, 359, 367, 370, 375, 376– 7, 381

Case 10/ 56, Meroni & Co. v. High Authority, [1957–1958] ECR 157 . . . . . . . . . . . . . . . . . . . . .177Joined Cases 32/ 58 & 33/ 58, SNUPAT v. High Authority, [1959] ECR 127 . . . .230, 331– 4, 356– 7Case 3/ 59, Germany v. High Authority, [1960] ECR 53 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .212Case 18/ 60, Louis Worms v. High Authority, [1962] ECR 195 . . . . . . . . . . . . . . . . . 356– 7, 361, 367Case 25/ 62, Plaumann & Co. v. Commission, [1963] ECR 95 . . . . . . . . . . . . . . . . . . . . . . . . . . . 361Opinion of AG Gand in Case 19/ 67, Van der Vecht, [1967] ECR 345 . . . . . . . . . . . . . . . . . . . . .253Case 19/ 67, Van der Vecht, [1967] ECR 345 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .253– 4Case 48/ 69, Imperial Chemical Industries Ltd. v. Commission, [1972] ECR 619 . . . . . . . . . . . . . .234Case 25/ 70, Köster, [1970] ECR 1161 . . . . . . . 194, 204, 206, 209, 221, 222, 225, 228, 233, 234,

237, 240, 241, 247, 248, 252, 275, 283, 285, 287, 381Case 46/ 72, De Greef v. Commission, [1973] ECR 543 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .137Case 131/ 73, Criminal proceedings against Giulio et Adriano Grosoli, [1973] ECR 1555 . . . . . . .234Opinion 1/ 76, re the draft Agreement establishing a European laying- up fund for

inland waterway vessels, [1977] ECR 741 . . . . . 200– 1, 206, 213– 17, 224, 228, 234– 5, 238– 9Case 5/ 77, Tedeschi, [1977] ECR 1555 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194Case 92/ 78, Simmenthal v. Commission, [1979] ECR 777 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .212Case 138/ 79, Roquette Frères v. Council, [1980] ECR 3333 . . . . . . . . . . . . . . . . . . . . .275, 280, 285Opinion of AG Warner in Case 98/ 80, Romano, [1981] ECR 1259 . . . . . . . . . . . . . . . . . 251– 3, 257Case 98/ 80, Romano, [1981] ECR 1241 . . . . . . . . . . 146, 196, 207, 209, 211, 235, 240– 1, 249– 58,

275, 289, 293, 297– 8, 327, 370, 377Joined cases 188 to 190/ 80, France, Italy and UK v. Commission, [1982] ECR 2545 . . . . . . . . . .263Case 16/ 81, Alaimo v. Commission, [1982] ECR 1559 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12Joined Cases 213 to 215/ 81, Norddeutsches Vieh- und Fleischkontor Herbert Will,

[1982] ECR 3583 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .234Opinion of AG Rozès in Case 302/ 81, Eggers, [1982] ECR 3443 . . . . . . . . . . . . . . . . . . . . . . . .256Case 13/ 83, Parliament v. Council, [1985] ECR 1513 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .288Case 294/ 83, Parti écologiste ‘Les Verts’ v. Parliament,

[1986] ECR 1339 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252, 263, 288, 328– 30, 362, 367Joined Cases 279/ 84, 280/ 84, 285/ 84 & 286/ 84, Walter Rau

Lebensmittelwerke e.a. v. Commission, [1987] ECR 1069 . . . . . . . . . . . . . . . . . . . . . . . . . . .223Case 5/ 85, AKZO Chemie v. Commission, [1986] ECR 2585 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217Case 20/ 85, Roviello, [1988] ECR 2805 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .288Case 149/ 85, Wybot, [1986] ECR 2391 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .275, 285Joined Cases 281, 283– 5, 287/ 85, Germany e.a. v. Commission, [1987] ECR 3203 . . . . . . . . . . . 141Case 415/ 85, Commission v. Ireland, [1988] ECR 3097 . . . . . . . . . . . . . . . . . . . . . . . . . . . .275, 285Case 68/ 86, UK v. Council, [1988] ECR 855 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .282Case 204/ 86, Greece v. Council, [1988] ECR 5323 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .280Case 21/ 87, Borowitz, [1988] ECR 3715 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

Table of Casesxvi

Opinion of AG Lenz in Case 236/ 87, Bergemann, [1988] ECR 5125 . . . . . . . . . . . . . . . . . . . . . 255Opinion of AG Darmon in Case 302/ 87, Parliament v. Council, [1988] ECR 5615 . . . . . . . . . .288Case 302/ 87, Parliament v. Council, [1988] ECR 5615 . . . . . . . . . . . . . . . . . . . . . . . .273, 275, 362Case 22/ 88, Vreugdenhil BV, [1989] ECR 2049 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .279Opinion of AG Van Gerven in Case C- 70/ 88, Parliament v. Council,

[1990] ECR I- 2041 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202, 271, 274Case 70/ 88, Parliament v. Council, [1990] ECR I- 2041 . . . . . . . . . . 202, 269, 270– 1, 273– 4, 279,

281, 283, 284– 6, 288, 290, 327, 329, 362, 363Case C- 322/ 88, Grimaldi, [1989] ECR 4407 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .294Case 353/ 88, Briantex Sas and Antonio Di Domenico v. EEC and

Commission, [1989] ECR 3623 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357Case C- 370/ 89, SGEEM and Etroy v. EIB, [1992] ECR I- 6211 . . . . . . . . . . . 9, 355, 357, 361, 364Joined cases T- 68/ 89, T- 77/ 89 & T- 78/ 89, Società Italiana Vetro SpA,

e.a. v. Commission, [1992] ECR II- 1403 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .275, 285Case C- 240/ 90, Germany v. Commission, [1992] ECR I- 5383 . . . . . . . . . . . . . . . . . . .223, 234, 283Opinion of AG Jacobs in case C- 269/ 90, Technische Universität München,

[1991] ECR I- 5469 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315Opinion of AG Darmon in Case C- 282/ 90, Vreugdenhil BV v. Commission,

[1992] ECR I- 1937. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .279Case C- 282/ 90, Vreugdenhil BV v. Commission, [1992] ECR I- 1937 . . . . . . . . . . . . . .279, 280, 283Opinion of AG Gulmann in Joined Cases C- 15/ 91 & C- 108/ 91,

Buckl e.a. v. Commission, [1992] ECR I- 6074 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353Case C- 102/ 91, Knoch, [1992] ECR I- 4341 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255Case C- 201/ 91, Grisvard and Kreitz, [1992] ECR I- 5009 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255Case C- 316/ 91, Parliament v. Council, [1994] ECR I- 625 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .268Opinion 1/ 92, re the draft agreement relating to the creation of the European

Economic Area, [1992] ECR I- 2821 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .239Opinion of AG Darmon in Case C- 41/ 92, The Liberal Democrats v.

Parliament, [1993] ECR I- 3153 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353Case C- 91/ 92, Faccini Dori, [1994] ECR I- 3325 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154Opinion of AG Van Gerven in Case C- 137/ 92 P, Commission v.

BASF AG e.a., [1994] ECR I- 2555 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217, 229, 242, 359Case C- 359/ 92, Germany v. Council, [1994] ECR I- 3681 . . . . . . . . . . . . . . . . . . . . . . . 146– 8, 299Joined Cases C- 46/ 93 & C- 48/ 93, Brasserie du Pêcheur, [1996] ECR I- 1029 . . . . . . . . . . . . . . .283Case C- 65/ 93, Parliament v. Council, [1995] ECR I- 643 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .280Opinion 2/ 94, re Accession by the Community to the ECHR, [1996] ECR I- 1759 . . . . . . . . . . . . .138Case C- 25/ 94, Commission v. Council, [1996] ECR I- 1469 . . . . . . . . . . . . . . . . . . . . . . . . . . . .62– 3Case C- 58/ 94, Netherlands v. Council, [1996] ECR I- 2169 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .283Case T- 243/ 94, British Steel plc v. Commission, [1997] ECR II- 1887 . . . . . . . . . . . . . .213, 280, 283Joined Cases T- 369/ 94 & T- 85/ 95, DIR International Film

e.a. v. Commission, [1998] ECR II- 357 . . . . . . . . . . . . . . . . . . . . . . . . . . . 218– 19, 237, 331– 4Case C- 191/ 95, Commission v. Germany, [1998] ECR I- 5449 . . . . . . . . . . . . . . . . . . . . . . . . . . .229Joined Opinion of AG Fennelly in Case C- 286/ 95 P and Joined Cases C- 287/ 95

P & C- 288/ 95 P, Commission v. ICI and Commission v. Solvay SA [2000] ECR I- 2341 . . . . 230Joined Cases T- 132/ 96 & T- 143/ 96, Freistaat Sachsen, Volkswagen AG and

Volkswagen Sachsen GmbH v. Commission, [1999] ECR II- 3663 . . . . . . . . . . . . . . . . . . . . .280Case C- 95/ 97, Région wallonne v. Commission, [1997] ECR I- 1787 . . . . . . . . . . . . . . . . . . . . . . .278Case C- 180/ 97, Regione Toscana v. Commission, [1997] ECR I- 5245 . . . . . . . . . . . . . . . . . . . . . .278Case C- 202/ 97, Fitzwilliam, [2000] ECR I- 883 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255Case T- 148/ 97, Keeling v. OHIM, [1998] II- 2217 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .329Case T- 277/ 97, Ismeri Europa Srl v. Court of Auditors, [1999] ECR II- 1825 . . . . . . . . . . . . . . . . 217Opinion of AG Léger in Case C- 452/ 98, Nederlandse Antillen v. Council,

[2001] ECR I- 8853 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .281Case C- 452/ 98, Nederlandse Antillen v. Council, [2001] ECR I- 8973 . . . . . . . . . . . . . . . . . . . . .281

Table of Cases xvii

Case T- 163/ 98, The Procter & Gamble Company v. OHIM, [1999] ECR II- 2383 . . . . . . . . . . . .338Case T- 13/ 99, Pfizer Animal Health SA v. Council, [2002] ECR II- 3318 . . . . . . . . . . . . . . . . . . . .32Case T- 120/ 99, Kik v. OHIM, [2001] ECR II- 2235 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28Opinion of AG Ruiz- Jarabo Colomer in Case C- 315/ 99 P,

Ismeri Europa Srl v. Court of Auditors, [2001] ECR I- 5281 . . . . . . . . . . . . . . . . .218, 362, 364Case C- 315/ 99 P, Ismeri Europa Srl v. Court of Auditors, [2001] ECR I- 5281 . . . . . . . . . . . . . . . 218Case T- 326/ 99, Olivieri v. Commission and EMEA, [2003] ECR II- 6053 . . . . . . . . . . . . . . . . . . 353Case T- 331/ 99, Mitsubishi HiTec Paper Bielefeld GmbH v. OHIM,

[2001] ECR II- 433 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .344Opinion 1/ 00, re the proposed agreement on the establishment of a

European Common Aviation Area, [2002] ECR I- 3493 . . . . . . . . . . . . . . . . . . . . . . . . . . . .239Opinion 2/ 00, re the Cartagena Protocol, [2001] ECR I- 9713 . . . . . . . . . . . . . . . . . . . . . . . . . . . 149Case C- 15/ 00, Commission v. EIB, [2003] ECR I- 7281 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .329Case C- 50/ 00 P, Unión de Pequeños Agricultores v. Council, [2002] ECR I- 6677 . . . . . . . . . . . .327Case C- 345/ 00 P, FNAB, Setrab & Est Distribution Biogam SARL v. Council,

[2001] ECR I- 3811 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131, 281Opinion of AG Geelhoed in Case C- 378/ 00, Commission v. Parliament and

Council, [2003] ECR I- 937 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218Case C- 378/ 00, Commission v. Parliament and Council, [2003] ECR I- 937 . . . . . . . . . . . . . . . .239Case T- 17/ 00 R, Rothley e.a. v. Parliament, [2000] ECR II- 2085 . . . . . . . . . . . . . . . . . . . . . . . . 219Case T- 17/ 00, Rothley e.a. v. Parliament, [2002] ECR II- 579 . . . . . . . . . . . . . . . . . . . . . . . . . . . 219Case T- 74/ 00, Artegodan v. Commission [2000] ECR II- 2583 . . . . . . . . . . . . . . . . . . . . . . . . . . . 353Case T- 106/ 00, Streamserve Inc. v. OHIM, [2002] ECR II- 723 . . . . . . . . . . . . . . . . . . . . . . . . .344Case T- 123/ 00, Dr Karl Thomae GmbH v. Commission, [2002] ECR II- 5193 . . . . . . . . . 131, 331– 3Case T- 209/ 00, Lamberts v. Ombudsman, [2002] ECR II- 765 . . . . . . . . . . . . . . . . . . . . . . . . . . 355Case T- 319/ 00, Borremans e.a. v. Commission, [2002] ECR II- 905 . . . . . . . . . . . . . . . . . . . . . . . . 18Case T- 353/ 00 R, Le Pen v. Parliament, [2001] ECR II- 125 . . . . . . . . . . . . . . . . . . . . . . . . . . . .348Joined Cases T- 377/ 00, T- 379/ 00, T- 380/ 00, T- 260/ 01 & T- 272/ 01,

Philip Morris International e.a. v. Commission, [2003] ECR II- 1 . . . . . . . . . . . . . . . . 277, 281Case C- 239/ 01, Germany v. Commission, [2003] ECR I- 10333 . . . . . . . . . . . . . . . . . . . . . . . . . .283Case C- 416/ 01, ACOR, [2003] ECR I- 14083 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .234Opinion of AG Geelhoed in Case C- 491/ 01, British American Tobacco

(Investments) Ltd, [2002] ECR I- 11453 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .282Case C- 491/ 01, British American Tobacco (Investments) Ltd, [2002] ECR I- 11453 . . . . . . . . . . .282Case T- 63/ 01, The Procter & Gamble Company v. OHIM, [2002] ECR II- 5255 . . . . . . . . . . . . .339Joined Cases T- 64/ 01 & T- 65/ 01, Afrikanische Frucht- Compagnie

GmbH v. Council and Commission, [2004] ECR II- 521 . . . . . . . . . . . . . . . . . . . . . . . .223, 279Case T- 77/ 01, Diputación Foral de Álava e.a. v. Commission,

[2002] ECR II- 81 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .334Opinion of AG Stix- Hackl in Case C- 118/ 02, Industrias de

Deshidratación Agrícola SA, [2004] ECR I- 3073 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .235Case C- 134/ 02 P, Ombudsman v. Lamberts, [2004] ECR I- 2803 . . . . . . . . . . . . . . . . . . . . 274, 355Case C- 167/ 02 P, Rothley e.a. v. Parliament, [2004] ECR I- 3149 . . . . . . . . . . . . . . . . . . . . . . . . 219Case C- 233/ 02, France v. Commission, [2004] ECR I- 2759 . . . . . . . . . . . . . 276, 284, 285, 286– 8Opinion of AG Léger in Case C- 301/ 02 P, Tralli v. ECB,

[2005] ECR I- 4071 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189, 221– 2Case C- 301/ 02 P, Tralli v. ECB, [2005] ECR I- 4071 . . . . 175, 193, 221– 5, 235, 245, 248, 280, 286Joined Cases T- 124/ 02 & T- 156/ 02, The Sunrider Corp. v. OHIM,

[2004] ECR II- 1149 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .342Case T- 273/ 02, Krüger GmbH & Co. KG v. OHIM, [2005] ECR II- 1271 . . . . . . . . . . . . . . . . . .339Joined Cases C- 154/ 04 & C- 155/ 04, Alliance for Natural Health, [2005]

ECR I- 6451 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175, 193, 220, 222– 5, 235, 248Opinion of AG Léger in Case C- 40/ 03 P, Rica Foods NV v. Commission, [2005]

ECR I- 6814 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .205

Table of Casesxviii

Case C- 94/ 03, Commission v. Council, [2006] ECR I- 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .286Case C- 131/ 03 P, Reynold Tobacco e.a. v. Commission, [2006] ECR I- 7795 . . . . . . . . . . . . 277, 281Opinion of AG Poiares Maduro in Case C- 160/ 03, Spain v. Eurojust, [2005]

ECR I- 2077 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .330Case C- 160/ 03, Spain v. Eurojust, [2005] ECR I- 2077 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .330Opinion of AG Stix- Hackl in Case C- 436/ 03, Parliament v. Council,

[2006] ECR I- 3733 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143Case C- 436/ 03, Parliament v. Council, [2006] ECR I- 3733 . . . . . . . . . . . . . . . . . . . . . . . . 143, 148Case C- 533/ 03, Commission v. Council, [2006] ECR I- 1025 . . . . . . . . . . . . . . . . . . . . . . . . . . . .275Case T- 123/ 03, Pfizer Ltd. v. Commission, [2004] ECR II- 1631 . . . . . . . . . . . . . . . . . . . . . . . . .354Case T- 133/ 03, Schering- Plough Ltd v. Commission and EMEA,

ECLI:EU:T:2007:365 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .330, 333Case T- 346/ 03, Krikorian e.a. v. Parliament, Council and Commission,

[2003] ECR II- 6037 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355Case T- 369/ 03 R, Arizona Chemical BV e.a. v. Commission, [2004] ECR II- 205 . . . . . . . . . . . . 212Case C- 18/ 04 P, Krikorian e.a. v. Parliament, Council and Commission,

ECLI:EU:C:2004:691 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355Case C- 66/ 04, UK v. Parliament and Council, [2005] ECR I- 10553. . . . . . . . . . . . . . . . . . . . . .148Opinion of AG Kokott in Case C- 217/ 04, UK v. Parliament and Council,

[2006] ECR I- 3771 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .143– 4, 168Case C- 217/ 04, UK v. Parliament and Council, [2006] ECR I- 3771 . . . . . . . 121, 131, 142, 143– 6,

148, 152– 3, 166, 174, 370, 378– 80Case C- 310/ 04, Spain v. Council, [2006] ECR I- 7285 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170Order of the Court in Case C- 317/ 04, Parliament v. Council, [2005] ECR I- 2457 . . . . . . . . . . .366Opinion of AG Mengozzi in Case C- 354/ 04 P, Gestoras Pro Amnistía

e.a. v. Council, [2007] ECR I- 1579 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274Case C- 417/ 04 P, Regione Siciliana v. Commission, [2006] ECR I- 3881 . . . . . . . . . . . . . . . . . . .278Case T- 49/ 04, Hassan v. Council and Commission, [2006] ECR II- 52 . . . . . . . . . . . . . . . . . . . . .333Case T- 474/ 04, Pergan Hilfsstoffe v. Commission, [2007] ECR II- 4225 . . . . . . . . . . . . 276, 285, 329Opinion of AG Sharpston in Case C- 29/ 05 P, OHIM v. Kaul GmbH,

[2007] ECR I- 2213 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .338Case C- 29/ 05 P, OHIM v. Kaul GmbH, [2007] ECR I- 2213 . . . . . . . . . . . . . . . . . . . . . . . . . . .338Case C- 77/ 05, UK v. Council, [2007] ECR I- 11459 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .67Case C- 214/ 05 P, Sergio Rossi SpA v. OHIM, [2006] ECR I- 7057 . . . . . . . . . . . . . . . . . . . . . . . .344Case T- 69/ 05, Evropaïki Dynamiki v. EFSA, ECLI:EU:T:2007:314 . . . . . . . . . . . . . . . . . . . . . .330Case T- 70/ 05, Evropaïki Dynamiki v. EMSA, [2010] ECR II- 313 . . . . . . . . . . . . . . . . . . . .330, 337Case C- 38/ 06, Commission v. Portugal, [2010] ECR I- 1569 . . . . . . . . . . . . . . . . . . . . . . . . . . . .277Case C- 55/ 06, Arcor AG & Co. KG, [2008] ECR I- 2931 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .313– 4Joined cases C- 120/ 06 P and C- 121/ 06 P, FIAMM and Fedon v. Council

and Commission, [2008] ECR I- 6513. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .274– 5Opinion of AG Poiares Maduro in Case C- 133/ 06, Parliament v. Council,

[2008] ECR I- 3189 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .280Case C- 133/ 06, Parliament v. Council, [2008] ECR I- 3189 . . . . . . . . . . . . . . . . . . . . 265, 277, 285Opinion of AG Poiares Maduro in Case C- 411/ 06,

Commission v. Parliament and Council, [2009] ECR I- 7585 . . . . . . . . . . . . . . . . . . . . . . . .278Case T- 63/ 06, Evropaïki Dynamiki v. EMCDDA, [2010] ECR II- 177 . . . . . . . . . . . . . . . . . . . .337Case T- 107/ 06, Inet Hellas v. Commission, [2009] ECR II- 4591 . . . . . . . . . 220, 224, 247, 251, 332Case T- 165/ 06, Fiorucci v. OHIM, [2009] ECR II- 1375 . . . . . . . . . . . . . . . . . . . . . . .276, 284, 285Case T- 187/ 06, Schräder v. CPVO, [2008] ECR II- 3151 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .206Case T- 258/ 06, Germany v. Commission, [2010] ECR II- 2027 . . . . . . . . . . . . . . . . . . . . . . . . . .348Case T- 311/ 06, FMC Chemical and Arysta Lifesciences v. EFSA, [2008] ECR II- 88 . . . . . . 219, 330Case T- 312/ 06, FMC Chemical v. EFSA, [2008] ECR II- 89 . . . . . . . . . . . . . . . . . . . . . . . . . . . .330Case T- 397/ 06, Dow AgroSciences v. EFSA, [2008] ECR II- 90 . . . . . . . . . . . . . . . . . . . . . . . . . .330

Table of Cases xix

Case T- 411/ 06, Sogelma v. EAR, [2008] ECR II- 2771 . . . . . . . . . . . . . . . . . . . . .329– 30, 333, 367Case C- 424/ 07, Commission v. Germany, [2009] ECR I- 11431 . . . . . . . . . . . . . . . . . . . . . . 312, 315Case C- 518/ 07, Commission v. Germany, [2010] ECR I- 1885 . . . . . . . . . . . . . . . . . . . . . . . . . . . 312Joined Cases T- 252/ 07, 271/ 07 & 272/ 07,

Sungro, SA e.a. v. Council and Commission, [2010] ECR II- 55 . . . . . . . . . . . . . . . . . . . . . . 170Case T- 264/ 07, CSL Behring v. Commission and EMA, [2010] ECR II- 4469 . . . . . . . . . . . . . . .333Case C- 58/ 08, Vodafone Ltd e.a., [2010] ECR I- 4999 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .170– 1Opinion of AG Trstenjak in Case C- 101/ 08, Audiolux e.a., [2009]

ECR I- 9823 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .259, 268, 269Case T- 135/ 08, Schniga GmbH v. CPVO, [2010] ECR II- 5089 . . . . . . . . . . . . . . . . . .276, 285, 363Case T- 196/ 08, Srinivasan v. Ombudsman, [2009] ECLI:EU:T:2008:470 . . . . . . . . . . . . . . . . .329Case T- 228/ 08, Szomborg v. Commission, [2009] ECR II- 224 . . . . . . . . . . . . . . . . . . . . . . . . . . . 353Case T- 539/ 08, Etimine SA and Ab Etiproducts Oy v. Commission,

[2010] ECR II- 4017 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .330Case C- 279/ 09, DEB Deutsche Energiehandels- und Beratungsgesellschaft

GmbH, [2010] ECR I- 13849 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .263Case T- 36/ 09, dm- drogerie markt GmbH & Co. KG v. OHIM, [2011] ECR II- 6079 . . . . . . . . . .341Case T- 52/ 09, Nycomed Danmark v. EMA, [2011] ECR II- 8133 . . . . . . . . . . . . . . . . . . . . . . . . .337Case T- 292/ 09, Mugraby v. Council and Commission, [2011] ECR II- 255 . . . . . . . . . . . . . . . . . . 353Case C- 137/ 10, Communautés européennes v. Région de Bruxelles- Capitale,

[2011] ECR I- 3515 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .236Case T- 1/ 10, SNF SAS v. ECHA, [2011] ECR II- 6576 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .354Case T- 93/ 10, Bilbaína de Alquitranes e.a. v. ECHA, ECLI:EU:T:2013:106 . . . . . . . . . . . . . . . .349Case T- 94/ 10, Rütgers Germany e.a. v. ECHA, ECLI:EU:T:2013:107 . . . . . . . . . . . . . . . . . 349, 354Case T- 95/ 10, Cindu Chemicals e.a. v. ECHA, ECLI:EU:T:2013:108 . . . . . . . . . . . . . . . . . . . . .349Case T- 96/ 10, Rütgers Germany GmbH e.a. v. ECHA, ECLI:EU:T:2013:109 . . . . . . .206, 349, 350Case T- 262/ 10, Microban v. Commission, [2011] ECR II- 7697 . . . . . . . . . . . . . . . . . . . . . . . . . .350Opinion of AG Bot in Case C- 221/ 10 P, Artegodan GmbH v. Commission,

ECLI:EU:C:2011:744 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195, 279Case C- 221/ 10 P, Artegodan GmbH v. Commission, ECLI:EU:C:2012:216 . . . . . . . . . . . . . . . . .279Case C- 77/ 11, Council v. Parliament, ECLI:EU:C:2013:559 . . . . . . . . . . . . . . . . . . . . . . . . . . . .275Case C- 199/ 11, Europese Gemeenschap v. Otis NV e.a., ECLI:EU:C:2012:684 . . . . . . . . . . . . . .236Case T- 496/ 11, UK v. ECB, ECLI:EU:T:2015:133 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .345Case C- 370/ 12, Pringle, ECLI:EU:C:2012:756 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118Opinion of AG Jääskinen in Case C- 270/ 12, UK v. Parliament and Council,

ECLI:EU:C:2013:562 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146– 8, 243– 5, 257, 290, 292– 3Case C- 270/ 12, UK v. Parliament and Council, ECLI:EU:C:2014:18 . . . . . . . . . . 1, 28, 38, 39, 61,

112, 118, 124– 6, 131, 146– 9, 155, 176, 187, 194, 196– 9, 201, 204, 206, 207, 210, 221, 236, 241, 242– 9, 256– 8, 285, 289,

293– 8, 312, 319, 326, 350, 359– 60, 367, 370– 1, 379Opinion of AG Cruz Villalón in Case C- 427/ 12, Commission v. Parliament

and Council, ECLI:EU:C:2013:871 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .206, 235Case C- 427/ 12, Commission v. Parliament and Council,

ECLI:EU:C:2014:170 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287, 289, 295Opinion 2/ 13, re Accession to the ECHR, ECLI:EU:C:2014:2454 . . . . . . . . . . . . . . . . . . . . . . . .239Case C- 65/ 13, Parliament v. Commission, ECLI:EU:C:2014:2289 . . . . . . . . . . . . . . . . . . . . . . .287Case C- 146/ 13, Spain v. Parliament and Council, ECLI:EU:C:2015:298 . . . . . . . . . . . . . .224, 295Case C- 147/ 13, Spain v. Council, ECLI:EU:C:2015:299 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .224Case C- 409/ 13, Council v. Commission, ECLI:EU:C:2015:217 . . . . . . . . . . . . . . . . . . . . . . . . . . 317Opinion of AG Jääskinen in Case C- 507/ 13, UK v. Parliament and Council,

ECLI:EU:C:2014:2394 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .241Case T- 102/ 13, Heli- Flight GmbH & Co. KG v. EASA,

ECLI:EU:T:2014:1064. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206, 247, 338, 343, 351

Table of Casesxx

Order of the President in Case T- 201/ 13, Rubinum v. Commission, ECLI:EU:T:2013:562 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .366

Case T- 201/ 13, Rubinum v. Commission, ECLI:EU:T:2013:562 . . . . . . . . . . . . . . . . . . . . . . . . .366Opinion of AG Mengozzi in Case C- 88/ 14, Commission v.

Parliament and Council, ECLI:EU:C:2015:304 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .287Case C- 88/ 14, Commission v. Parliament and Council, ECLI:EU:C:2015:499 . . . . . . . . . . . . . .287Case C- 363/ 14, Parliament v. Council, ECLI:EU:C:2015:579 . . . . . . . . . . . . . . . . . . . . . . . 39, 265Case T- 660/ 14, SV Capital v. EBA, ECLI:EU:T:2015:608 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .341

DECISIONS OF THE BOA R DS OF A PPE A L

ECHA Board of Appeal, Case A- 005- 2011, Honeywell Belgium N.V. . . . . . . . . . . . . . . . . . 351, 352ECHA Board of Appeal, Case A- 001- 2012, Dow Benelux . . . . . . . . . . . . . . . . . . . . . . . . . . 350, 351ECHA Board of Appeal, Case A- 004- 2012, Lanxess . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .343EASA Board of Appeal, Case AP/ 01/ 2012, Heli- Flight GmbH & Co. KG . . . . . . . . . . . . . . . . . . 351EASA Board of Appeal, Case AP/ 03/ 2012, Luck . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .342EASA Board of Appeal, Case AP/ 04/ 2013, Robinson Helicopter Company . . . . . . . . . . . . . . . . .341ESAs Joint Board of Appeal, Case 2013- 8, SV Capital OÜ v. EBA . . . . . . . . . . . . . . . . . . . . . . .342ESAs Joint Board of Appeal, Case 2013- 14, Global Private Rating

Company ‘Standard Rating’ Ltd v. ESMA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361

FOR EIGN A ND MEMBER STATES’ COURTS’ DECISIONS

US Supreme Court, Marbury v. Madison, [1803] 5 United States Reports 137 . . . . . . . . . . . . . . .302US Supreme Court, McCulloch v. State of Maryland,

[1819] 17 United States Reports 316 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .300US Supreme Court, Union Bridge Co. v. US, [1907] 204 United States Reports 364 . . . . . . . . . .301US Supreme Court, United States v. Delaware & Hudson Co., [1909] 213

United States Reports 366 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .302US Supreme Court, Myers v. US, [1926] 272 United States Reports 52 . . . . . . . . . . . . . . . . . . . .300US Supreme Court, Hampton & Co. v. US, [1928] 276 United States Reports 394 . . . . . . . . . . .301US Supreme Court, United States v. Chicago, [1931] 282 United States Reports 311 . . . . . . . . . . .301US Supreme Court, New York Central S. Corp. v. United States, [1932]

287 United States Reports 12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .301US Supreme Court, Federal Radio Commission v. Nelson Bros. B. & M. CO.,

[1933] 289 United States Reports 266 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .301US Supreme Court, Panama Refining Co. v. Ryan, [1935] 293 United States Reports 388 . . . . . .301US Supreme Court, A.L.A. Schechter Poultry Corporation v. United States,

[1935] 295 United States Reports 495 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .301US Supreme Court, Humphrey’s Executor v. US, [1935] 295 United States Reports 602 . . . . . . . .300US Supreme Court, Skidmore v. Swift & Co., [1944] 323 United States Reports 134 . . . . . . . . . .302US Supreme Court, Atlas Roofing Co., Inc. v. Occupational Safety Commission,

[1977] 430 United States Reports 442 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .301US Supreme Court, Industrial Union Department, AFL- CIO v. American Petroleum

Institute, [1980] 448 United States Reports 607 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .302US Supreme Court, INS v. Chadha, [1983] 462 United States Reports 919 . . . . . . . . . . . . . . . . .301US Supreme Court, Chevron U.S.A., Inc. v. Natural Resources Defense

Council, Inc., [1984] 467 United States Reports 837 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .302– 3US Supreme Court, United States v. Varig Airlines, [1984] 467 United States Reports 797 . . . . . . 359US Supreme Court, Bowsher v. Synar, [1986] 478 United States Reports 714 . . . . . . . . . . . . . . . .300US Supreme Court, South Dakota v. Dole, [1987] 483 United States Reports 203 . . . . . . . . . . . . 161US Supreme Court, Berkovitz v. United States, [1988] 486 United States Reports 531 . . . . . . . . . 359US Supreme Court, Morrison v. Olson, [1988] 487 United States Reports 654 . . . . . . . . . . . . . . .300

Table of Cases xxi

US Supreme Court, Mistretta v. United States, [1989] 488 United States Reports 361 . . . . .301, 305US Supreme Court, New York v. United States, [1992] 505 United States Reports 144 . . . . . . 49, 160US Supreme Court, MCI Telecommunications Corp. v. American Telephone and

Telegraph Co., [1994] 512 United States Reports 218 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .303US Supreme Court, Printz v. United States, [1997] 521 United States

Reports 898. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49, 160, 266, 306US Supreme Court, Clinton v. City of New York, [1998] 524 United States

Reports 417 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .301, 306US Supreme Court, Reno v. Condon, [2000] 528 United States Reports 141 . . . . . . . . . . . . . . . . .49US Supreme Court, FDA v. Brown & Williamson Tobacco Corp., [2000]

529 United States Reports 120 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .302, 302US Supreme Court, Christensen v. Harris County, [2000] 529 United States

Reports 576 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .303US Supreme Court, Whitman v. American Trucking Associations, Inc.,

[2001] 531 United States Reports 457 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .305US Supreme Court, US v. Mead Corp., [2001] 533 United States Reports 218 . . . . . . . . . . . . . . .303US Supreme Court, Barnhart v. Walton, [2002] 535 United States Reports 212 . . . . . . . . . . . . . .303US Supreme Court, Free Enterprise Fund v. Public Company Accounting Oversight

Board, [2010] 561 United States Reports 477 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .300

Bundesverfassungsgericht, Kreditwesen, [1962] 14 BVerfGE 197 . . . . . . . . . . . . . . . . . . . . . 309, 310Bundesverfassungsgericht, Behördliches Beschwerderecht, [1973] 35 BVerfGE 263 . . . . . . . . . . . . 314Bundesverfassungsgericht, Mitbestimmungsgesetz Schleswig- Holstein, [1995]

93 BVerfGE 37 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312Bundesverfassungsgericht, Moratorium Gorleben, [2001] 104 BVerfGE 238 . . . . . . . . . . . . . . . .309Bundesverfassungsgericht, Lippeverband, [2002] 107 BVerfGE 59 . . . . . . . . . . . . . . . . . . . . . . . 312Bundesverfassungsgericht, Zollkriminalamt, [2004] 110 BVerfGE 33 . . . . . . . . . . . . . . . . . . . . . 310Bundesverfassungsgericht, Hartz IV- Arbeitsgemeinschaften, [2007] 119 BVerfGE 331 . . . . . . . .307

Table of Legislation

Décision (CECA) 33/ 53 de la Haute Autorité, J.O. 1953 8 137 . . . . . . . . . . 177

Décision (CECA) 22/ 54 de la Haute Autorité, J.O. 1954 4 286 . . . . . .176, 178

Décision (CECA) 14/ 55 de la Haute Autorité, J.O. 1955 8 685. . . . . . . 177, 178

Regulation (EEC) 574/ 72 of the Council, OJ 1972 L 74/ 1 . . . . . . . . . . .251

Regulation (EEC) 907/ 73 of the Council, OJ 1973 L 89/ 2 . . . . . . . . . . . .17

Directive (EEC) 75/ 319 of the Council, OJ 1975 L 147/ 13 . . . . . . . . .102

Regulation (EEC) 337/ 75 of the Council, OJ 1975 L 39/ 1 . . . . . . . . . . . . . . .13, 16, 335, 349

Regulation (EEC) 1365/ 75 of the Council, OJ 1975 L 139/ 1 . . . . . . . 13, 16,

335, 349Directive (EEC) 81/ 851 of the

Council, OJ 1981 L 317/ 1 . . . . . . . . . .102Regulation (EEC) 3245/ 81 of the

Council, OJ 1981 L 328/ 1 . . . . . . . . . . .17Regulation (EC) 2062/ 94 of the

Council, OJ 1994 L 216/ 1 . . .16, 335, 349Regulation (EC) 2100/ 94 of the

Council, OJ 1994 L 227/ 1 . . . . . . 16, 335, 338– 46

Regulation (EC) 2965/ 94 of the Council, OJ 1994 L 314/ 1 . . . . . . . . . . .16

Directive (EC) 95/ 46 of the European Parliament and of the Council, OJ 1995 L 281/ 31 . . . . . . . . . . . . . . . .163

Joint Action (JHA) 96/ 277 of the Council, OJ 1996 L 105/ 1 . . . . . . . . . .103

Joint Action (JHA) 97/ 12 of the Council, OJ 1997 L 7/ 5 . . . . . . . . . . . . 41

Joint Action (JHA) 98/ 428 of the Council, OJ 1998 L 191/ 4 . . . . . . . . . .103

Directive (EC) 98/ 8 of the European Parliament and of the Council, OJ 1998 L 123/ 1 . . . . . . . . . . . . . . . . . 32

Joint Action (JHA) 98/ 245 of the Council, OJ 1998 L 99/ 8 . . . . . . . . . . . 41

Decision (EC, ECSC, Euratom) 1999/ 352 of the Commission, OJ 1999 L 136/ 20 . . . . . . . . . . . . . . . . . . . . . . . .14

Directive (EC) 2000/ 43 of the Council, OJ 2000 L 180/ 22 . . . . . . . . . . . . . . . .163

Regulation (EC) 2666/ 2000 of the Council, OJ 2000 L 306/ 1 . . . . . . . . . .11

Regulation (EC) 2667/ 2000 of the Council, OJ 2000 L 306/ 7 . . . . . . . . . .11

Regulation (EC) 45/ 2001 of the European Parliament and of the Council, OJ 2001 L 8/ 1 . . . . . . . . . . . 239

Directive (EC) 2001/ 83 of the European Parliament and of the Council, OJ 2001 L 311/ 67 . . . . . . . . . . . . . . . . . . .25, 26, 31

Regulation (EC) 1049/ 2001 of the European Parliament and of the Council, OJ 2001 L 145/ 43 . . . . .211, 239

Regulation (EC) 178/ 2002 of the European Parliament and of the Council, OJ 2002 L 31/ 1 . . . . . . . . .16, 51

Decision (JHA) 2002/ 187 of the Council, OJ 2002 L 63/ 1 . . . . . . . . 16, 40

Framework Decision (JHA) 2002/ 465 of the Council, OJ 2002 L 162/ 1 . . . . . 42

Regulation (EC) 1406/ 2002 of the European Parliament and of the Council, OJ 2002 L 208/ 1. . . . . . . .16, 33

Regulation (EC) 58/ 2003 of the Council, OJ 2003 L 11/ 1 . . . . . . . .10, 157

Regulation (EC) 1829/ 2003 of the European Parliament and of the Council, OJ 2003 L 268/ 1. . . . . . .25, 335

Regulation (EC) 1831/ 2003 of the European Parliament and of the Council, OJ 2003 L 268/ 29 . . . . . . . . 32

Regulation (EC) 2065/ 2003 of the European Parliament and of the Council, OJ 2003 L 309/ 1 . . . . . . . . . . 32

Directive 2004/ 49 of the European Parliament and of the Council, OJ 2004 L 164/ 44 . . . . . . . . . . . . . . . . 36

Regulation (EC) 460/ 2004 of the European Parliament and of the Council, OJ 2004 L 77/ 1 . . . . . . . . . . . .11

Regulation (EC) 726/ 2004 of the European Parliament and of the Council, OJ 2004 L 136/ 1. . . . . . . 16, 43,

331, 332, 337Regulation (EC) 851/ 2004 of the

European Parliament and of the Council, OJ 2004 L 142/ 1 . . . . . . .16, 335

Table of Legislationxxiv

Regulation (EC) 881/ 2004 of the European Parliament and of the Council, OJ 2004 L 164/ 1 . . . . . . . . . .16

Regulation (EC) 883/ 2004 of the European Parliament and of the Council, OJ 2004 L 166/ 1 . . . . . . . . . 250

Regulation (EC) 1935/ 2004 of the European Parliament and of the Council, OJ 2004 L 338/ 4 . . . . . . . . . . . . . . . . .32, 335, 349

Regulation (EC) 2007/ 2004 of the Council, OJ 2004 L 349/ 1. . . . . . . . . . .17

Decision (JHA) 2005/ 681 of the Council, OJ 2005 L 256/ 63 . . . . . . . . . .16

Regulation (EC) 768/ 2005 of the Council, OJ 2005 L 128/ 1 . . . . . .17, 42– 3


Regulation (EC) 1907/ 2006 of the European Parliament and of the Council, OJ 2006 L 396/ 1 . . . . . . . . . . . . . .17, 337, 338– 46

Regulation (EC) 1920/ 2006 of the European Parliament and of the Council, OJ 2006 L 376/ 1 . . . . . . .16, 337

Regulation (EC) 1922/ 2006 of the European Parliament and of the Council, OJ 2006 L 403/ 9 . . . . . . . . . .17


Regulation (EC) 168/ 2007 of the Council, OJ 2007 L 53/ 1 . . . . . . . . . . . .17

Regulation (EC) 658/ 2007 of the European Commission, OJ 2007 L 155/ 10 . . . . . . . . . . . . . . . . . . . . . . . . 33

Regulation (EC) 717/ 2007 of the European Parliament and of the Council, OJ 2007 L 171/ 32 . . . . . . . . .170

Decision (EC) 70/ 2008 of the European Parliament and of the Council, OJ 2008 L 23/ 21 . . . . . . . . . . 40

Directive (EC) 2008/ 95 of the European Parliament and of the Council, OJ 2008 L 299/ 25 . . . . . . . . . 37

Regulation (EC) 216/ 2008 of the European Parliament and of the Council, OJ 2008 L 79/ 1 . . . . . . . . . . . . . . 16, 337, 338– 46

Regulation (EC) 294/ 2008 of the European Parliament and of the Council, OJ 2008 L 97/ 1 . . . . . . . . 17, 337

Decision (JHA) 2008/ 976 of the Council, OJ 2008 L 348/ 130 . . . . . . .103

Regulation (EC) 1007/ 2008 of the European Parliament and of the Council, OJ 2008 L 293/ 1 . . . . . . . . . . .11


Regulation (EC) 1339/ 2008 of the European Parliament and of the Council, OJ 2008 L 354/ 82. . . . . . . . . .16

Directive (EC) 2009/ 72 of the European Parliament and of the Council, OJ 2009 L 211/ 55 . . . . . . . . .117

Directive (EC) 2009/ 73 of the European Parliament and of the Council, OJ 2009 L 211/ 94 . . . . . . . . .117

Regulation (EC) 73/ 2009 of the Council, OJ 2009 L 30/ 16 . . . . . . . . . . 41

Directive (EC) 2009/ 140 of the European Parliament and of the Council OJ 2009 L 337/ 37 . . . . . . . . .117

Regulation (EC) 207/ 2009 of the Council, OJ 2009 L 78/ 1 . . . . . . . 16, 152,

335, 338– 46Decision (JHA) 2009/ 371 of the

Council, OJ 2009 L 121/ 37 . . . . 17, 39, 42Regulation (EC) 401/ 2009 of the

European Parliament and of the Council, OJ 2009 L 126/ 13 . . . . . . . . . .16

Decision (EC) 633/ 2009 of the European Parliament and of the Council, OJ 2009 L 190/ 1 . . . . . . . . . . . 9

Regulation (EC) 713/ 2009 of the European Parliament and of the Council, OJ 2009 L 211/ 1 . . . . . . 17, 337,

338– 46Decision (JHA) 2009/ 935 of the

Council, OJ 2009 L 325/ 12 . . . . . . . . . 39Regulation (EC) 1060/ 2009 of the

European Parliament and of the Council, OJ 2009 L 302/ 1 . . . . . . . . . 38


Regulation (EC) 1107/ 2009 of the European Parliament and of the Council, OJ 2009 L 309/ 1. . . . . . . . . . 32

Regulation (EC) 1211/ 2009 of the European Parliament and of the Council, OJ 2009 L 337/ 1 . . . . . . . . . . .17

Regulation (EC) 1224/ 2009 of the Council, OJ 2009 L 343/ 1 . . . . . . . . . 42

Table of Legislation xxv

Decision (EU) 2010/ 427 of the Council, OJ 2010 L 201/ 30 . . . . . . . . . .15

Regulation (EU) 439/ 2010 of the European Parliament and of the Council, OJ 2010 L 132/ 11 . . . . . . .17, 42

Regulation (EU) 912/ 2010 of the European Parliament and of the Council, OJ 2010 L 276/ 11 . . . . . . . . . .16

Regulation (EU) 1093/ 2010 of the European Parliament and of the Council, OJ 2010 L 331/ 12 . . . . . . . . . .17, 38, 337, 338– 46



Regulation (EU) 182/ 2011 of the European Parliament and of the Council, OJ 2011 L 55/ 13 . . . . . .239, 292

Decision (CFSP) 2011/ 411 of the Council, OJ 2011 L 183/ 16 . . . . . . . . . .16

Regulation (EU) 580/ 2011 of the European Parliament and of the Council, OJ 2011 L 165/ 3 . . . . . . . .11

Regulation (EU) 1077/ 2011 of the European Parliament and of the Council, OJ 2011 L 286/ 1 . . . . . . . .17

Regulation (EU) 1169/ 2011 of the European Parliament and of the Council, OJ 2011 L 304/ 18 . . . . . . . . . 33

Directive (EU) 2012/ 34 of the European Parliament and of the Council, OJ 2012 L 343/ 32 . . . . . . . . .117

Regulation (EU) 236/ 2012 of the European Parliament and of the Council, OJ 2012 L 86/ 1 . . . . . . . . . . . . . . 28, 39, 124, 242

Regulation (EU) 528/ 2012 of the European Parliament and of the Council, OJ 2012 L 167/ 1 . . . . . . . . . . 32

Regulation (EU) 648/ 2012 of the European Parliament and of the Council, OJ 2012 L 201/ 1 . . . . .37, 38

Regulation (EU) 100/ 2013 of the European Parliament and of the Council, OJ 2013 L 39/ 30 . . . . . . .33, 110

Regulation (EU) 526/ 2013 of the European Parliament and of the Council, OJ 2013 L 165/ 41 . . . . . . . . . .11

Regulation (EU) 1024/ 2013 of the Council, OJ 2013 L 287/ 63 . . . . . . . . .167

Decision (CFSP) 2014/ 75 of the Council, OJ 2014 L 41/ 13 . . . . . . . . . . .16

Decision (CFSP) 2014/ 401 of the Council, OJ 2014 L 188/ 73 . . . . . . . . . .16

Regulation (EU) 806/ 2014 of the European Parliament and of the Council, OJ 2014 L 225/ 1 . . . . . . . . . . . . . 17, 323, 338– 46

Decision (CFSP) 2015/ 1835 of the Council, OJ 2015 266/ 55 . . . . . . . . . . .16

Regulation (EU) 2015/ 2219 of the European Parliament and of the Council, OJ 2015 L 319/ 1 . . . . . . . . . . .16

List of Abbreviations

ACER Agency for the Cooperation of Energy RegulatorsAEC Association européenne pour la coopérationAFDI Annuaire français de droit internationalAFSJ Area of Freedom, Security and JusticeAöR Archiv des öffentlichen RechtsAVR Archiv des VölkerrechtsBEREC Body of European Regulators for Electronic CommunicationsBEREC Office Office of the Body of European Regulators for Electronic CommunicationsCAP Common Agricultural PolicyCCP Common Commercial PolicyCDE Cahiers de droit européenCdT Translation Centre for the Bodies of the European UnionCedefop European Centre for the Development of Vocational TrainingCEPOL European Police CollegeCFI Court of First InstanceCFSP Common Foreign and Security PolicyCIREA Centre for Information, Discussion and Exchange on AsylumCIREFI Centre for Information, Discussion and Exchange on the Crossing of

Frontiers and ImmigrationCJEUL Columbia Journal of European LawCLRev Columbia Law ReviewCLS Council Legal ServiceCMLRev Common Market Law ReviewCPVO Community Plant Variety OfficeCUP Cambridge University PressCYELS Cambridge Yearbook of European Legal StudiesDÖV Die Öffentliche VerwaltungEAC European Agency for CooperationEAR European Agency for ReconstructionEASA European Aviation Safety AgencyEASO European Asylum Support OfficeEBA European Banking AuthorityECB European Central BankECDC European Centre for Disease Prevention and ControlECHA European Chemicals AgencyECLR European Competition Law ReviewECMA European Electronic Communications Market AuthorityECSC European Coal and Steel CommunityEDA European Defense AgencyEEA European Environment AgencyEEAS European External Action ServiceEFCA European Fisheries Control AgencyEFDO European Film Distribution Office

List of Abbreviationsxxviii

EFSA European Food Safety AuthorityEFTA European Free Trade AreaEIB European Investment BankEIGE European Institute for Gender EqualityEIOPA European Insurance and Occupational Pensions AuthorityEIT European Institute of Innovation and TechnologyELJ European Law JournalELRev European Law ReviewEMA European Medicines AgencyEMCDDA European Monitoring Centre for Drugs and Drug AddictionEMCF European Monetary Cooperation FundEMSA European Maritime Safety AgencyENISA European Network and Information Security AgencyENP European Neighbourhood PolicyENPI European Neighbourhood and Partnership InstrumentEPL European Public LawEPO European Patent OrganisationERA European Railway AgencyERG European Regulators GroupERGECNS European Regulators Group for Electronic Communications Networks

and ServicesESA European Supervisory AuthorityESDC European Security and Defence CollegeESGAB European Statistical Governance Advisory BoardESMA European Securities and Markets AuthorityETF European Training FoundationEUI European University InstituteEUISS European Union Institute for Security Studieseu- LISA Agency for the operational management of large- scale IT systems in the

Area of Freedom, Security and JusticeEUMC European Monitoring Centre on Racism and XenophobiaEU- OSHA European Agency for Safety and Health at WorkEurofound European Foundation for the Improvement of Living and Working

ConditionsEurojust European Union’s Judicial Cooperation UnitEuropol European Police OfficeEUSC European Union Satellite CentreEuZW Europäische Zeitschrift für WirtschaftsrechtFRA Fundamental Rights AgencyFRC Federal Radio CommissionFrontex European Agency for the Management of Operational Cooperation at the

External Borders of the Member States of the European UnionFTC Federal Trade CommissionGG GrundgesetzGSA European GNSS AgencyIAC Interstate Aviation CommitteeICC Interstate Commerce CommissionIGC Intergovernmental ConferenceIIA Inter- institutional Agreement

List of Abbreviations xxix

JCMS Journal of Common Market StudiesJEPP Journal of European Public PolicyJZ JuristenzeitungLIEI Legal Issues of Economic IntegrationMJECL Maastricht Journal of European and Comparative LawNATO North Atlantic Treaty OrganisationNTER Nederlands Tijdschrift voor Europees RechtOHIM Office for Harmonisation in the Internal MarketOLAF European Anti- Fraud OfficeOUP Oxford University PressPSC Political and Security CommitteePSO Public Sector OrganisationRAE Revue des Affaires EuropéennesRDUE Revue du Droit de l’Union EuropéenneREALaw Review of European Administrative LawRFDA Revue française de droit administratifRISA Revue Internationale des Sciences AdministrativesRMC(UE) Revue du Marché Commun (et de l’Union européenne)RTDE Revue trimestrielle de droit européenRUE Revue de l’Union européenneSCIFA Strategic Committee on Immigration, Frontiers and AsylumSEA Single European ActSEW Tijdschrift voor Europees en economisch rechtSRB Single Resolution BoardSRM Single Resolution MechanismTBP Tijdschrift voor Bestuurswetenschappen en PubliekrechtTEU Treaty on European UnionTFEU Treaty on the Functioning of the European UnionWA Working ArrangementWEAG Western European Armaments GroupWEAO Western European Armaments OrganisationWEP West European PoliticsWEU Western European UnionWTO World Trade OrganizationYEL Yearbook of European Law

Introduction

This book aims to give a comprehensive overview of the phenomenon of agencifi-cation at EU level, both from a political and a legal perspective.

Today, EU agencies exist in the fields of food safety, environment, air safety, maritime safety, border control, asylum, fisheries, intellectual property, police cooperation, judicial cooperation, railways, medicines, human rights, chemicals, financial services, energy, telecoms, etc. The deepening of EU integration in these fields has its repercussions on agencification and since these fields are subject to continuous development, agencification is also in constant flux.

In the introductory chapter some basic concepts are defined, elaborated, and put in their context. To better understand the role EU agencies play, the dif-ferent tasks which may be entrusted to them are also explored. The ad hoc de-velopment of agencification is a distinctive characteristic of this institutional phenomenon, with the institutions so far failing to come to a formal model. However, a debate on an agency model often implicitly starts from the premise that such a model should in effect be pursued. The reasons why the search for an EU agency model is an approvable endeavour are therefore also spelled out and found in the EU’s need for legitimacy. An overview of some of the more important provisions in EU agencies’ mandates is then presented to give a state of play and to allow a provisional assessment of the Common Approach on Decentralised Agencies.

In the chapter on the political limits, the political forces driving the process of agencification are explored. Understanding these forces and looking at the strategic behaviour of EU institutions allow for the identification of the political limits to the process of agencification.

The next chapter focuses on the legal limits. First, agencies and agencification are looked at in the light of the general principles of EU law. Following this, the Meroni and Romano rulings will be analysed and their relevance for EU agen-cies assessed. For a long time these cases were only discussed between academics but recently, questions of fundamental constitutional importance in relation to agencification were put to the Court in the Short- selling case. Because the institu-tional balance is found to be relevant to agencification, this principle of EU law is first itself assessed and analysed in order to apply it in a scrutiny of agencifica-tion. Similarly, the concept of delegation is often used to qualify the empower-ment of EU agencies by the EU legislator even if the concept of conferral seems more appropriate. How the empowerment of EU agencies should be understood

Introduction2

is therefore also explored in light of the different delegation regimes that exist in the EU legal order.

Since the notion of control (over EU agencies) is found to be an important legal limit to agencification, the existing control over EU agencies and agencifica-tion is assessed. In order to have a sufficiently broad perspective on these issues, a comparative analysis is undertaken in relation to (independent) agencies in two federal polities which are often taken as examples to the EU, that is, the USA and Germany. Since an important aspect of this control is the judicial control exercised by the Court of Justice, a significant section is also devoted to this issue.

In a final chapter, insights gathered from the preceding sections are wrapped up and conclusions as well as an outlook on the future, including a proposal for a legal basis in the Treaties, are presented.

References to case law and legislation have been updated up to July 2015.

ISetting the Scene

EU Agencies, Agencification, and the EU Administration

1 EU Agencies and the Agency Phenomenon

The first step in properly delimiting the subject of this book is defining the EU agency concept. In the context of national agencification, Verhoest et al generally describe agencies as:

Organisations that operate at arm’s length of the government to carry out public tasks […]. They are structurally disaggregated from their parent ministries, are said to face less hierarchical and political influence on their daily operations and have more managerial freedom in terms of finances and personnel, compared to ordinary ministries or depart-ments. Agencies usually are not totally independent, because in many cases political ex-ecutives have ultimate responsibility for their activities.1

While this is not a clear definition it does list the typical characteristics of agencies and, in fact, a more precise definition is most likely impossible. The OECD noted that ‘As even a cursory review of international experience reveals, the term public agency, when used by a national government, really carries whatever meaning that government wishes to give to it’.2 Pollit et al enter into the question of why it is so hard to come to a short yet watertight definition of agencies,3 but this issue will not be further explored. The international context merely serves to show that agencification is not a unique EU process and that EU agencies share the basic characteristics listed by Verhoest et al. Yet, as will be shown, most EU agencies also fulfil a role which is different from the function which national agencies exercise.

Although agencies have been around in some states for several decades,4 a ver-itable process of agencification was kicked off following the promotion of the

1 Koen Verhoest, Sandra Van Thiel, Geert Bouckaert, and Per Lægred, ‘Agencification as a Global Phenomenon: Introduction’, in Verhoest, Van Thiel, Bouckaert, and Lægred (eds), Government Agencies: Practice and Lessons from 30 Countries, Basingstoke, Palgrave Macmillan, 2012, p 3.

2 OECD, ‘Financial Management and Control of Public Agencies’, (2001) Sigma Papers 32, http:// dx.doi.org/ 10.1787/ 5kml60vk0h9x- en, p 14.

3 Christopher Pollit et al, ‘Agency Fever? Analysis of an International Policy Fashion’, (2001) 3 Journal of Comparative Policy Analysis: Research and Practice 3, pp 273– 5.

4 Carsten Greve, Matthew Flinders, and Sandra Van Thiel, ‘Quangos— What’s In a Name? Defining Quangos from a Comparative Perspective’, (1999) 12 Governance 2, p 130.

http://dx.doi.org/10.1787/5kml60vk0h9x-en

http://dx.doi.org/10.1787/5kml60vk0h9x-en

Setting the Scene4

New Public Management (NPM) paradigm since the 1980s.5 However, relying on NPM hardly helps in clarifying the agency phenomenon, since NPM itself and what it constitutes is the subject of debate as well.6 Rommel and Christiaens note that ‘NPM is notorious for its amorphous nature and just about the hardest thing about NPM is to actually define and characterize it’.7

Even without a clear definition or consensus on the notion of agency and on the paradigm of NPM, it is accepted that agencies and agencification fit well with the ideas which are generally subsumed under the NPM paradigm. Yet although the general description by Verhoest et al may equally be applied to EU agencies, this does not necessarily mean EU agencification is also driven by NPM.

Verrier’s account of how national agencies fit so well with NPM is elucidating in this regard, since it shows how NPM and the agency model are an attack against (the stereotype of) a traditional unwieldy bureaucracy,8 but such a bur-eaucracy cannot be identified at EU level.9 Further, unless national agencification is simply being overlooked in EU studies, it is remarkable that only a few authors writing on EU agencies make the link between the latter and NPM.10

If NPM cannot explain EU agencification, proper attention should be devoted to the real driving force behind the process. Firstly, however, a clear definition of the notion of EU agency should be worked out. While a definition of what con-stitutes an ‘agency’ may not be necessary for the purposes of this study, it is clear that there is no point in defining political and legal limits to EU agencification without properly defining what is meant by EU agency.

5 Gérard Marcou, ‘Le Thème de l’Agence et la Réforme des Administrations Centrales’, in Molinier (ed), Les agences de l’Union européenne, Bruxelles, Bruylant, 2011, p 6.

6 According to Gruening, NPM cannot even be qualified as a paradigm. See Gernod Gruening, ‘Origin and Theoretical Basis of New Public Management’, (2001) 4 International Public Management Journal 1, p 19.

7 Jan Rommel and Johan Christiaens, ‘Beyond the Paradigm Clashes in Public Management’, (2006) 28 Administrative Theory & Praxis 4, p 613. For attempts at defining NPM, see Christopher Hood, ‘The New Public Management in the 1980’s: Variations on a Theme’, (1995) 20 Accounting, Organizations and Society 2/ 3, pp 94– 7; David Mathiasen, ‘The New Public Management and its Critics’, (1999) 2 International Public Management Journal 1, p 92; OECD, Governance in Transition: Public Management Reforms in OECD Countries, Paris, OECD, 1995, p 8.

8 Benoît Verrier, ‘Les Enjeux du Recours aux Agences: Quelques Tensions Problématiques’, in Molinier (ed), Les agences de l’Union européenne, Bruxelles, Bruylant, 2011, p 41.

9 Because EU agencies operate in a multi- organizational network, Everson et al also rejected NPM’s usefulness in the reform of European Public Management in the wake of the Santer crisis. See Michelle Everson, Giandomenico Majone, Les Metcalfe, and Adriaan Schout, The Role of Specialised Agencies in Decentralising EU Governance— Report Presented to the Commission, September 1999, p 183.

10 Faroud Shirvani, ‘New Public Management und europäische Agenturen’, (2008) 61 DÖV 1; Martin Shapiro, ‘Independent Agencies’, in Craig and de Búrca (eds), The Evolution of EU Law, Oxford, OUP, 2011, p 113; Adriaan Schout and Fabian Pereyra, ‘The Institutionalization of EU Agencies: Agencies as “Mini Commissions”’, (2011) 89 Public Administration 1, p 3. Helfritz simply notes that the EU has not been immune from the agencification developments at national level. See Vark Helfritz, Verselbständigte Verwaltungseinheiten der Europäischen Union, Berlin, Weissensee Verlag, 2000, p 79.

EU Agencies and the Agency Phenomenon 5

1.1 Definition

Before the academic debate on EU agencies gained real prominence, Everson re-marked that an official agency definition was lacking:11 an observation which was repeated by Vos ten years later.12 Today, so many years, institutional practice, and reflection later, there is still no official definition.13 In part this is unsurprising, since the EU institutions which have set up these bodies were foremost concerned with more practical and mundane issues than the more fundamental question of how these bodies should fit into the EU’s primary law- defined institutional architecture.

Yet, to answer this question, a more precise identification of the EU agencies is indispensable. A first attempt was triggered by the Commission’s White Paper on Governance.14 The Commission implicitly acknowledged that EU agencies are a policy response to the insufficient application of EU law15 and, in order to structure this response, the Commission suggested the idea of a framework on EU agencies.16 This required a definition of the EU agency instrument which the Commission put forward in its 2002 Communication on the operating framework for agencies: ‘agencies […] were created by regulation in order to perform tasks clearly specified in their constituent Acts, all have legal personality and all have a certain degree of organisational and financial autonomy.’17 The Commission then made a distinction between the ‘executive agencies’ and the so- called ‘regulatory agencies’.18 Both types of agencies are indeed radically different, and only the latter type is the subject of the present study. As a result, the Commission’s em-bryonic definition needed further refining. In 2005 the Commission eventually proposed a framework,19 and suggested that the term EU agency ‘sh[ould] mean any autonomous legal entity set up by the legislative authority in order to help regulate a particular sector at European level and help implement a Community policy […] This definition [does] not include the so- called “executive” agencies’.20

The 2005 definition was a step backwards in precision from the 2002 definition and the distinction between the ‘regulatory’ agency and the ‘executive’ agency

11 Michelle Everson, ‘Independent Agencies: Hierarchy Beaters?’, (1995) 1 ELJ 2, p 185.12 Ellen Vos, ‘Independence, Accountability and Transparency of European Regulatory

Agencies’, in Geradin, Muñoz, and Petit (eds), Regulation through Agencies in the EU, Cheltenham, Edward Elgar, 2005, p 122.

13 Article 1a of the EU’s Staff Regulations defines ‘agencies’ as ‘Union bodies to whom these Staff Regulations apply under the Union acts establishing them’. For evident reasons, this definition will not be relied upon.

14 European Commission, COM (2001) 428 final. For a brief discussion of the White Paper, see Jacqueline Dutheil de la Rochère, ‘Quelques réflexions à propos du Livre blanc de la Commission “Gouvernance européenne”’, (2002) RMCUE 454, pp 10– 15.

15 European Commission, COM (2001) 428 final, pp 23– 4. 16 Ibid, p 24.17 European Commission, COM (2002) 718 final, p 3.18 For a discussion of the executive agencies, see Edoardo Chiti, ‘Les agences et l’administration

directe dans l’Union européenne’, in Auby and Dutheil de la Rochère (eds), Droit administratif euro-péen, Bruxelles, Bruylant, 2007, pp 99– 107; Michael Koch, ‘Mittelbare Gemeinschaftsverwaltung in der Praxis’, (2005) 16 EuZW 15, pp 455– 9.

19 European Commission, COM (2005) 59 final. 20 Ibid, p 11.

Setting the Scene6

was still not clearly made, meaning the Commission’s definition was not further adopted in legal doctrine.21 This also has to do with the Commission’s confusing terminology, distinguishing between executive and regulatory agencies. Within the latter, the Commission made a further distinction between the ‘decision- making regulatory agencies’ and the ‘executive regulatory agencies’.22 But because of the definition’s vagueness, it is unclear why an ‘executive regulatory agency’ should not simply be an ‘executive agency’.

A part of the terminology problem may be retraced to the fuzzy concept of ‘regulation’. Morgan and Yeung note that ‘regulation is a phenomenon that is notoriously difficult to define with clarity and precision, as its meaning and the scope of its inquiry are unsettled and contested’.23 They claim regulation may then have multiple meanings, which can be put on a continuum ranging from a strict definition such as ‘deliberate attempts by the state to influence socially valuable behaviour […] by establishing, monitoring and enforcing legal rules’ to a broad one such as ‘all forms of social control, whether intentional or not and whether imposed by the state or not’.24 According to Morgan and Yeung, lawyers tend to focus on the stricter definition, and indeed Wessel and Wouters claim to understand regulation ‘in a broad sense’ as the ‘setting of rules, standards or prin-ciples that govern conduct by public and/ or private actors’.25

The Commission’s confusing terminology in relation to the EU agencies may have to do with the apparent link it makes between EU agencies and national agencies,26 borrowing some of its terminology from the agency experiences in the United States (US), the birthplace of the ‘independent regulatory agency’, according to authors such as Zwart and Delzangles.27 Although there is no gen-eral consensus on exact terminology in the US either, three basic groups of fed-eral agencies are generally distinguished: cabinet agencies (eg the Food and Drug Administration), executive agencies (eg the Environmental Protection Agency), and independent agencies (eg the Federal Trade Commission).28 The cabinet

21 See however Daniel Kelemen, ‘The Politics of Eurocracy: Building a New European State?’, in Jabko and Parsons (eds), The State of the European Union: With US or against US? European Trends in American Perspective, Oxford, OUP, 2004, Vol. 7, p 175.

22 European Commission, COM (2002) 718 final, p 8.23 Bronwen Morgan and Karen Yeung, An Introduction to Law and Regulation: Text and

Materials, Cambridge, CUP, 2007, p 3.24 Ibid, pp 3– 4.25 Ramses Wessel and Jan Wouters, ‘The Phenomenon of Multilevel Regulation’, in

Wouters, Wessel, and Follesdal (eds), Multilevel Regulation and the EU: The Interplay between Global, European, and National Normative Processes, Leiden, Martinus Nijhoff Publishers, 2008, p 12 (emphasis added).

26 See for instance COM (2001) 428 final, p 23.27 Tom Zwart, ‘Independent Regulatory Agencies in the US’, in Zwart and Verhey (eds),

Agencies in European and Comparative Law, Antwerp, Intersentia, 2003, p 3; Hubert Delzangles, ‘L’encadrement des agences américaines de régulation, un modèle pour l’Europe?’, in Peraldi Leneuf and Normand (eds), La légistique dans le système de l’Union européenne, Bruxelles, Bruylant, 2012, p 163.

28 Abraham Ribicoff, ‘Congressional Oversight and Regulatory Reform’, (1976) 28 Administrative Law Review 3, p 416; Peter Strauss, ‘The Place of Agencies in Government: Separation of Powers and the Fourth Branch’, (1984) 84 CLRev 3, p 583.


agency operates under a regular department, whereas the latter two do not. However, both the cabinet and executive agencies should be situated within the executive branch, and thus under the US President’s control,29 whereas the inde-pendent agencies have been referred to, since the Brownlow Commission’s report, as ‘a new and headless fourth branch of the Government’.30

These labels only relate to the position which these bodies occupy within the fed-eral administration and as a result their importance should not be overestimated,31 since all three may be regulatory agencies.32 From a US perspective, the Commission denoting some EU agencies as ‘regulatory’ is then surprising, given they lack proper powers. On the other hand, both the US and EU ‘executive agencies’ share the char-acteristic of a reduced independence from the main ‘executive’ actor. As a result, in the US context, ‘executive’ refers to the agency’s place in the constitutional struc-ture, but in the EU context, also given the problems of defining the EU’s structure along a traditional separation of powers, ‘executive’ has also come to be understood as qualifying the nature of the powers conferred on the agency.33

Coming back to the problem of defining EU agencies, it was already observed that the Commission’s definition and typology were not taken over in legal doctrine. The use of the term ‘regulatory’ by the Commission has further been criticized in legal doctrine34 and by the European Parliament, who proposed to refer to these bodies as ‘decentralized agencies’,35 a choice on which the institu-tions seem to have settled following the Common Approach on Decentralised Agencies.36 However, it should be noted that the agencies are as much decentral-ized as they are regulatory. As Scott notes, the only decentralizing aspect to the agencies is that they are not located in Brussels.37 This has led Driessen to refer to

29 Paul Verkuil, ‘The Purpose and Limits of Independent Agencies’, (1988) 37 Duke Law Journal 2– 3, p 259.

30 President’s Committee on Administrative Management, Report on the Administrative Management in the Government of the United States, 1937, p 346.

31 See Strauss, n 28, p 596.32 Marshall Breger and Gary Edles, ‘Established by Practice: The Theory and Operation of

Independent Federal Agencies’, (2000) 52 Administrative Law Review 4, p 1038.33 Although in the early days the US cabinet agency was, in theory, also confined to managerial

tasks. See Dominique Custos, ‘The Rulemaking Power of Independent Regulatory Agencies’, (2006) 54 The American Journal of Comparative Law Fall Suplement, p 617.

34 See inter alia Jacques Ziller, ‘Les autorités administratives indépendantes entre droit interne et droit de l’Union européenne’, (2010) 26 RFDA 5, p 904; Jorrit Rijpma, ‘Hybrid Agencification in the Area of Freedom, Security and Justice and Its Inherent Tensions: The Case of Frontex’, in Busuioc, Groenleer, and Trondal (eds), The Agency Phenomenon in the European Union, Manchester, Manchester University Press, 2012, pp 85– 6; Edoardo Chiti, ‘An Important Part of the EU’s Institutional Machinery: Features, Problems and Perspectives of European Agencies’, (2009) 46 CMLRev 5, p 1398; Herwig Hofmann, ‘Agency Design in the European Union’, (2010) 28 Windsor Yearbook of Access to Justice, p 314.

35 See point 6 of the Resolution of the European Parliament, OJ 2009 L 255/ 206.36 Stressing the difference between decentralized agencies and executive agencies, see Maroš

Šefčovič, ‘Ein Ordnungsrahmen für die dezentralisierten Agenturen der EU’, (2012) EuZW 21, pp 801– 2.

37 Colin Scott, ‘Agencies for European Regulatory Governance: A Regimes Approach’, in Geradin, Petit, and Muñoz (eds), Regulation through Agencies in the EU, Cheltenham, Edward Elgar, 2005, pp 70– 1. See also Dubos, who notes: ‘La création des agences n’est pas un mécanisme de

Setting the Scene8

them as statutory agencies,38 but for simplicity’s sake it is proposed here to simply refer to them as EU agencies.

With the issue of the bodies’ designation out of the way, the issue of defining EU agencies can be addressed. Some of the definitions proposed in legal doc-trine will be commented on before the definition used in this study is presented. Juxtaposing the latter with some alternatives will contribute to a better under-standing of what an EU agency is and what it is not.

Barbieri and Ongaro see EU agencies from a public management perspec-tive as structurally disaggregated organizational solutions established by the Community institutions, which take a specific place in the conceptual space of autonomy versus control and which show the following characteristics: dispersed over the entire EU territory; not a Commission service; not mentioned in the Treaties; not created outside the Treaties.39 Although the notion of agencies as organizational solutions is an interesting perspective, the definition lacks preci-sion and simplicity. Since any kind of organization has ‘a place’ in the conceptual space of autonomy- control, the reference to this aspect does not help in defining EU agencies. The lack of precision is furthermore apparent from the list which the authors distil from their definition, which does not include the executive agencies, although they still squarely fall within their proposed definition.

Van Ooik and Eijsbouts define EU agencies as all organs which (i) are legally linked to the EU and (ii) have been entrusted executive tasks under EU policies without much discretionary powers, but (iii) are not directly involved in the insti-tutional structure of the EU, resulting in a certain independence.40

However, while the first element aims to work around the problem that some EU agencies are provided for under the Treaties, this part of the definition could even encompass international organizations or private enterprises as long as there is ‘a legal link’ with the EU. The second element of the definition is even more problematic. By relying on the dichotomy between executive and discretionary powers, the authors import the problems inherent in the Meroni doctrine (cf. Chapter IV).

The third element is equally problematic because being ‘directly involved in the EU’s institutional structure’ is rather vague. The authors illustrate their definition by reference to the European Investment Bank (EIB), stating that it operates outside the EU’s institutional structure and has purely executive tasks. Apart from the fact

décentralisation […], mais plutôt un mécanisme de centralisation au détriment des États’: Olivier Dubos, ‘Objectif d’efficacité de l’exécution du droit de l’union européenne: la tectonique des com-pétences’, in Neframi (ed), Objectifs et compétences dans l’Union européenne, Bruxelles, Bruylant, 2012, pp 300– 1. Further, Dariusz Adamski, ‘The ESMA Doctrine: A Constitutional Revolution and the Economics of Delegation’, (2014) 39 ELRev 6, p 812.

38 Bart Driessen, Interinstitutional Convention as Checks and Balances in EU Law, Leuven, KUL, 2006, PhD Thesis, p 332.

39 Dario Barbieri and Eduardo Ongaro, ‘Les agences de l’UE: points communs et diffé-rences avec les agences publiques agissant au niveau national’, (2008) 74 RISA 3, p 423.

40 Ronald van Ooik and Tom Eijsbouts, ‘De wonderbaarlijke vermenigvuldiging van Europese agentschappen. Verklaring, analyse, perspectief ’, (2006) 54 SEW 3, p 105.


that the first observation is questionable in the light of SGEEM & Etroy v. EIB,41 being outside the EU’s institutional structure is also different from not being in-volved directly therein. Secondly, looking at the EIB’s mandate, it is difficult to conclude that the EIB only has executive tasks. For instance, part of the EIB’s man-date is to finance projects outside the EU.42 As Hachez and Wouters point out,43 this implies that the EIB must also strive to fulfil the general objectives of the EU’s external relations listed in Article 21 TEU. It is precisely this exercise of balancing different general objectives in a single decision which, according to the Court in Meroni, necessitates the exercise of discretionary powers (cf. section IV 5.1.4.2.2).

The definition which could be found on the EU’s own information site on agencies was used by a number of authors.44 According to that website an EU agency is (i) a body governed by European public law; (ii) distinct from the in-stitutions; (iii) with its own legal personality; (iv) set up by an act of secondary legislation; (v) to accomplish a very specific technical, scientific, or managerial task. Groenleer has criticized this definition because it excludes some agencies from the former second and third pillar which may allegedly be set up by an act of primary law.45 This point is debatable, however, since the agencies hinted at are merely foreseen in primary law, which is something different from being set up under primary law (cf. section I 1.1.3).

The definition which has inspired the definition used in this study builds on the one originally proposed by Fischer- Appelt and further refined by Griller and Orator.46 According to the former, EU agencies are established on a secondary law basis as relatively independent permanent bodies of the EU with their own legal personality and entrusted with specific tasks.47 For the purposes of the present

41 Case C- 370/ 89, SGEEM and Etroy v. EIB, [1992] ECR I- 6211, para 13.42 Decision (EC) 633/ 2009 of the European Parliament and of the Council, OJ 2009 L 190/ 1.43 Nicolas Hachez and Jan Wouters, ‘A Responsible Lender? The European Investment

Bank’s Environmental, Social and Human Rights Accountability’, (2012) 49 CMLRev 1, p 51.44 See inter alia Sharon Frank, A New Model for European Medical Device

Regulation: A Comparative Legal Analysis in the EU and the USA, Groningen, Europa Law Publishing, 2003, p 117; Govin Permanand and Ellen Vos, ‘EU Regulatory Agencies and Health Protection’, in Mossialos (ed), Health Systems Governance in Europe: The Role of European Union Law and Policy, New York, CUP, 2010, p 153; Martijn Groenleer, Michael Kaeding, and Esther Versluis, ‘Regulatory Governance through Agencies of the European Union? The Role of the European Agencies for Maritime and Aviation Safety in the Implementation of European Transport Legislation’, (2010) 17 JEPP 8, p 1215; Edoardo Chiti and Ramses Wessel, ‘The Emergence of International Agencies in the Global Administrative Space: Autonomous Actors or State Servants?’, in Collins and White (eds), International Organizations and the Idea of Autonomy, New York, Routledge, 2011, p 149. Although Busuioc does not refer as such to the information site, her definition is the same: see Madalina Busuioc, The Accountability of European Agencies: Legal Provisions and Ongoing Practices, Delft, Eburon, 2010, p 13. The information on the website which contained the defining characteristics has now been taken offline.

45 Martijn Groenleer, The Autonomy of European Agencies: A Comparative Study of Institutional Development, Delft, Eburon, 2009, p 19.

46 See Stefan Griller and Andreas Orator, ‘Everything under Control? The “Way Forward” for European Agencies in the Footsteps of the Meroni Doctrine’, (2010) 35 ELRev 1.

47 Dorothee Fischer- Appelt, Agenturen der Europäischen Gemeinschaft: eine Studie zu Rechtsproblemen, Legitimation und Kontrolle europäischer Agenturen mit interdisziplinären und re-chtsvergleichenden Bezügen, Berlin, Duncker & Humblot, 1999, p 38.

Setting the Scene10

study, Fischer- Appelt’s definition is only adapted in relation to the references to ‘specific tasks’ and the agencies’ ‘relative independence’.

Since every government body fulfils certain tasks, it would be difficult to im-agine the EU institutions taking the effort to establish a new body only to not endow it with a certain mission. The agencies’ independence then is problematic, because independence and autonomy do not take binary values. As a result, al-though these notions may describe EU agencies, they are ill suited to define agen-cies. In addition, the agencies’ independence may be subsumed under another defining feature, that is, their separate legal personality.

As a result, EU agencies may be defined as (i) permanent bodies, (ii) under EU public law, (iii) established by the institutions through secondary legislation, and (iv) endowed with their own legal personality. Each of these elements making up the definition will now be looked at in detail.

1.1.1 Permanent bodiesA first essential requirement is that EU agencies are permanent bodies. This does not mean that they can never be abolished, but instead that they are established for an indeterminate period and that at the moment of their establishment they are indeed created to remain part of the EU’s institutional machinery. This char-acteristic sets the EU agencies apart from a large number of other bodies active within the EU.

Perhaps firstly, it distinguishes the EU agencies from the EU executive agencies which otherwise meet the other requirements of the definition. The framework regulation for the executive agencies48 provides that the Commission determines the lifetime of each executive agency. Still, the Commission may extend this life-time as many times as necessary, meaning the executive agencies may de facto be as permanent as regular EU agencies, but the difference is that they are not established as permanent bodies, the Commission’s decisions always containing a sunset clause. Similarly, the acts establishing Joint Undertakings, provided for in Article 187 TFEU, also always contain a sunset clause.

A unique case in this regard is the European Network and Information Security Agency (ENISA); the Commission lists this as a ‘regulatory agency’,49 and is fol-lowed in this by most authors.50 Since the ENISA has only been established for a fixed period, this is remarkable for those authors referring to the permanent character of EU agencies,51 while others note that the ENISA is an exception to

48 See Article 3(1) of Regulation (EC) 58/ 2003 of the Council, OJ 2003 L 11/ 1.49 European Commission, COM (2008) 135 final, p 7.50 See inter alia van Ooik and Eijsbouts, n 40, p 105; Takis Tridimas, ‘Community Agencies,

Competition Law, and ECSB Iniatiatives on Securities Clearing and Settlement’, (2009) 28 YEL, p 231; Herwig Hofmann, Gerard Rowe, and Alexander Türk, Administrative Law and Policy of the European Union, Oxford, OUP, 2011, p 288.

51 Groenleer, n 45, p 384; Griller and Orator, n 46, pp 8– 9; Merijn Chamon, ‘EU Agencies: Does the Meroni Doctrine Make Sense?’, (2010) 17 MJECL 3, p 285.


the rule.52 The ENISA is indeed a special case distinguishable from the European Agency for Reconstruction (EAR). That agency was also generally classified as a (non- executive) EU agency while being non- permanent.53 Had the EAR been established today, it would have been an executive agency,54 since the EAR was established to implement the CARDS programme,55 just like the executive agen-cies are established to manage specific EU programmes. The ENISA on the other hand has a more general mission.56 Article 27 of the ENISA’s original establishing regulation57 provided it would be established for a period of five years— a term that has already been extended or reviewed three times.58 As Groenleer notes, agencies are apparently not easily abolished,59 with the cases of ENISA (see sec-tion II 3.6.1) and CEPOL (see section II 4.1) being exemplary.

1.1.2 EU public law bodiesThe requirement that EU agencies are EU public law bodies might seem super-fluous but is necessary nonetheless.

If the EU institutions establish a body by concluding an international agreement on behalf of the EU, the body created, falling under international law, is not an EU agency even if it exercises powers conferred on it by the EU institutions. For instance, had the European laying- up fund for inland waterway vessels (cf. section IV 5.5.1.1) been established, it could not have been qualified as an EU agency. In any event, the agreement establishing the laying- up fund would have been a mixed agreement and would bring not only the Member States and the EEC together, but also the Swiss Confederation. In the case of Europol, however, no third country was involved when Europol operated under the 1995 Convention. The Convention was a typical inter-national agreement,60 albeit that the contracting parties were the ‘Member States of the EU’. As a result, Europol was not a body under EU public law, but simply an inter-national organization. Only when Europol’s establishing instrument changed from the Convention to a decision under the former third pillar did it become an EU agency.

A more recent example is the European Financial Stability Facility (EFSF), which was set up pursuant to a decision of the representatives of the Governments of the Member States of the Euro Area meeting within the Council,61 but as a

52 Groenleer, n 45, p 27.53 See Article 16 of Regulation (EC) 2667/ 2000 of the Council, OJ 2000 L 306/ 7.54 See also Groenleer, n 45, p 27.55 See Regulation (EC) 2666/ 2000 of the Council, OJ 2000 L 306/ 1.56 Unlike the EAR, the ENISA was also established following the adoption of the framework

regulation for the executive agencies.57 Regulation (EC) 460/ 2004 of the European Parliament and of the Council, OJ 2004 L 77/ 1.58 See Regulation (EC) 1007/ 2008 of the European Parliament and of the Council, OJ 2008 L

293/ 1; Regulation (EU) 580/ 2011 of the European Parliament and of the Council, OJ 2011 L 165/ 3; Regulation (EU) 526/ 2013 of the European Parliament and of the Council, OJ 2013 L 165/ 41.

59 Groenleer, n 45, p 19.60 Alexandra De Moor and Gert Vermeulen, ‘The Europol Council Decision: Transforming

Europol into an Agency of the European Union’, (2010) 47 CMLRev 4, p 1092.61 Decision of the Representatives of the Governments of the Euro Area Member States Meeting

within the Council of the European Union, 10 May 2010, Doc. 9614/ 10.

Setting the Scene12

Luxemburgish public limited liability company.62 Because of this, and notwith-standing the undeniable link the EFSF has with the EU, the EFSF cannot be qualified as an EU public body.

That the EU agencies could be qualified as EEC bodies was not clear initially. In the Alaimo case, the question was put to the Court whether years in service of the Cedefop counted as years in service of the Communities. The Commission thought not, inter alia referring to the Cedefop’s own legal personality. The Court remarked in this regard that it was ‘necessary and sufficient to ascertain whether the Centre is part of the “European Communities”’,63 answering positively. Even if this was a (mere) staff case, the solution adopted by the Court resulted in the judicial recognition of the agencies as EU bodies, despite their own legal person-ality (cf. section I 1.1.4).

1.1.3 Established through secondary legislationA third defining characteristic of EU agencies is that they are established through secondary legislation. Above it was noted that some agencies are foreseen in the Treaties, that is, the European Defence Agency (EDA), the European Police Office (Europol), and the EU’s Judicial Cooperation Unit (Eurojust). These ref-erences are immaterial for the question of defining agencies since being men-tioned or foreseen in primary law does not mean these bodies are also established thereby.

Hilf ’s observations on the organizational structure of the EU is elucidating in this regard. Hilf sees three different organizational layers (Organisationsschichten) within the EU: the primary structure of the EU is laid down in the Treaties. The EU’s secondary structure is elaborated by the EU institutions based on specific en-abling clauses in the Treaties. Lastly there is the EU’s tertiary structure, for which the Treaties do not contain any explicit authorizations.64 Europol, Eurojust, and the EDA should be situated in the secondary structure of the EU, on which Hilf notes that the Treaties’ enabling clauses do not enter into the details of the tasks, competences, the external legal form, the internal structure, and the seat of such bodies.65

While the Treaties do provide some guidance on these agencies’ tasks, they leave it to the legislator to concretely define these tasks and to define the structure and functioning of these agencies.66 As a result these three bodies are not estab-lished by the Treaties and therefore still qualify as EU agencies.67 Although their

62 See Journal Officiel du Grand- Duché de Luxembourg, 8 juin 2010, C- No 1189, p 57026.63 Case 16/ 81, Alaimo v. Commission, [1982] ECR 1559, para 7.64 Meinhard Hilf, Die Organisationsstruktur der Europäischen Gemeinschaften, Berlin, Springer

Verlag, 1982, p 4.65 Ibid, p 65.66 See Article 45 TEU for the EDA, Article 85 TFEU for Eurojust, and Article 88 TFEU for

Europol.67 The major difference between the EDA and Eurojust and Europol lies in the precise definition

of the EDA’s tasks, whereas the tasks of the other two agencies are defined in secondary law. One could argue that the EDA therefore does not squarely fall within the secondary structure of the EU.


being part of the EU’s secondary rather than tertiary structure is irrelevant for the issue of qualifying them as EU agencies, the distinction is not without import-ance. Since the Treaties mandate the legislature to establish these agencies, the legality of their existence from a constitutional perspective is beyond any doubt (see also section IV 1.1).

For the purpose of clarifying the present element of the definition, EU agencies may be juxtaposed with other ‘agency- like’ organs such as the European Central Bank and the EIB. Both these bodies have legal personality separate from the EU68 and both were not formal institutions of the EU before the entry into force of the Lisbon Treaty. Still these bodies were established in primary law, forming part of the EU’s primary structure.69

1.1.4 Separate legal personalityA last defining characteristic of EU agencies is their separate legal personality. This characteristic really sets them apart from a number of other permanent and inde-pendent EU bodies. The EU agencies’ legal personality is explicitly provided for in their establishing acts following a typical provision which reads: ‘The [agency] shall have legal personality. It shall enjoy in all the Member States the most extensive legal capacity accorded to legal persons under their laws.’ Cedefop and Eurofound are an exception to this, since their establishing regulations only speak of the ‘most extensive legal capacity accorded to legal persons’.70 However, this difference in wording is not as such to deny these EU agencies their own legal personality.71

The inclusion of the element of legal personality also allows for a more pre-cise, clear, and unambiguous definition of an EU agency, since it makes refer-ences to the autonomy or independence of agencies redundant. It is undisputed that EU agencies have a certain autonomy and independence, but this character-istic may be covered by a reference to their distinct legal personality. Cherishing economy of statement, the element of a separate legal personality adds not only to the simplicity of the proposed definition but also to its clarity. This option is also preferred since qualifying the EU agencies’ autonomy and independence is not an easy undertaking.72 Since a governmental body is never completely independent, several authors refer to EU agencies as ‘relatively independent’,73

68 See Article 282(3) TFEU for the ECB and Article 308 TFEU for the EIB.69 See Protocol No 4 on the statute of the European system of central banks and of the European

Central Bank and Protocol No 5 on the statute of the European Investment Bank.70 Article 1 of Regulation (EEC) 337/ 75 of the Council, OJ 1975 L 39/ 1. Article 4(1) of

Regulation (EEC) 1365/ 75 of the Council, OJ 1975 L 139/ 1.71 Richard Lauwaars, ‘Auxiliary Organs and Agencies in the E.E.C.’, (1979) 16 CMLRev 3,

p 369; Fischer- Appelt, n 47, p 41.72 More generally, the use of the notion of ‘independence’ has been criticized by Scholten. See

Miroslava Scholten, ‘“Independent, hence Unaccountable”?— The Need for a Broader Debate on Accountability of the Executive’, (2011) 4 REALaw 2, pp 5– 44.

73 Fischer- Appelt, n 47, p 38; Gráinne de Búrca, ‘The Institutional Development of the EU: A Constitutional Analysis’, in Craig and de Búrca (eds), The Evolution of EU Law, Oxford, OUP, 1999, p 76; Griller and Orator, n 46, p 9.

Setting the Scene14

‘semi- independent’,74 ‘semi- autonomous’,75 or ‘quasi- autonomous’.76 Although these labels may be correct, they do not help towards identifying the EU agencies, since this would first require a clarification of the prefixes ‘relatively’, ‘semi’, and ‘quasi’.77 Granted, such an enquiry would be interesting— and has indeed been undertaken78— but it is not essential to come to a proper definition. It would also be cumbersome to classify bodies as EU agencies, since this would first necessitate a detailed analysis of their functioning.

A separate legal personality also juxtaposes the EU agency with other inde-pendent or autonomous EU bodies. For instance, other such bodies without legal personality are the European Security and Defence College (ESDC), the European Anti- Fraud Office (OLAF), the European External Action Service (EEAS), etc.

The ESDC is a Common Foreign and Security Policy (CFSP) body which is quite comparable to the existing CFSP agencies.79 Just like the latter, it finds its roots in the military cooperation in the context of the Western European Union (WEU). As early as 1991, the WEU Member States proposed to trans-form the WEU Institute of Security Studies into a European Security and Defence Academy.80 Because this would have necessitated allocating more resources,81 the Ministers’ intention never materialized until the ESDC was set up in 2005, within the EU rather than within the WEU. The ESDC provides training to civilian and military personnel to make them more familiar with ESDP matters. It is fully com-parable with the CFSP agencies or CEPOL in all respects apart from the fact that the Council did not provide it with legal personality.

OLAF was set up by the Commission to exercise the latter’s administrative investigations in the fight against fraud and corruption. Article 3 of the OLAF decision provides that it ‘shall exercise the powers of investigation […] in com-plete independence [and] the Director of the Office shall neither seek nor take instructions from the Commission, any government or any other institution or body’.82 Independence is indeed a fundamental prerequisite for OLAF to

74 Thomas Gehring and Sebastian Krapohl, ‘Supranational Regulatory Agencies between Independence and Control: the EMEA and the Authorization of Pharmaceuticals in the European Single Market’, (2007) 14 JEPP 2, p 222.

75 Everson, n 11, p 199.76 Deirdre Curtin, ‘Holding (Quasi- ) Autonomous EU Administrative Actors to Public

Account’, (2007) 13 ELJ 4.77 Barbara Remmert, ‘Die Gründung von Einrichtungen der mittelbaren Gemeinschaftsverwaltung’,

(2003) 37 Europarecht 1, p 135.78 See Groenleer, n 45, 2009.79 The EEAS’ online organizational chart even briefly listed the ESDC as a proper CFSP

agency.80 See Declaration on the Role of the WEU and Its Relations with the EU and with the Atlantic

Alliance, 10 December 1991, http:// www.weu.int/ documents/ 911210en.pdf.81 Sven Biscop, De integratie van de WEU in de Europese Unie, Leuven, Acco, 2000, p 66.82 Decision (EC, ECSC, Euratom) 1999/ 352 of the Commission, OJ 1999 L 136/ 20. As the pre-

amble to the decision setting up OLAF explains, OLAF succeeded the Task Force for Coordination of Fraud Prevention which itself succeeded the Unit for the Coordination of Fraud Prevention (UCLAF). The UCLAF had been established in 1987; see European Commission, COM (1987)

http://www.weu.int/documents/911210en.pdf

http://www.weu.int/documents/911210en.pdf


properly fulfil its function, but still OLAF has not been granted its own legal personality.83

The EEAS was established by the Council as a ‘functionally autonomous body’ of the EU.84 Although the EEAS has also been granted ‘the legal capacity neces-sary to perform and attain its objectives’, this cannot be equated with a separate legal personality.85 The cases of the Cedefop and Eurofound (discussed previously) cannot serve as precedents either, since at the time of the EEAS’ establishment the standard clause used in the agencies’ acts was already well established. Even textually, there is a major difference between the legal capacity of the Cedefop and Eurofound and that of the EEAS, since the latter only has the ‘necessary’ rather than the ‘most extensive’ legal capacity.

Being endowed with an own legal personality may be a defining characteristic of an EU agency, but this does not yet reveal the nature of that legal personality. Ruffert has drawn attention to this issue in light of the EU’s organizational plurality and in relation to the Euratom Supply Agency, Hilf noted that its legal personality pur-suant to Article 54 Euratom could be understood as national legal personality, inter-national legal personality, or legal personality under EU law.86 The same question is raised by the grant of legal personality to EU agencies. However, because it is not necessary to qualify the nature of the legal personality of EU agencies for the purpose of coming to a definition, this issue will not be addressed here (cf. section II 3.4).

1.1.5 ConclusionLacking an official definition, EU agencies are defined here as permanent bodies under EU public law, established by the institutions through secondary legislation and endowed with their own legal personality.

The definition’s four elements are easily operationalizable and as a result it dif-fers from other definitions which may be found in legal doctrine, which often do not allow an easy identification of the EU agencies. This will be done in the following section.

1.2 EU agencies

In the general debate on government agencies (cf. section I 1), Van Thiel has categorized public sector organizations along six types. Type 0 is a unit or a direc-tory of the central governments (comparable to a Commission DG at EU level);

572 final, pp 14– 15. OLAF was set up as a result of the poor performance of the UCLAF, which was assessed in detail by the Court of Auditors; see Special Report 8/ 98 of the Court of Auditors, OJ 1998 C 230/ 1.

83 Originally, the Commission proposed to establish OLAF as a body meeting the agency defin-ition. See European Commission, COM (1998) 717 final.

84 Decision (EU) 2010/ 427 of the Council, OJ 2010 L 201/ 30.85 Bart Van Vooren, ‘A Legal- institutional Perspective on the European External Action

Service’, (2010) Cleer Working Papers 7, pp 13– 15.86 Matthias Ruffert, ‘Personality under EU Law: A Conceptual Answer towards the

Pluralisation of the EU’, (2014) 20 ELJ 3, pp 357– 8; Hilf, n 64, pp 59– 64.

Setting the Scene16

Type 1 corresponds to a (semi- )autonomous body without legal personality; and Type 2 bodies are ‘legally independent organisations/ bodies […] either based on public (2a) or private (2b) law’.87 From a comparative perspective, then, EU agencies fall within the Type 2 (a) category.

Now that the different components of the working definition have been ex-plained in detail, the last question to solve is which bodies meet the four require-ments as presented and therefore qualify as EU agencies. The resulting list counts thirty- six bodies (in chronological order of establishment):

The European Centre for the Development of Vocational Training (Cedefop),88 the European Foundation for the Improvement of Living and Working Conditions (Eurofound),89 the European Training Foundation (ETF),90 the European Environment Agency (EEA),91 the European Medicines Agency (EMA),92 the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA),93 the Office for Harmonisation in the Internal Market (OHIM),94 the European Agency for Safety and Health at Work (EU- OSHA),95 the Community Plant Variety Office (CPVO),96 the Translation Center for the Bodies of the European Union (CdT),97 the European Police College (CEPOL),98 the European Union Institute for Security Studies (EUISS),99 the European Union Satellite Centre (EUSC),100 the European Food Safety Authority (EFSA),101 the European Union’s Judicial Cooperation Unit (Eurojust),102 the European Maritime Safety Agency (EMSA),103 the European Aviation Safety Agency (EASA),104 the European Centre for Disease Prevention and Control (ECDC),105 the European Railway Agency (ERA),106 the European Defence Agency (EDA),107 the European GNSS Agency (GSA),108 the

87 Sandra Van Thiel, ‘Comparing Agencies across Countries’, in Verhoest, Van Thiel, Bouckaert, and Lægred (eds), Government Agencies: Practice and Lessons from 30 Countries, Basingstoke, Palgrave Macmillan, 2012, p 19.

88 Regulation (EEC) 337/ 75, OJ 1975 L 39/ 1.89 Regulation (EEC) 1365/ 75, OJ 1975 L 139/ 1.90 Regulation (EC) 1339/ 2008 of the European Parliament and of the Council, OJ 2008 L

354/ 82.91 Regulation (EC) 401/ 2009 of the European Parliament and of the Council, OJ 2009 L 126/ 13.92 Regulation (EC) 726/ 2004 of the European Parliament and of the Council, OJ 2004 L 136/ 1.93 Regulation (EC) 1920/ 2006 of the European Parliament and of the Council, OJ 2006 L 376/ 1.94 Regulation (EC) 207/ 2009 of the Council, OJ 2009 L 78/ 1.95 Regulation (EC) 2062/ 94 of the Council, OJ 1994 L 216/ 1.96 Regulation (EC) 2100/ 94 of the Council, OJ 1994 L 227/ 1.97 Regulation (EC) 2965/ 94 of the Council, OJ 1994 L 314/ 1.98 Decision (JHA) 2005/ 681 of the Council, OJ 2005 L 256/ 63. Following the conclusion

of the manuscript, this Decision was replaced by Regulation (EU) 2015/ 2219 of the European Parliament and of the Council, OJ 2015 L 319/ 1.

99 Decision (CFSP) 2014/ 75 of the Council, OJ 2014 L 41/ 13.100 Decision (CFSP) 2014/ 401 of the Council, OJ 2014 L 188/ 73.101 Regulation (EC) 178/ 2002 of the European Parliament and of the Council, OJ 2002 L 31/ 1.102 Decision (JHA) 2002/ 187 of the Council, OJ 2002 L 63/ 1.103 Regulation (EC) 1406/ 2002 of the European Parliament and of the Council, OJ 2002 L 208/ 1.104 Regulation (EC) 216/ 2008 of the European Parliament and of the Council, OJ 2008 L 79/ 1.105 Regulation (EC) 851/ 2004 of the European Parliament and of the Council, OJ 2004 L 142/ 1.106 Regulation (EC) 881/ 2004 of the European Parliament and of the Council, OJ 2004 L 164/ 1.107 Decision (CFSP) 2011/ 411 of the Council, OJ 2011 L 183/ 16. Following the conclusion of

the manuscript, this Decision was replaced by Decision (CFSP) 2015/ 1835 of the Council, OJ 2015 266/ 55.

108 Regulation (EU) 912/ 2010 of the European Parliament and of the Council, OJ 2010 L 276/ 11.


European Agency for the Management of Operational Cooperation at the External Borders of the Member States of the European Union (Frontex),109 the European Fisheries Control Agency (EFCA),110 the European Chemicals Agency (ECHA),111 the European Institute for Gender Equality (EIGE),112 the Fundamental Rights Agency (FRA),113 the European Institute of Innovation and Technology (EIT),114 the European Police Office (Europol),115 the Agency for the Cooperation of Energy Regulators (ACER),116 the Office of the Body of European Regulators for Electronic Communications (Berec Office),117 the European Asylum Support Office (EASO),118 the European Banking Authority (EBA),119 the European Insurance and Occupational Pensions Authority (EIOPA),120 the European Securities and Markets Authority (ESMA),121 the Agency for the operational management of large- scale IT systems in the Area of Freedom, Security and Justice (eu- LISA),122 and the Single Resolution Board (SRB) for the banking union’s single resolution mechanism.123

In addition, two further bodies have been established in the past which also meet the requirements of the definition but which have now been (de facto) abolished. These bodies are the European Monetary Cooperation Fund (EMCF)124 and the European Agency for Cooperation (EAC).125

The EMCF, the very first agency, was established in the 1970s as the precursor to the ECB.126 The origins of the EMCF dated back to the Summit of The Hague in 1969, when it was decided that the possibility of setting up a European Reserve Fund would be investigated. This exercise was undertaken by the Werner and Ansiaux expert groups. Based on their reports, the Council invited the committee of Central Bank Governors and the Monetary Committee to elaborate a further report on an EMCF.127 Following the adoption of this report, the Council ex-pressed its views on the tasks which the EMCF should fulfil and on the legal base to establish such a body.128 When the Commission made the formal proposal for

109 Regulation (EC) 2007/ 2004 of the Council, OJ 2004 L 349/ 1.110 Regulation (EC) 768/ 2005 of the Council, OJ 2005 L 128/ 1.111 Regulation (EC) 1907/ 2006 of the European Parliament and of the Council, OJ 2006 L 396/ 1.112 Regulation (EC) 1922/ 2006 of the European Parliament and of the Council, OJ 2006 L 403/ 9.113 Regulation (EC) 168/ 2007 of the Council, OJ 2007 L 53/ 1.114 Regulation (EC) 294/ 2008 of the European Parliament and of the Council, OJ 2008 L 97/ 1.115 Decision (JHA) 2009/ 371 of the Council, OJ 2009 L 121/ 37.116 Regulation (EC) 713/ 2009 of the European Parliament and of the Council, OJ 2009 L 211/ 1.117 Regulation (EC) 1211/ 2009 of the European Parliament and of the Council, OJ 2009 L 337/ 1.118 Regulation (EU) 439/ 2010 of the European Parliament and of the Council, OJ 2010 L 132/ 11.119 Regulation (EU) 1093/ 2010 of the European Parliament and of the Council, OJ 2010 L 331/ 12.120 Regulation (EU) 1094/ 2010 of the European Parliament and of the Council, OJ 2010 L 331/ 48.121 Regulation (EU) 1095/ 2010 of the European Parliament and of the Council, OJ 2010 L 331/ 84.122 Regulation (EU) 1077/ 2011 of the European Parliament and of the Council, OJ 2011 L 286/ 1.123 Regulation (EU) 806/ 2014 of the European Parliament and of the Council, OJ 2014 L 225/ 1.124 Regulation (EEC) 907/ 73 of the Council, OJ 1973 L 89/ 2.125 Regulation (EEC) 3245/ 81 of the Council, OJ 1981 L 328/ 1.126 For a detailed account of the establishment of the EMCF, see Jean- Victor Louis, ‘Le Fonds

Européen de Cooperation Monetaire’, (1973) 9 CDE 3, pp 255– 97.127 Résolution du Conseil et des Représentants des Gouvernments des Etats membres, OJ 1971

C 28/ 1.128 See the Declaration following the Paris Summit, Bulletin of the European Communities, 1972/

10, p 17.

Setting the Scene18

the EMCF,129 most issues had therefore already been worked out. Elucidating for agencification in general are the following observations in the Ansiaux Report:

We can already ask ourselves whether, among the different paths possible for progressively realising these objectives [of the final stage of an economic and monetary union], it would not be opportune to rely on methods which not only guarantee consultation [between central banks], but which also allow the Community to perfect those mechanisms, which are necessary under any hypothesis, to achieve the final stage […] From this point of view, the introduction of a Community organ managed by the Governors of the central banks might be very effective to allow the process to be completed, compared to the simple con-sultation among central banks […] It is a matter of appreciation to determine at which moment the establishment of such an organ would be most useful. On this topic, opinions may diverge and the experts, taking the view that this is a political matter rather than a technical one, do not take position on this issue.130

The Ansiaux group thus foretold how EU agencies both bring together national administrations and constitute an (independent) executive capacity at EU level, equally noting how this decision ultimately is a political rather than a technical one. The Werner group noted that the EMCF would be destined to further in-tegrate into a Community system of central banks and that each Member State should therefore be represented within the EMCF.131

The EAC for its part was established in the 1980s to take over the tasks exercised by the Association européenne pour la cooperation (AEC), a body under private law est-ablished in 1964. The members of the AEC were all civil servants of the Commission and the AEC managed EEC personnel sent to developing countries.132 The EAC was established to embed the AEC in the EEC framework,133 but it never went into actual operation because no budget was allocated for it to recruit personnel.134

Finally, although this was already touched upon briefly previously, two agen-cies which were generally considered to be ‘regulatory agencies’ have been ex-cluded from the list of thirty- six. These are the EAR and the ENISA because of their non- permanent character. However, the ENISA will not be completely excluded from this study, since the ENISA case is useful to shed light on some issues concerning agencification through ‘genuine’ EU agencies.

1.2.1 Classifying EU agenciesClassifications are useful for bringing structure and insight to an otherwise dis-orderly reality. After presenting the typical classifications of EU agencies found in

129 European Commission, COM (73) 69 final.130 Groupe d’experts sous la présidence de Baron Ansiaux, Rapport au Conseil et à la Commission

concernant la réalisation par étapes de l’union économique et monetaire dans la Communauté, 8 octobre 1970, p 17. Translated from French by the author.

131 Groupe d’experts sous la présidence de M. Mertens de Wilmars, Rapport sur le Fonds Européen de Coopération, 1er juin 1972, pp 19– 20.

132 See the Statutes of the AEC, Belgisch Staatsblad/ Moniteur Belge, 3 oktober/ octobre 1964, pp 10536– 7.

133 European Commission, COM (1978) 93 final.134 See Case T- 319/ 00, Borremans e.a. v. Commission, [2002] ECR II- 905, para 3.


legal doctrine, it will be found that a classification of agencies’ powers rather than of the agencies themselves is more elucidating.

Just like there is no official definition of EU agencies, there is no official typ-ology or academic consensus thereon. The typologies advanced in legal doctrine suffer from much the same deficiencies as the definitions discussed previously. Craig, concerned by the limited added value of these typologies, remarked that ‘we should not allow the desire for “order” lead to the imposition of a taxonomic Procrustean frame that forces agencies into categories that are ill- fitting’.135 Indeed, a good typology should inter alia reduce complexity; help identify both similarities and differences between cases; form the descriptive basis for analysis and explanation; and remain manageable, meaning that it should not be too spe-cific or too general so as to become useless for further analysis.136

The typology proposed by the Commission was already touched upon previ-ously, whereby Geradin and Petit agree that the Commission’s typology was non-sensical.137 They further emphasize the Commission’s confused understanding of the concept of ‘regulation’, since it observed in 2003 that the ERA would only have an advisory role but not a direct regulatory role.138 However, under its understanding of ‘regulation’ in its 2005 proposal for an interinstitutional agree-ment,139 advisory functions also come under the ‘regulation’ umbrella.

Because the Commission’s typology was not really helpful, different authors have proposed their own taxonomies, which Griller and Orator have classified in four types. The dimension used to classify agencies may be temporal, struc-tural, functional, or instrumental.140 A temporal classification would divide the EU agencies into four categories, often labelled as the four waves of agencifica-tion matching the periods of the 1970s, the 1990s, the early 2000s, and the late 2000s/ early 2010s.141 This temporal approach brings limited added value since there is no real correlation between the time of establishment and an agency’s key characteristics.142 The only trend which may be noted is that the more recent agen-cies are indeed endowed with more significant powers than the earlier agencies. Today the importance of agencification mainly lies in its qualitative dimension (cf. section I 1.3), whereby more (far- reaching) powers are entrusted to agencies. Qualitative agencification then (mostly) occurs when existing agencies are granted

135 Paul Craig, EU Administrative Law, Oxford, OUP, 2006, p 152.136 Kenneth Bailey, Typologies and Taxonomies: An Introduction to Classification Techniques,

Thousand Oaks, Sage, 1994, pp 12– 16.137 Damien Geradin and Nicolas Petit, ‘The Development of Agencies at EU and National

Levels: Conceptual Analysis and Proposals for Reform’, (2005) 23 YEL, p 181.138 European Commission, COM (2003) 270 final, p 51. 139 See n 19.140 Griller and Orator, n 46, p 10.141 Most identify the 1970s as the first wave. See eg Fischer- Appelt, n 47, p 46; Geradin

and Petit, n 137, p 171; Craig, n 135, pp 148– 50; Tridimas, n 50, pp 230– 2; Griller and Orator, n 46, p 12. Others see the 1990s as the first wave: see Berthold Rittberger and Arndt Wonka, ‘Credibility, Complexity and Uncertainty: Explaining the Institutional Independence of 29 EU Agencies’, (2010) 33 WEP 4, p 730; Mark Thatcher, ‘The Creation of European Regulatory Agencies and Its Limits: A Comparative Analysis of European Delegation’, (2011) 18 JEPP 6, p 796.

142 See Bailey, n 136, p 2.

Setting the Scene20

further powers, but the temporal approach only captures the initial establishment of an agency.

The added value of a structural approach, which refers to the different pillars under which the agencies operate, is even more limited,143 even if it should be noted that the Common Approach on decentralised agencies of 2012 still distin-guishes the CFSP agencies from the other agencies.144

This leaves the functional and instrumental approaches. Both offer potentially interesting insights and the former in particular has proved to be very popular. Following a functional approach, the defining characteristic of an agency depends on the function it is tasked to exercise. In her functional classification, Vos makes a distinction between four tasks, resulting in three types of agencies: (i) infor-mation, (ii) management, and (iii) regulatory agencies.145 Van Ooik saw three functions: (i) collecting information, (ii) applying EU rules in specific cases, and (iii) assisting the Commission in implementing EU programmes.146

Geradin and Petit distinguished between five functions:147 (i) implement-ing EU law in specific regimes, (ii) observing specific policy fields, (iii) pro-moting social dialogue, (iv) operating as a subcontractor to the Commission, and (v) ensuring network safety/ interoperability. Geradin and Petit stress the difficulties in coming to a functional characterization, since one single agency may fall under different categories.148 Chiti drew up three categories of agen-cies: (i) those coordinating common information systems, (ii) those that have an assistance function, and (iii) and those taking binding decisions in adminis-trative procedures.149

Lastly, Busuioc distinguished between (i) information agencies, (ii) manage-ment agencies, (iii) operational cooperation agencies, (iv) decision- making agen-cies, and (v) quasi- regulatory agencies. Busuioc also struggled with the problem that a single agency may fulfil several functions and as a result she has categorized a number of agencies several times,150 as is evident from the graphical representa-tion of these five typologies.151

Figure 1.1 itself allows some conclusions to be drawn. Firstly, while all au-thors claim to have followed a functional approach, both the number of categories

143 Griller and Orator, n 46, p 10.144 See Joint Statement of the Council of the EU, the European Commission and the European

Parliament on Decentralised Agencies, 19 July 2012.145 Ellen Vos, ‘Agencies and the European Union’, in Zwart and Verhey (eds), Agencies in

European and Comparative Perspective, Antwerpen, Intersentia, 2003, pp 119– 21.146 Ronald van Ooik, ‘The Growing Importance of Agencies in the EU: Shifting

Governance and the Institutional Balance’, in Curtin and Wessel (eds), Good Governance and the European Union: Reflections on Concepts, Institutions and Substance, Antwerpen, Intersentia, 2005, pp 139– 45.

147 Geradin and Petit, n 137, pp 177– 80. 148 Ibid, p 180.149 Edoardo Chiti, n 34, pp 1403– 4.150 Busuioc, n 44, pp 26– 7.151 Evidently, these five classifications only categorize those agencies which existed at the time

of writing. For even earlier classifications, see inter alia Everson, n 11, pp 186– 8; Alexander Kreher, ‘Agencies in the European Community— A Step towards Administrative Integration in Europe’, (1997) 4 JEPP 2, pp 236– 8; Fischer- Appelt, n 47, pp 47– 9.

EU

SC

ISS

FR

AE

EA

EM

CD

DA

Ced

efo

pE

uro

foun

dE

U-O

SH

A

B1

A1

C1

D1

E1

C2

B2

E2

D2

C5

E5

A2

C4

E4

C3

A3

B3

D3

E3

EF

SA

EN

ISA

EC

DC

ER

A

EM

SA

EA

SA

EC

HA

OH

IMC

PV

O

EF

CA

EM

A

Fro

ntex

Eur

op

ol

Eur

oju

stE

DA

GS

AC

epo

l

Cd

T

Exe

cutiv

e ag

enci

es

ET

FE

AR

A. V

os

B. V

an O

oik

C. G

erad

in &

Pet

itD

. Chi

tiE

. Bus

uio

c

Figu

re 1

.1 F

unct

iona

l cla

ssifi

catio

ns o

f EU

age

ncie

s

Setting the Scene22

they induce and the specific classification of individual agencies are different. The figure shows that there is a general consensus on the ‘information’ func-tion of agencies such as the EEA, EMCDDA, FRA, Cedefop, Eurofound, and EU- OSHA and on the function of the EMA, OHIM, CPVO, and ECHA, even if there is no agreement on how that function should be labelled. There is also some agreement on the ‘executive’ or ‘management’ function of the former EAR, the CdT, and the ETF, although Vos sees the latter two as information agencies. While the classifications of Geradin and Petit and of Van Ooik predate CEPOL and GSA, it is safe to assume these authors would also classify these bodies under their ‘executive’ categories.

The figure also shows how certain agencies are notoriously difficult to classify. This is especially true for the EFSA, EASA, and ERA. According to Van Ooik, the ERA is an information agency, whereas Busuioc regards it as a quasi- regulatory agency. The same goes for the EFSA, which again is an information agency to Van Ooik and an observatory agency to Geradin and Petit, but a regulatory agency to Vos and a quasi- regulatory agency to Busuioc. The EASA then may be found in as many as three of Busuioc’s categories, both as an operational, a decision- making, and a quasi- regulatory agency, but Chiti classifies it in the same category as the CdT, since for him the EASA (merely?) fulfils a technical assistance function vis- à- vis the Commission.

1.2.1.1 A functional classificationGiven the existing plethora of functional classifications, caution is in order when-ever another functional classification is proposed. The added value of such a clas-sification was also questioned by Geradin and Petit, who note that a functional typology is helpful from an analytical point of view, but cannot be a basis for proposals for reform.152 However, that view is not subscribed to here. Although the extent of the powers conferred to EU agencies is indeed fundamental for a legal assessment, the function of these bodies and the specific moments at which they intervene in the policy cycle is also relevant, the constitutional limits varying depending on the specific activity/ function exercised by an EU institution, body, or agency. This will be elaborated upon later (see section I 1.2.1.2.1), following a presentation of an instrumental classification.

Because a functional classification is also relevant in the light of a possible reform of EU agencies, such a general classification will also be presented here. A first category then comprises the agencies which provide services to other EU institutions or bodies or to which management tasks have been outsourced. To set them apart from the executive agencies under Regulation 58/ 2003 (cf. section I 1.1), these bodies will be called ‘service agencies’. They are: the CdT, the ETF, the GSA, the BEREC Office, the EIT, and the eu- LISA. These bodies are more or less outsiders in the constellation of EU agencies. Some of these agencies could ac-tually be established as ‘executive agencies’ under Regulation 58/ 2003, or as joint

152 Geradin and Petit, n 137, p 180. Similarly, see Griller and Orator, n 46, p 12.


undertakings.153 From a reform perspective, then, these agencies are the least (if at all) problematic.

A second category of EU agencies is that of the ‘information agencies’ whose main task is to produce, gather, and disseminate information. The third category builds further on the second and comprises the assistance agencies. As Majone ob-served, the agencies of the second wave, most of which were information agencies, did not have far- reaching powers but could still play a role in indirect regulation, by shaping the information on which policy actors base their decisions.154 This is possibly the least intrusive way to induce multiple actors to apply a joint policy. At the same time it presupposes that these actors are rational and have the same interests,155 which is not always the case.

For both categories, an important question is who the agencies work for. Do agencies gather information for the EU institutions or for the Member States?156 This is linked to the question of whether agencies are involved in policy formu-lation or implementation. The first group is least problematic and the question mainly is whether the EU has any competence in the area concerned (ie what is the EU’s competence in the field of gender equality?). The same question is per-tinent for the second group, and in addition the question may be raised if and how EU agencies may affect the Member States’ autonomy in implementing EU law. Similarly, does an EU agency assist the EU institutions or the Member States?157 An additional question here is what ‘assistance’ means. It could mean supporting operational cooperation (EFCA, EMSA, Frontex, Europol, Eurojust) or adopt-ing soft law and/ or providing training to national civil servants (ERA, EASO). From a reform perspective, the relevant question is what ‘assistance’ means in practice. Depending on who is being assisted (institutions or Member States), this question should be looked at from a separation- of- powers (or institutional bal-ance) perspective or a federal perspective.

A fourth category of agencies may be termed the ‘internal market agencies’. These agencies are directly involved in defining the terms and conditions of the internal market. The best and earliest examples of these agencies are the CPVO and the OHIM, which grant intellectual propriety rights in designs, models, and plant varieties for the entire internal market. For these agencies, it is not only their core function, but also the only one they have. This is different for agencies

153 As regards the EIT for instance, the EESC, during the legislative process, suggested that it be established as a Joint Undertaking. See point 4.12.1 of the Opinion of the European Economic and Social Committee, OJ 2007 C 161/ 28.

154 Giandomenico Majone, ‘The New European Agencies: Regulation by Information’, (1997) 4 JEPP 2, p 265.

155 It is clear such ‘indirect regulation’ can only work efficiently if actors act (sufficiently) ration-ally and are motivated (predominantly) by the general policy objectives.

156 Information agencies falling under the first type are the Cedefop, Eurofound, EU- OSHA, FRA, EIGE, EUSC, and ISS. Information agencies belonging to the second type are the EEA, EMCDDA, and the ECDC.

157 Assistance agencies falling under the first type are the EMA and EFSA. Information agencies belonging to the second type are the EFCA, Frontex, EMSA, ERA, EASO, Europol, Eurojust, and the EDA.

Setting the Scene24

such as the EASA, ECHA, ACER, EBA, EIOPA, ESMA, and SRB, which also regulate the internal market, combining these functions with those of ‘assistance agencies’. Often they will ‘assist’ both the Member States and the Commission. From a reform perspective, it is relevant here to ascertain how far these agencies may be established based on the internal market provision of the TFEU (Article 114). In addition, the question arises of how these atypical bodies may fulfil their mandates in full respect of the prerogatives of the institutions and the Member States.

From this overview it becomes clear how hard it is to classify agencies along impermeable lines. It is also apparent that giving agencies ‘higher functions’ gen-erally implies giving them ‘lower functions’. After all, it is hard to imagine how agencies could assist the Commission or the Member States without at the same time, and as a precondition to their assistance, having acquired the relevant ex-pertise similarly to an ‘information agency’. As was already hinted at previously, the dichotomy of assisting or providing information to either the EU institutions or the Member States is also too simplistic. Figure 1.2 probably gives a more nu-anced yet still clear picture of the types of EU agencies.

1.2.1.2 An instrumental classificationBecause it is easier to qualify the powers of an agency, there is less divergence be-tween the instrumental classifications proposed by different authors.

The instrumental approaches proposed by Geradin and Petit and Griller and Orator are rather coarse since a single criterion is used, that is, whether the agency can take a binding decision or not.158 Evidently this is a first criterion, but it needs to be complemented by others, inter alia the scope of the decisions and how au-thoritative they are. Again, it will be impossible to classify agencies in exclusive categories, since a single agency may wield different powers. As a result, the ques-tion is not so much whether an agency is ‘decision- making’ or not, but whether it acts as such in a specific case.

Service Information

Market

Internal

EUMS

Assistance

Figure 1.2 A functional typology of EU agencies

158 Geradin and Petit identify three categories: (i) executive agencies without decision- making powers; (ii) decision- making agencies, including those without formal decision- making powers but which nonetheless enjoy a considerable influence over the adoption of the Commission’s final de-cisions; and (iii) regulatory agencies with discretionary powers. See Geradin and Petit, n 137, pp 182– 3. Griller and Orator list four categories: (i) ordinary agencies without decision- making powers, (ii) pre- decision- making agencies, (iii) genuine decision- making agencies, and (iv) rulemak-ing agencies. See Griller and Orator, n 46, p 13.


A first category is the non- decision- making agencies, that is, agencies without an independent decision- making power. A question here is whether these agencies may act on their own initiative159 or whether they act on the initiative of another actor.160 This is especially important in the light of Majone’s argument on ‘regu-lation by information’ (cf. section I 1.2.1.1).

Griller and Orator’s pre- decision- making agencies are subsumed under this first category, but are also further differentiated. Griller and Orator, citing Geradin and Petit,161 argue that these agencies ‘enjoy a considerable influence over the adoption of final decisions by the Commission’.162

On this point a first differentiation should be made turning on the nature of the opinion. Here the agency may have been empowered to give a ‘noncommittal’ opinion163 or an authoritative opinion which comes with strings attached, that is, because legislation prescribes that the authority adopting the final decision is under a special obligation to take account of the opinion.164 A second differenti-ation among these agencies is whether they are involved in the process leading to an individual decision or whether the process leads to the adoption of a generally applicable decision. A third differentiation relates to the scope of the agency’s in-volvement. The agency may thus be empowered to give a simple opinion on (part) of a measure, or the agency may be empowered to propose the draft measure

159 Article 5 of the ACER Regulation for instance provides that ‘[t] he Agency may, upon a re-quest of the European Parliament, the Council or the Commission, or on its own initiative, provide an opinion or a recommendation [… ] on any of the issues relating to the purpose for which it has been established’ (emphasis added).

160 An example of this may be found in Article 2(d) of the EEA Regulation which provides that the agency may ‘advise individual Member States, upon their request […] on the development, estab-lishment and expansion of their systems for the monitoring of environmental measures, […]; such advice may also include peer reviews by experts at the specific request of Member States’ (emphasis added).

161 Geradin and Petit, n 137, p 183.162 Griller and Orator, n 46, p 13.163 Tridimas, citing Craig, has referred to the EASA as a quasi- regulatory agency because of

its ‘strong recommendatory power’ when it drafts rules for the Commission to be adopted under a comitology procedure. See Tridimas, n 50, p 235; Craig, n 135, p 155. Article 19(1) of the EASA Regulation, however, only provides that the agency may submit its drafts to the Commission as opinions. The EASA Regulation itself therefore does not lay down a special obligation on the Commission to adhere to these opinions.

164 In the regime for GMO authorization, the Commission prepares a draft decision under a comitology procedure based on the opinion of the EFSA. If its draft decision deviates from the opinion, the regulation prescribes that ‘the Commission shall provide an explanation for the dif-ferences’. See Article 7(1) of Regulation (EC) 1829/ 2003 of the European Parliament and of the Council, OJ 2003 L 268/ 1. Differently, when another Member State disagrees with the reference Member State’s assessment report of a medicinal product under the decentralized procedure, the Commission will take a decision under a comitology procedure following a scientific opinion by the EMA. The relevant directive thereby provides: ‘Where, exceptionally, the draft decision is not in accordance with the opinion of the Agency, the Commission shall also annex a detailed explan-ation of the reasons for the differences’ (emphasis added). See Article 33 of Directive (EC) 2001/ 83 of the European Parliament and of the Council, OJ 2001 L 311/ 67. The ESA Regulations probably curtail the Commission’s power to adopt a final decision to the greatest extent; Articles 10 of the Regulations provide the Commission may adopt delegated acts, called regulatory technical stand-ards, based on the ESA’s opinions. The recitals in the preambles to the Regulations provide that these draft opinions should as a rule not be amended by the Commission. See n 210.

Setting the Scene26

itself. Similarly, instead of being involved in the process leading to a decision, an agency could also be empowered to provide authoritative interpretations of deci-sions (possibly adopted by another authority) or it could adopt ‘guidelines’ related to these norms.165

The same gradation exercise may be done for a second category, the ‘oper-ational agencies’ represented graphically in Figure 1.3. These agencies are re-ferred to as ‘operational’ because their powers are hard to qualify as binding or non- binding. The operational powers allow them to facilitate operational cooperation between the Member States or they allow the EU to have its own operational capacity. When it comes to joint operations by the Member States, agencies may coordinate these (in a first step) or even organize them (in a second step). A final step is the power for an agency to set up its own operations inde-pendent from the Member States. The same differentiation may be made for another important operational task exercised by agencies, that is, the power of inspection or monitoring. An agency may thus perform these tasks at the request of another authority (Commission, Member State) or it may perform them on its own authority.166

165 Apart from the opinions under Article 19(1) of the EASA Regulation (cf. n 163), the EASA is also empowered under Article 19(2) to adopt ‘certification specifications and acceptable means of compliance and guidance material’ to be used when the EASA exercises its powers under the certi-fication process, that is, when it adopts individual decisions on applications.

166 Another situation may be noted in the pharmacovigilance directive, which does not provide inspection powers to the EMA, but does provide that the agency may request inspections to be car-ried out on manufacturers or distributors of active substance by the national authorities. See Article 111 of Directive (EC) 2001/ 83, OJ 2001 L 311/ 67.

4. Organizingoperations/Organizing inspections

3. Organizing joint operations2. Coordinating joint operations/carrying

out inspections

1. Information exchangeOperational agencies

Figure 1.3 Hierarchy of agencies’ operational powers


10. Final general decision9. Final individual decisionDecision-making agencies

8. Drafting of general decision7. Authoritative interpretation of statutes/

Authoritative guidelines6. Opinion on general decision

5. Drafting of individual decision4. Authoritative opinion on individual decision

3. Noncommittal opinion on individual decisionMandatory involvement of agencies

2. Pro-active exercise of powers, possibility of acting own initiative1. Passive exercise of powers

Non-decision-making agencies

Figure 1.4 Hierarchy of agencies’ (non- )decision- making powers

The third category consists of the decision- making agencies, bodies ultimately responsible (de iure) for the adoption of the final binding decision. Again one can differentiate here between the power to adopt individual decisions and norma-tive acts. A general decision may be any material legislation (including delegated acts under Article 290 TFEU) or even international agreements. While the non- decision- making agencies’ and the decision- making agencies’ powers have been presented in Figure 1.4 in a pyramidal structure to express the idea of ‘intensifying’ powers, it should be clear that the power to draft general decisions is often more significant than the power to adopt a binding individual decision.

In 2010 Griller and Orator noted that their fourth category, the rulemaking agency, was a theoretical one, since no such agency had been established at EU level.167 For instance, although Europol can conclude international agreements on its own behalf with international organizations and third countries (cf. section II 3.4.3), this power is subject to the Council’s approval.

In late 2010, however, the ESMA was also established, its regulation contain-ing an interesting enabling clause in Article 9(5) which reads: ‘The Authority may temporarily prohibit or restrict certain financial activities […] in the cases speci-fied and under the conditions laid down in the [body of EU financial legislation] or if so required in the case of an emergency situation in accordance with and under the conditions laid down in Article 18.’ Since the existence of an emergency

167 Griller and Orator, n 46, p 14.

Setting the Scene28

situation can only be established by the Council under that Article 18, the second possibility did not result in an independent power of the ESMA.

This changed with the adoption of Regulation 236/ 2012 on short selling.168 This regulation added to the body of financial legislation and provides additional powers for the ESMA, granting it normative powers in exceptional circumstances. Although the power thus provided is still an exception to the rule (ie national au-thorities being competent), the ESMA itself now decides whether it may act or not. Indeed, Article 27(3) explicitly provides that in a case of disagreement be-tween the ESMA and a national authority, the ‘ESMA shall consider whether the conditions are satisfied and it is an appropriate case for the use of its powers of intervention under Article 28’. The ESMA may therefore well act as a rulemaking agency, although most of the time its powers are less intense. This power of the ESMA was contested but sanctioned in Short- selling (cf. Chapter IV).

1.2.1.2.1 The instrumental classification as a tool for assessing the constitutional limits to agencification

As was already noted previously, an instrumental classification of the powers granted to agencies is in itself not sufficient to assess whether that grant of power respects the constitutional limits as set out by the Treaties.

These limits are shaped by a combination of the principle of conferred powers, the principle of proportionality, the principle of subsidiarity, and the institutional balance. It is therefore not possible to simply assess the powers granted to agen-cies to make conclusions as to the constitutionality of their empowerment. This may be illustrated by a simple example: if the constitutional limits could be de-duced purely from an instrumental approach, the power to adopt an individual decision would be assessed identically in every situation. In the early 1990s, there were already agencies that had been granted the power to adopt individual de-cisions: the OHIM and CPVO decide on applications for intellectual property rights. Although the granting of such powers could be contested,169 it has not been the subject of much controversy,170 let alone being contested on constitu-tional grounds. This may be juxtaposed with the much more contested debate on a European Cartel Office which could take over the Commission’s role as the EU’s competition authority.171 The reason here is straightforward: Article 105 TFEU explicitly empowers the Commission to enforce the rules laid down in Articles 101

168 Regulation (EU) 236/ 2012 of the European Parliament and of the Council, OJ 2012 L 86/ 1.169 See for instance Chamon, n 51, pp 294– 5.170 There has been some critique on other grounds such as the specific language regime of the

OHIM, see Case T- 120/ 99, Kik v. OHIM, [2001] ECR II- 2235 (upheld on appeal).171 The arguments in favour of such a competition agency have mainly come from Germany.

For one of the earliest suggestions, see Klaus Holderbaum, ‘Chancen für eine europäische Kartellbehörde’, (1967) 2 Europarecht 2, pp 116– 33. The German government even put the issue on the table during the negotiations of the 1996 IGC. Karel Van Miert, then Commissioner for Competition Policy, was vehemently opposed to this proposal. See Karel Van Miert, ‘The Proposal for a European Competition Agency’, (1996) 2 Competition Policy Newsletter 2, pp 1– 4. For the original German proposal, see Conference of the Representatives of the Governments of the Member States, Creation of Restrictive Practices Office, 22 May 1996, CONF 3855/ 96.


and 102 TFEU, whereas a similar provision in primary law is lacking as regards the powers granted to the OHIM and CPVO.172

Apart from this straightforward example, the outcome of the assessment may also depend on the policy field concerned and whether it comes under an exclu-sive, a shared, or a supplementing competence of the EU.173 Next to this material requirement, formal requirements also play a role. Compare for instance the atyp-ical acts adopted by the OHIM and the CPVO with the delegated and imple-menting acts under Articles 290 and 291 TFEU. The latter Articles prescribe the different actors which may be involved in adopting those acts (Commission, Member States, Parliament, Council) and since the agencies are not mentioned in these Articles, the limits on their involvement should be different from the limits in those cases where comparable express provisions in primary law do not exist. Insights gathered from a functional classification of EU agencies may also be used here, since the assessment of agencies assisting rather than informing national administrations or assisting national administrations rather than assist-ing EU institutions will also be different, for instance under a subsidiarity and/ or proportionality perspective (cf. sections IV 2– 3).

1.2.1.3 A typology of agencies’ powers instead of agenciesThe problem in neatly categorizing agencies in any instrumental classification is that agencies rarely have one single power. As a result, one and the same agency may, for example, act as a decision- making agency at one time and as an oper-ational agency the next.174 Because of this it is more elucidating to look at some of the powers which have been conferred on agencies.

1.2.1.3.1 Non- decision- making agenciesThis first group of agencies do not adopt binding acts but are involved in the adoption of soft law, that is, ‘instruments which have not been attributed legally binding force as such, but nevertheless may have certain (indirect) legal effects, and that are aimed at and may produce practical effects’.175 As will be discussed presently, some of these instruments may have more legal and practical effects than others.176

a. Drawing up reports Probably the most modest power is that of presenting reports, opinions, studies, etc. The significance of this power is that the agency is (supposed to be) the EU’s centre of expertise in its field, giving some weight to its findings.

172 See also the opinion of the Council Legal Service on the OHIM proposal, section IV 5.4.3.2.173 See Articles 3– 6 TFEU.174 See also Merijn Chamon, ‘The Influence of “Regulatory Agencies” on Pluralism in

European Administrative Law’, (2012) 5 REALaw 2, p 71.175 See Linda Senden, Soft Law in European Community Law, Oxford, Hart Publishing,

2004, p 112.176 Alberti discussed the instruments adopted by EU agencies in a similar way, making a distinc-

tion between (i) preparatory acts for binding measures, (ii) guidelines, (iii) general recommenda-tions, and (iv) reports and best practices. See Jacopo Alberti, ‘L’utilisation d’actes de soft law par les agences de l’Union européenne’, (2014) RUE 576, pp 163– 6.

Setting the Scene30

As it is, all agencies which fulfil an information function (ie virtually every agency) will exercise this power. An exception to this is the service agencies. The significance of the present power will largely depend on the factual context, but it may still be interesting to verify an agency’s independence in its information task. Is an agency free, within its mandate, to gather information? Does it have to rely on already existing information sources or is it equipped or resourced to conduct its own studies? Does it (legally) depend on other actors to present information or may it present and disseminate information on its own initiative, etc?

The information function exercised by agencies becomes genuinely interesting when it is lifted out of the noncommittal sphere by linking the agency’s action with a legal effect (however modest), creating a ‘power’ on the part of the agency.

b. Non- committal opinion on an individual decision Under this power, an agency may address opinions to a second authority for the purposes of adopting an individual decision, even if there is no formal obligation on the latter authority to take account of the agency’s opinion. In addition, unless the agency is virtually the sole knowledgeable organization, there is no de facto obligation on the com-petent authority to take the opinion into account either.

Two examples serve to illustrate. A first is the interplay between competition policy and economic regulation. While no Cartel Office has been established, there are EU agencies in the energy, financial, telecoms, railway, air transport, etc sectors. While efficiency considerations dominate competition policy, the goals of economic regulation are broader,177 raising the question how possible tensions be-tween these fields may be resolved.178 Without entering into this debate, it is clear that the agencies’ expertise may be beneficial to competition authorities who need to have a proper understanding of the markets in which they scrutinize (anti- )competitive behaviour. Examples may be found in the ACER’s power to submit opinions on barriers to the completion of the internal market.179

Secondly, the FRA is empowered to adopt opinions on the topics within its field of action. Interestingly, the FRA Regulation further provides that the FRA’s opinions will not interfere with the procedures in Articles 258 and 263 TFEU, implicitly acknowledging that the FRA’s reports could identify a failure to fulfil an obligation under EU law on the part of a Member State.180 Similarly, pro-visions in the EASA and the ESAs Regulations may be found, protecting the

177 See Okeoghene Odudu, ‘The Wider Concerns of Competition Law’, (2010) 30 Oxford Journal of Legal Studies 3, p 602.

178 See José Carlos Laguna de Paz, ‘Regulation and Competition Law’, (2012) 33 European Competition Law Review 2, pp 77– 83.

179 For instance, in the section on gas wholesale market integration of its 2011 report, the ACER noted certain non- EU gas producers applied a ‘divide and rule’ strategy in the EU internal market. The European Commission began proceedings to open a competition investigation against Gazprom on 4 September 2012. See ACER, ACER/ CEER Annual Report on the Results of Monitoring the Internal Electricity and Natural Gas Markets in 2011, 29 November 2012, p 157.

180 This restriction of the FRA’s mandate was already foreseen in the Commission’s original pro-posal. See European Commission, COM (2005) 280 final. Different amendments were proposed by MEPs but none of them made it to the final text. See Amendments 72- 193 to the Draft report (PE 369.836v02- 00) by Kinga Gál, 5 April 2006, Committee on Civil Liberties, PE 370.083v02- 00.


Commission’s discretion to start proceedings under Article 258 TFEU, even if the agencies concerned have identified infringements of EU law.181

The power to adopt noncommittal opinions may seem insignificant at first sight, since the competent authority is under no obligation to even take account of the opinion. However, the above examples indicate that in reality the situation is less clear. Some agencies have been established to ‘monitor’ actions or inactions of other actors. Not taking account of such monitoring report then undermines the raison d’ être of these agencies. This idea is apparent in the FRA regulation, whereby the importance of the FRA’s reports is implicitly recognized by empha-sizing the Commission’s constitutionally enshrined independence in infringe-ment proceedings.182 A second element is the status of the agency. Comparing the illustrations given previously, whereas a number of other high- profile human rights institutions exist in Europe apart from the FRA, this is much less the case for organizations with expertise comparable to the EASA and the ESAs. Although the legal value of these bodies’ opinions and reports is not altered by this element, in practice the Commission may find it much more difficult to ignore the latter agencies’ findings.

c. Authoritative opinion on an individual decision The authoritative quality of an opinion results from the unique position of the agency in providing part of the information necessary to adopt the final decision. Applied to the examples given presently: the EMA and EFSA opinions are authoritative because they are the primary European experts in the fields concerned,183 and their position is further reinforced by a legal obligation on the Commission to properly argue its decision when it does not follow the advice of the agencies. Still, authoritative opinions are ‘merely’ authoritative and may often not deal with every aspect of the final de-cision, as is the case for the examples given presently, since the agencies are only responsible for risk assessment and not risk management.

The EMA provides the Commission with scientific opinions on applications for authorization of medicines under the centralized procedure.184 Article 10 of the Regulation prescribes that following the adoption of the final opinion by EMA, the Commission prepares a draft decision within fifteen days. The opinion is

181 See Article 54(1) EASA and recitals 31 of the preambles to the EBA and ESMA Regulations and recital 30 of the preamble to the EIOPA Regulation.

182 See Article 4(2) of the FRA Regulation.183 Although no other European centres of expertise exist, relevant expertise will usually still

exist at national level. As regards the EFSA, Alemanno notes that the agency ‘has not been entrusted with the power to act as the ultimate body of scientific advice in the EU.’ See Alberto Alemanno, ‘Food Safety and the Single European Market’, in Ansell and Vogel (eds), What’s the Beef? The Contested Governance of European Food Safety, Cambridge, MIT Press, 2006, p 250. This is different for agencies such as the ACER and ESAs, which do act as such ultimate bodies.

184 The EMA Regulation makes the centralized procedure for authorization mandatory for the types of medicine referred to in its annex. For medicines containing an active substance which was not yet authorized in 2004 and for certain types of innovative medicinal products the procedure is optional. For other products, the decentralized or mutual recognition procedure may be used. These two procedures are provided for in Directive 2001/ 83, the decentralized procedure being introduced by the amending Directive 2004/ 27. See Directive (EC) 2001/ 83, OJ 2001 L 311/ 67.

Setting the Scene32

non- binding, but the Regulation provides that ‘[w] here the draft decision differs from the opinion of the Agency, the Commission shall attach a detailed explan-ation of the reasons for the differences’.

The EFSA has been granted the power of opinion under a number of legis-lative acts.185 By way of example, Article 9 of the Regulation on materials in-tended to come into contact with food provides that an application should be submitted to the national competent authority which forwards the information to the EFSA. According to Article 10 the EFSA adopts its opinion on the ap-plication within six months. If the Commission does not follow the EFSA’s opinion, the Commission must ‘provide without delay an explanation for the reasons for the differences’.

The way in which the provisions on the advice of the EMA and EFSA are drafted shows that the Commission is expected to follow the advice of the agen-cies and that it should otherwise properly motivate its decision.186 Ultimately, however, the Commission is free to take a different decision than the one advo-cated by the agency.187

The Regulation on biocidal products188 illustrates clearly (i) that an agency’s mandate is not simply defined in its establishing regulation and (ii) that agencies’ powers were often exercised at national level prior to the agency being created (cf. section III 1). Under the previous regime on the authorization of biocidal products,189 Member States granted such authorizations under a mutual recog-nition regime. In cases of disagreement, the Commission would settle the dis-pute by adopting a decision pursuant to a comitology procedure. The new regu-lation leaves the national procedure for authorization intact but also provides for a Union authorization. Such applications are submitted to the ECHA, even if the actual evaluation is still undertaken by a national authority. Following the receipt of a complete and valid application, the national authority has 365 days to make an evaluation and propose its findings to the ECHA. Based on these findings the ECHA will prepare a report for the Commission, which in turn will decide whether or not to authorize the product under the examination comitology procedure. The ECHA also acquires a role in the resolution of disputes between Member States under the mutual recognition procedure. The Commission is still competent to resolve these disputes, but in order to do so it can now ask the

185 See inter alia Regulation (EC) 1831/ 2003 of the European Parliament and of the Council, OJ 2003 L 268/ 29; Regulation (EC) 2065/ 2003 of the European Parliament and of the Council, OJ 2003 L 309/ 1; Regulation (EC) 1935/ 2004 of the European Parliament and of the Council, OJ 2004 L 338/ 4; Regulation (EC) 1924/ 2006 of the European Parliament and of the Council, OJ 2006 L 404/ 9; Regulation (EC) 1331/ 2008 of the European Parliament and of the Council, OJ 2008 L 354/ 1; Regulation (EC) 1069/ 2009 of the European Parliament and of the Council, OJ 2009 L 300/ 1; Regulation (EC) 1107/ 2009 of the European Parliament and of the Council, OJ 2009 L 309/ 1.

186 However, even in absence of a precise provision this obligation exists. See Case T- 13/ 99, Pfizer Animal Health SA v. Council, [2002] ECR II- 3318, paras 196– 9.

187 See also Marco Zinzani, Market Integration Through ‘Network Governance’: The Role of European Agencies and Networks of Regulators, Antwerp, Intersentia, 2012, p 41.

188 Regulation (EU) 528/ 2012 of the European Parliament and of the Council, OJ 2012 L 167/ 1.189 Directive (EC) 98/ 8 of the European Parliament and of the Council, OJ 1998 L 123/ 1.


ECHA to make a scientific assessment of the claims in dispute.190 As a result, the ECHA acquires both an important role in the Union authorization procedure, which has been constructed as a complex and balanced composite procedure,191 as well as under the national authorization procedure.

Lastly, a special type of individual decision is the imposition of a financial pen-alty. Two agencies have been granted a role in procedures whereby the Commission may impose fines on undertakings for non- compliance with the material rules in their respective fields. Article 84(3) of the EMA Regulation and Article 25 of the EASA Regulation both provide that the Commission may impose financial penal-ties at the request of the agency.192 Both Regulations provide that the Commission should adopt further implementing measures to establish a framework.193 Even if the agencies’ requests are authoritative, the Regulations leave the formal discretion to impose sanctions to the Commission by providing that it may, rather than shall, act on the request of the agency.194

d. Opinion on general decision Some agencies have been granted the power to submit opinions in the decision- making process leading to general decisions. This category is still quite diverse, since the final general decision may for example be a legislative, delegated, or implementing act. For the latter two, the agency’s opinion will be submitted directly to the responsible authority (the Commission), but the opinions for legislative acts will only be opinions leading to the Commission’s proposal, which is then taken under consideration by the Parliament and Council.

Although the EMSA’s tasks are predominantly focused on assisting the Member States and the Commission at an operational level, its establishing regulation also provides that it will assist the Commission ‘in the preparatory work for up-dating and developing relevant legal acts of the Union, in particular in line with the development of international legislation’.195 The 2013 revision of the EMSA Regulation196 rephrased the earlier provision which only referred to the ‘updating and developing of Community legislation’, thus clarifying that the EMSA may provide input for all kinds of legal acts, and not just (formal) legislation.

190 The ECHA’s (subsidiary) role in solving disputes between (national authorities of) Member States would suggest that unlike the EFSA (cf. n 183), the ECHA is the ultimate body of scientific advice in this field.

191 The ReNEUAL Model Rules on EU Administrative Procedure define the composite pro-cedure as follows: ‘an administrative procedure where EU authorities and the authorities of a Member State or of different Member States have distinct functions which are inter- dependent.’ Further on the notion of composite administrative procedures, see section I 1.2.1.3.4.

192 See also Article 49(3) of Regulation (EC) 1901/ 2006 of the European Parliament and of the Council, OJ 2006 L 378/ 1.

193 For these implementing rules, see Implementing Regulation (EU) 646/ 2012 of the European Commission, OJ 2012 L 187/ 29; Regulation (EC) 658/ 2007 of the European Commission, OJ 2007 L 155/ 10.

194 At the moment, only the EMA has used this power. In 2012 it started a procedure against Roche, adopting its (confidential) final report for the Commission in April 2014. See EMA Press Office, EMA/ 211922/ 2014.

195 See Article 2(2)a EMSA Regulation. A similar provision regarding the EFSA may be found in Article 5 of the Regulation on food information. See Regulation (EU) 1169/ 2011 of the European Parliament and of the Council, OJ 2011 L 304/ 18.

196 See Regulation (EU) 100/ 2013 of the European Parliament and of the Council, OJ 2013 L 39/ 30.

Setting the Scene34

Even if an agency’s mandate does not explicitly provide that it will be involved in the preparation of further legislation in its sector, the agency will often still be involved. Under its mandate the ERA only assists the Commission in adopting measures implementing the existing framework. However, the Commission Staff Working Document on the proposals for the fourth railway package refers to a number of ERA studies, including a preliminary assessment on the impact on the ERA conducted by the ERA itself.197

The power to adopt opinions leading to general decisions may vary. Some agen-cies’ mandates explicitly provide that they should be involved by the Commission when new legal acts are drafted; other agencies are naturally involved because they are among the few bodies of expertise; still other agencies may indirectly influence the Commission’s proposals through their general reports and opinions.

e. Authoritative interpretation of statutes and authoritative guidelines A number of different powers were all subsumed under this single heading because their exercise results in non- binding but authoritative acts of general scope. This is the case when agencies have the power to adopt ‘guidelines’. An example may be found in the regulation on materials which are intended to come into contact with foodstuffs, where the EFSA is tasked to adopt guidelines on the submission of applications.198 The time period for the EFSA to conclude its review starts to run when a ‘valid application’ is submitted. Although the guidelines cannot be deter-minative for deciding when an application is valid, it is clear that market operators will have reasons to accommodate to the guidelines to ensure a swift procedure.

Further obligations may also be imposed because guidelines may be broader in scope than the relevant provisions in secondary legislation. Sometimes this is even explicitly recognized, as is the case for the EBA guidelines on the assessment of the suitability of members of the management body and key function holders of financial institutions, which provides: ‘in order to achieve the necessary and desir-able degree of harmonisation in this area these Guidelines are deliberately broader in scope than Article 11 of Directive 2006/ 48/ EC.’199 On this point it should be noted that such additional rules should typically be adopted by implementing meas-ures under Article 291 TFEU or delegated acts under Article 290 TFEU. Adopting them as such would mean they would be de iure binding and therefore challenge-able before the CJEU, which is not the case if one relies on mere ‘guidelines’.

The EMA guideline on the procedure for adopting EMA guidelines is rather explicit on the possible binding effect of its guidelines. While it first acknowledges that ‘guidelines do not have legal force and the definitive legal requirements are those outlined in the relevant Community legislative framework’,200 it continues to

197 European Commission, SWD (2013) 8 final.198 See EFSA Panel on Food additives flavourings processing aids and materials in contact with food

(AFC), Guidance Document on the Submission of a Dossier on a Substance to be Used in Food Contact Materials for Evaluation, 30 July 2008, http:// www.efsa.europa.eu/ en/ efsajournal/ doc/ 21r.pdf.

199 EBA, Guidelines on the Assessment of the Suitability of Members of the Management Body and Key Function Holders, 22 November 2012, EBA/ GL/ 2012/ 06, p 5 (emphasis added).

200 EMEA, Procedure for European Union Guidelines and Related Documents within the Pharmaceutical Legislative Framework, 18 March 2009, EMEA/ P/ 24143/ 2004 Rev. 1 corr, p 4.

http://www.efsa.europa.eu/en/efsajournal/doc/21r.pdf

http://www.efsa.europa.eu/en/efsajournal/doc/21r.pdf


note that ‘guidelines are to be considered as a harmonised Community position’.201 When it comes to the possibility to deviate from these guidelines the paper pro-vides that ‘alternative approaches may be taken, provided that these are appropri-ately justified’.202 It is worth stressing here that while the non- binding character of EMA’s soft law is acknowledged by the agency itself, it still requires parties to justify their intended non- compliance with such non- binding guidance.203 These examples illustrate the use of soft law by agencies and raise the general problem of the use of soft law in regulation.204

Lastly, it should be noted that the decisions of agencies’ Boards of Appeal may also be authoritative. Although these decisions are binding, they do not have erga omnes effect. Still, since the Board of Appeal will often interpret relevant EU le-gislation, its decisions will have further repercussions. Its interpretations would obviously never be binding on the Court, but they will inform the Board’s own understanding and that of the agency of relevant legislation.205

f. Drafting general decisions As was pointed out previously, it may be hard in practice to draw the line between the power to draft a decision and the power to give an opinion. Drafting a decision means that an agency is empowered to pre-pare the decision in full and that it is left to another authority to formally adopt the act. The discretion left to that authority to amend the draft may be further restricted by statutory provision, but is any case restricted because the ratio in em-powering an agency to prepare drafts is that the expertise is housed in the agency and not in the (ultimate) competent authority.

Article 18(a) of the EASA Regulation provides that the agency may issue opin-ions addressed to the Commission, whereby Article 19(1) EASA clarifies that ‘[i] n order to assist the Commission in the preparation of proposals for basic principles, applicability and essential requirements to be presented to the [legislator] and the adoption of the implementing rules, the Agency shall prepare drafts thereof ’. Based on this description of the EASA’s power, these drafts may still only be opin-ions, especially when they concern formal legislation. In the case of the imple-menting rules, however, this is different. Even if there is no formal obligation on the part of the Commission to take account of these drafts, the technical nature of the implementing rules in the aviation sector will have a significant impact on the Commission’s margin of discretion. To deviate from the EASA’s draft, the

201 Ibid, p 4. 202 Ibid, p 5.203 The EMA further rejected the suggestion that non- compliance with the guidelines should

not be subject to a justification. See EMEA, Overview of Comments Received on Draft Guidelines Procedure for EU Guidelines and Related Documents within the Pharmaceutical Legislative Framework, 24 June 2005, EMEA/ 125817/ 2004, p 5.

204 On the interplay between the two new modes of governance of soft law and agencification, see Merijn Chamon, ‘Le recours à la soft law comme moyen d’éluder les obstacles constitutionnels au développement des agences de l’UE’, (2014) RUE 576, pp 152– 60.

205 In principle decisions of the Boards of Appeal do not set precedents, even if this has been sug-gested for the OHIM Boards of Appeal. See Max Planck Institute for Intellectual Property and Competition Law, ‘Study on the Overall Functioning of the European Trade Mark System’, (15 February 2011), http:// ec.europa.eu/ internal_ market/ indprop/ docs/ tm/ 20110308_ allensbach- study_ en.pdf, p 37.

http://ec.europa.eu/internal_market/indprop/docs/tm/20110308_allensbach-study_en.pdf




Setting the Scene36

Commission would need to convince the members of the EASA comitology com-mittee that its alternative is better than that of the experts at the agency.206

Another example may be found in the ERA regulation. Article 6 provides that the ERA ‘shall recommend to the Commission the common safety methods and the common safety targets’ foreseen in the Railway Safety Directive.207 Articles 6 and 7 of that directive provide that the Commission will adopt a mandate, under a comitology procedure, for the ERA to work out drafts of these measures which are then recommended by the ERA to the Commission, which adopts the final versions subject to a comitology procedure. The Commission may still make changes to these recommendations before proposing them to the comitol-ogy committee, but the bulk of the final decision is contained in the agency’s rec-ommendation.208 Here as well the agency may at times recommend that its own powers be extended.209

Drafting a general decision may also be done through different stages, involving different actors, resulting in a somewhat complex procedure, as is illustrated by the procedure to adopt network codes for the electricity and gas markets. Articles 6 of both Regulations 714/ 2009 and 715/ 2009 set out a cascading procedure for their adoption by the Commission, based on framework guidelines normally adopted by the ACER. The codes in turn are normally drafted by the relevant ENTSO. In this integrated (cascading) procedure, the ACER assists both the Commission and the ENTSO and the Commission scrutinizes the ACER, which in turn scrutinizes the ENTSO. If one of the authorities fails to fulfil its role under the procedure, its role is assumed by the authority next in line.

The power to draft general decisions has provisionally culminated in the powers granted to the ESAs to work out draft regulatory and implementing technical standards. The drafts of the implementing technical standards may only become final following a final decision by the Commission under a comitology procedure. The regulatory technical standards are delegated acts under Article 290 TFEU. The ESAs’ Regulations provide that the Commission should endorse the drafts and can only amend them in case of incompatibility with EU law, disproportion-ality, or if they go against the principles of the EU’s acquis in financial services

206 On unique occasions, this aspect of the EASA’s rulemaking powers even receives significant attention in popular media. This was the case when the EASA submitted a draft to regulate pilots’ flight times which was mediatized before the matter was discussed in the (comitology) Committee for the application of common safety rules in the field of civil aviation. See Opinion No 04/ 2012 of the European Aviation Safety Agency, 28 September 2012.

207 Directive 2004/ 49 of the European Parliament and of the Council, OJ 2004 L 164/ 44.208 Compare for instance the ERA’s recommendation on a common safety method on risk evalu-

ation and assessment and the final Commission Regulation. The most significant changes may be found in the ordering of the provisions and their inclusion in the body or the annex to the Regulation. See Recommendation 02/ 2007 of the European Railway Agency; Regulation (EC) 352/ 2009 of the European Commission, OJ 2009 L 108/ 4.

209 For instance, ERA’s recommendation to give it the power to adopt technical opinions on the information given by Member States following enforcement actions by the Commission. This suggestion was later included in the Commission’s final decision. See Recommendation 01/ 2008 of the European Railway Agency; Decision (EC) 2009/ 460 of the European Commission, OJ 2009 L 150/ 11.


legislation. Further, if the Commission changes the draft, it should first consult the ESA.210

These restrictions on the Commission’s competence under Article 290 TFEU were unprecedented. It should be clear that these requirements go much further than the typical requirement imposed on the Commission to ‘take account’ of an agency opinion. Under Articles 10 of the ESAs’ Regulations, the Commission may only adopt delegated acts on its own if the ESAs have refused to provide the Commission with a draft within the time limits prescribed by the legislator.

The procedure under Articles 15 for adopting implementing technical stand-ards under Article 291 TFEU is identical to the procedure of Articles 10 and the logic behind the procedure is very much the same: if the ESAs have worked out drafts, the Commission should not alter them any more save ‘where the Union’s interests so require’.211 And even in such cases, ‘[t] he Commission shall not change the content of a draft implementing technical standard prepared by the Authority without prior coordination with the Authority’.212

To prevent these restrictions on its role constituting a precedent for future legis-lation, the Commission issued a statement noting it had ‘serious doubts whether the restrictions on its role when adopting delegated acts and implementing meas-ures are in line with articles 290 and 291 TFEU’.213 Subsequently, however, the legislator has taken further steps by including certain provisions not only in the recitals of legislative acts but also in the actual provisions.214

1.2.1.3.2 Decision- making agenciesa. Adopting individual decisions The power to adopt individual decisions

may be passive, in the sense that the agency should first be asked to adopt a decision (OHIM, CPVO, EASA), or active, in the sense that the agency may adopt such decisions on its own initiative (ACER, ESAs). In addition, there is the question whether the power granted to an agency is concurrent to the power held by another actor. For the early agencies this was not the case, as for instance the OHIM grants an EU trademark, while national trademarks are granted by national authorities under the trademark Directive.215 The establish-ment of the ESAs has changed this. The ESAs may adopt individual decisions

210 See recitals 23 of the preambles to the EBA and ESMA Regulations and recital 22 of the pre-amble to the EIOPA Regulation.

211 See Articles 15(1) of the ESA Regulations. 212 Ibid.213 See the Commission’s statements in the addenda to Council documents: 15649/ 10, ADD 1, 10

November 2010; 15647/ 10 ADD 1, 10 November 2010 and 15648/ 10 ADD 1, 10 November 2010.214 See Article 82(3) of Regulation (EU) 648/ 2012 of the European Parliament and of the

Council, OJ 2012 L 201/ 1. For the Commission’s opposition, see European Commission, ‘Communication on the actions taken on opinions and resolutions adopted by Parliament at the March I and II 2012 part- sessions’, SP (2012) 323.

215 See Directive (EC) 2008/ 95 of the European Parliament and of the Council, OJ 2008 L 299/ 25. Of course, there still is some competition between the OHIM and the national authorities, since operators have a choice between filing for national and EU protection. Because of this national trademark offices, through the relevant comitology committee, had originally set the fees payable to the OHIM at an excessive level, making the (lower) national fees more attractive. This results in

Setting the Scene38

vis- à- vis financial institutions which normally come under the jurisdiction of national regulators. The three situations foreseen in the establishing regulations are a breach of EU law by a national authority,216 emergency situations,217 and when national authorities disagree in cross- border situations.218 Additionally, further legislation may elaborate an ESA’s mandate, such as the legislation at issue in Short- selling.

If there is a breach or a non- application of relevant EU legislation, the ESA may only adopt decisions vis- à- vis financial institutions if a national authority refuses to comply with the Commission’s formal opinion. In the case of an emergency situation, the ESA will receive the same powers, but the Council should first de-clare the existence of an emergency situation. The procedures are similar in the last case when a disagreement exists between national authorities. In such a case they may request the competent ESA, or in some specific cases the ESA may act itself,219 to act as a mediator in a ‘conciliation phase’. If no agreement may be found between the national authorities, the ESA will be able to step in and issue binding decisions.

As noted above, a special kind of decision is the decision to impose financial sanctions. Whereas the EASA and EMA may make requests to the Commission to impose such sanctions (discussed previously), the ESMA has been granted full powers in this respect under Regulation 648/ 2012 relating to OTC derivatives, central counterparties, and trade repositories,220 and Regulation 1060/ 2009 on credit rating agencies.221 Under both Regulations the Commission had origin-ally proposed a scheme identical to the one applicable to the EASA and EMA,222 but on both occasions the Parliament insisted on granting the ESMA this power directly,223 the Commission inter alia agreeing in light of the particularity of the financial services sector.

The fact that the ESMA has been granted power to fine undertakings is inter-esting since the same power was refused to the ACER by the Commission and the Council, following a suggestion to this end by the Parliament in order to ensure the ACER would have ‘effective enforcement mechanisms’.224 The Commission

huge surpluses booked by the OHIM since the fees were set at a significantly higher level than the actual costs incurred by the OHIM. See Madalina Busuioc, European Agencies: Law and Practices of Accountability, Oxford, OUP, 2013, p 174.

216 See Article 17 of the ESA Regulations.217 See Article 18 of the ESA Regulations.218 See Article 19 of the ESA Regulations.219 See Article 19(1) second sentence of the ESA Regulations.220 Regulation (EU) 648/ 2012 of the European Parliament and of the Council, OJ 2012 L 201/ 1.221 Regulation (EC) 1060/ 2009 of the European Parliament and of the Council, OJ 2009

L 302/ 1.222 See Article 36a in European Commission, COM (2010) 289 final; Article 55 in European

Commission, COM (2010) 484 final.223 European Commission, ‘Communication on the action taken on opinions and resolu-

tions adopted by the Parliament at the December 2010 part- session’, SP (2011) 1477; European Commission, n 214.

224 See justification given for Amendment 42 in the Report by MEP Giles Chichester, 4 June 2006, A6- 0226/ 2008 PE400.717v02- 00.


agreed to the possibility of imposing financial sanctions, but only if this power was provided to the Commission in consultation with the ACER rather than the other way around. The reason for this, according to the Commission, was that the Meroni doctrine prevented the legislator from granting such powers to an agency.225 The Council on the other hand did not like the idea of financial sanc-tions at all,226 and in the end the ACER was established without a provision on such ‘enforcement mechanisms’.

b. Adopting general decisions The power to adopt general decisions is arguably the most far- reaching power, since it allows normative acts to be adopted. Within the EU legal order, this power is normally reserved to the legislator, the Council, and because of its delegated and implementing acts the Commission also has an important (normative) function.227 As a result, and because the Treaties lack an enabling clause for establishing agencies, EU agencies with a normative compe-tence would seem impossible under current primary law.

While Europol may conclude agreements with third states, it is dependent for this power on the Council, since the Council (i) needs to determine with which states and bodies Europol may conclude agreements, (ii) needs to authorize the Director to negotiate the agreements, and lastly (iii) needs to give its approval to the Board for the conclusion of the agreement.228

This power is therefore different from the Council’s power vis- à- vis the ESAs to declare the existence of an emergency situation. After all, the Council’s consent to Europol’s agreements implies it takes on (political) responsibility for the agree-ment. However, as regards the ESAs the Council can only revoke its declaration of an emergency situation, ending the ESAs’ competence; but it does not sanction every act adopted by the ESAs in the emergency situation.

Regulation 236/ 2012 has also further empowered the ESMA,229 allowing it to prohibit or impose conditions on the entry into a short sale (of a financial instru-ment) by natural or legal persons. As discussed later (see section IV 5.8.1), it is ultimately up to the ESMA to decide when and how to act. The powers granted to ESMA are a good illustration of the continuing agencification (in a qualitative sense, cf. section I 1.3). This significant new step was, as a result, also challenged by the UK but sanctioned by the Court in Short- selling.

225 See European Commission, 18 June 2006, ‘Commission Position on EP Amendments at first reading’, SP (2008) 4439 http:// www.europarl.europa.eu/ oeil/ spdoc.do?i=15160&j=0&l=en.

226 See Common Position (EC) 10/ 2009 of the Council, OJ 2009 C 75E/ 1.227 The Commission’s powers under Articles 290 and 291 TFEU (may) result(s) in material legis-

lation. In addition, under Article 45(3)d TFEU and Article 106(3), the Commission has also been conferred specific (formal) legislative powers.

228 See Article 2 of the Act of the Council laying down rules governing Europol’s external re-lations with third States and non- European Union related bodies, OJ 1999 C 26/ 19. See Article 23 of Decision (JHA) 2009/ 371, OJ 2009 L 121/ 37. For the current list of states and bodies with which Europol can conclude agreements, see Decision (JHA) 2009/ 935 of the Council, OJ 2009 L 325/ 12. At the moment of writing a case was pending before the Court in which the Parliament challenges the latest Council decision amending this list. See Case C- 363/ 14, OJ 2014 C 329/ 10.

229 See Article 28 of Regulation (EU) 236/ 2012, OJ 2012 L 86/ 1.

http://www.europarl.europa.eu/oeil/spdoc.do?i=15160&j=0&l=en

http://www.europarl.europa.eu/oeil/spdoc.do?i=15160&j=0&l=en

Setting the Scene40

1.2.1.3.3 Operational tasksLastly, some agencies have (also) been entrusted with operational tasks which are a different matter from the decision- and non- decision- making powers presented previously.

a. Exchanging information In a way, most agencies are involved in the ex-change of information, since they are all centres of expertise and a lot of agencies are also tasked to facilitate the exchange of best practices among national admin-istrations. However, the exchange of information as an operational task is dif-ferent, since the purpose of the exchange is not to support policy implementation in general but to facilitate or even realize operational activities. Articles 6 and 7 of the Eurojust decision for instance provide that Eurojust ensures that the Member States exchange relevant information on their investigations and prosecutions. Article 5 of the Europol Decision in turn provides that Europol will ‘collect, store, process, analyse and exchange information and intelligence’. Such exchange of information allows more effective combatting of international and trans- border crime. However, under the establishing decisions, this exchange of information in a hub- and- spoke model is built around a rather passive hub which acts as a tool for the national actors.230 This model, originating in a politically sensitive field, could also be exported to other areas. National customs authorities for instance also engage in data sharing,231 and while the Commission has decided not to pursue the idea of a European Customs Agency,232 the possibility that such an agency would be established to further streamline the activities of the different national authorities cannot be excluded.

b. Providing training to national functionaries The fact that national author-ities are, as a rule, responsible for the implementation of EU law evidently creates a risk that this implementation will not be uniform across the EU (cf. section I 2.1). For this reason, powers have been granted to the Commission under Article 291 TFEU, but it is also a motive behind organizing training at EU level for national functionaries. These programmes may be ambitious and broad in scope,233 while training in more specific fields is often entrusted to EU agencies.

230 From Article 8 of the Eurojust decision, it is clear that there is no real obligation on the national authorities to share information following a request. Article 8(4) of the Europol dec-is ion, on the other hand, provides that ‘national units shall supply Europol on their own initia-tive with the information and intelligence necessary for it to carry out its tasks’. Still, De Moor and Vermeulen point out that the legal force of this obligation is unclear since the Europol dec-ision lacks direct effect. See De Moor and Vermeulen, n 60, p 1095. According to Vande Sompel the obligation was ‘created half- heartedly: non- compliance of the information obligation cannot be sanctioned and there is no coercive instrument to impose the respect of it.’ See Roger Vande Sompel, ‘Europol, the High Potential EU Law Enforcement Agency, despite the Member States: How Serious is the Problem, Knowing that Belgium is a “Good Pupil in the Europol Class”?’, in Verhage, Terpstra, Deelman, Muylaert, and Van Parys (eds), Policing in Europe, Antwerp, Maklu, 2010, p 143.

231 See Decision (EC) 70/ 2008 of the European Parliament and of the Council, OJ 2008 L 23/ 21.232 See Answer of the European Commission to Parliamentary Question (E- 9415/ 2010) by MEP

Valentin Bodu, 9 December 2010.233 See eg European Commission, COM (2011) 551 final. For an analysis, see Herman van

Harten, ‘Who’s Afraid of a True European Judicial Culture’, (2012) 5 REALaw 2, pp 131– 52.


Providing training is not the major function of most agencies, but for CEPOL it is, since the agency has been established as the EU’s training institute for police officers. Hereby CEPOL continues the work undertaken under the Oisin and Falcone training programmes in the 1990s,234 but now within a permanent structure.

Under its assistance function, the EFCA has been granted the power to pro-vide training to the members of national fisheries inspectorates (as well as to the EU inspectors) and to devise a core training curriculum. The same power may be identified in the EASO’s mandate, since it both provides training and develops a European asylum curriculum.

Even if the effects of centralized training of (decentralized) Member States’ officials may have interesting harmonizing effects in the field, it should be noted that the competence to train national officials is rather modest. After all, the training also depends on the Member States’ willingness to have their officials trained. Training will also be given in line with the agency’s work programme, which is adopted by the agency’s (Member State- dominated, cf. section II 3.1.2) Board. As a result, the training offered may be ‘needs- based’ (which it evidently should be), but the question is whether the needs recognized by the Member States are the same as the objective needs for effective and uniform implementa-tion of the relevant EU law in the field.

c. Co- ordinating or organizing joint MS operations Because, as a rule, EU Member States are responsible for the execution of EU law, they will also under-take operational activities to this end. Using the example just cited, the operational measures necessary for enforcing the Common Customs Code are undertaken by the national customs authorities. Similarly, in the Common Agricultural Policy the Member States are responsible for checking their farmers’ eligibility for direct support, with EU legislation only setting out a framework and minimum require-ments for the Member States.235

Whereas the trans- border element in controlling farmlands is minimal and Member States’ authorities will usually be able to deal with this task them-selves, other operational activities often require a combined effort by the Member States: the fight against international crime, the control of the EU’s external bor-ders, control of migration flows to the EU, etc. In these areas it may be more efficient for the Member States to cooperate with each other. Of course, such cooperation may always be set up ad hoc, bilaterally or even multilaterally, but if it ceases to be a mere exception to the rule, it becomes interesting to organize a permanent structure. Such a permanent structure may then be created in the form of an EU agency.

The joint operations are still Member States’ and not the agency’s operations, meaning the operational capacity is primarily delivered by the Member States.

234 See Joint Action (JHA) 97/ 12 of the Council, OJ 1997 L 7/ 5; Joint Action (JHA) 98/ 245 of the Council, OJ 1998 L 99/ 8.

235 See for instance Articles 20 and 22 of Regulation (EC) 73/ 2009 of the Council, OJ 2009 L 30/ 16.

Setting the Scene42

The legal framework for organizing joint investigation teams (JITs) in combatting crime has been laid down in the Framework Decision 2002/ 465, which provides in Article 1(12) that bodies such as Europol and Eurojust may participate in these investigation teams.236 However, the Framework Decision makes clear that it is up to the Member States to make use of this facility and that participation by Eurojust or Europol is subject to their approval. As such it cannot be said that these two agencies even have a coordinating role. Whereas the Eurojust deci-sion is quite succinct, the Europol decision contains more provisions on the JITs. Europol for instance has the power ‘to suggest the setting up of joint investigation teams in specific cases’, but even if such a suggestion is picked up by the Member States, Europol’s tasks are limited to a (mere) supportive role once the JIT is set up.237

This is different for the EASO, which may coordinate actions in support of Member States subject to particular pressure, albeit at the request of the Member State(s) concerned. This support may involve the EASO sending an asylum sup-port team.238 The deployment of such a team is then governed by an operational plan which is concluded between the Director of EASO and the Member State(s) concerned.239 Since the support teams are made up of national experts this may still be qualified as a joint Member State operation, the coordination of which is left to the EASO.

Even if the conservation of marine biological resources under the Common Fisheries Policy (CFP) is an exclusive EU competence,240 and the Commission has its own inspection capacity, the Member States still hold the primary res-ponsibility for monitoring fishing activities on their compliance with the CFP rules.241 In this respect, Article 95 of the Regulation on the Common Control System in the CFP provides that the Commission, under the comitology examin-ation procedure, may decide which fisheries should be subject to specific control and inspection programmes. Article 9 of the EFCA Regulation provides that it is up to the EFCA to coordinate these programmes under a joint deployment plan. These plans are adopted by the Director in consultation with the Member State(s) concerned but without the latter’s formal agreement. At the same time, the implementation of the deployment plan itself again depends on the Member States, since they have to provide the means of control and inspection to the agency. Lastly, whereas there is an obligation on the Member States to provide the means committed,242 there is no clear obligation to commit means.243 The latter (general) problem inspired the 2011 revision of the Frontex regulation, since that

236 See Framework Decision (JHA) 2002/ 465 of the Council, OJ 2002 L 162/ 1.237 See Articles 5(1)d and 6 of the Europol Decision.238 See Article 10 of the EASO Regulation.239 See Article 18 of the EASO Regulation. 240 See Article 3 TFEU.241 See especially Article 79(2) of Regulation (EC) 1224/ 2009 of the Council, OJ 2009 L 343/ 1.242 See Article 13(2) a of the EFCA Regulation.243 Article 11(1) of the EFCA Regulation only provides that the ‘Member States shall each year

before 15 October notify the Agency of the means of control and inspection that it has available for the purpose of control and inspection in the following year’.


agency also faced difficulties in its operational tasks. Frontex now has the com-petence to acquire its own equipment (boats, planes, etc) pursuant to Article 7 of the regulation.244 This is significant, since the competence to acquire its own equipment (given sufficient human and financial resources) allows the agency to become less dependent on the Member States and could eventually allow it to org anize operations independently.

d. Organizing operations The power to independently organize operations rests on a number of premises: the agency should have access to its own operational personnel and equipment and its mandate should foresee a clear framework for these operations whereby the decision to initiate an operation lies squarely with the agency, subject to evident requirements.

To this date, no agency has been granted such a competence. This will come as no surprise, since it would go against the principal rule that Member States are responsible for enforcing EU law ‘in the field’. In addition, there is one funda-mental practical reason: EU support mechanisms depend on an identified need on the part of the Member States. Giving autonomous enforcement powers to the EU would side line the notion that it is up to the Member States to identify this need, not for EU actors to determine whether the Member States are in need of support. Such autonomous operational enforcement power remaining hypothetical may also be seen in the legislative process for the 2011 revision of the Frontex regulation: the (hypothetical) options to make Frontex independent from the Member States were not even included in the Commission’s Impact Assessment.245

Still, if Member State enforcement in a given regime continuously remains un-satisfactory, it could be possible to create an autonomous operational capacity at EU level as it would trigger an exceptional intervention power. In a sense, such an operational power would be comparable to the intervention powers granted to the ESAs, which may address decisions to market players in case the national au-thority (regularly competent) fails to take appropriate action in accordance with EU legislation.

e. Carrying out and organizing inspections The difference between carrying out and organizing an inspection lies in the independent role of the agency to undertake the inspection. The agency may have such an independent power or it may depend on a request from another authority (Commission or Member States).

Inspection powers may however be even more modest, as evidenced by the EMA’s powers to request Member States’ authorities to inspect manufacturing sites of producers which have applied for an authorization at the EMA.246

Article 3 of the EMSA Regulation deals with the agency’s inspection visits to the Member States and gives the agency a large deal of autonomy in this respect.

244 This is similar to the EFCA’s power in this respect: see Article 8(2) of the EFCA Regulation.245 See European Commission, SEC (2010) 149.246 See Articles 8 and 33 of the EMA Regulation.

Setting the Scene44

The methodology of the inspection visits is laid down by the Board of EMSA,247 and the decision to visit a Member State is taken by the Director in consultation with the Commission. Following its visit, the EMSA will draw up a report for the Commission and the Member State concerned. Following a cycle of inspection visits, the EMSA will draw up a general report with horizontal findings.

The EASA may organize inspections of both Member States and undertak-ings. The decision to do so lies with the Director but otherwise the procedure is similar to the one for the EMSA (discussed previously). As regards investigations of undertakings, the Regulation provides the EASA may also assign this task to national aviation authorities. The detailed rules for these inspections were laid down by the Commission in an Implementing Regulation,248 which inter alia provides that the EASA should adopt an annual inspection programme and that the Commission may also request ad hoc inspections by the EASA. Following its inspection, the EASA will draw up a report for the Commission. On the basis of this report, the EASA and the national authority concerned have to agree on an action plan to remedy the deficiencies found. If no action plan may be agreed upon or the deficiencies are not remedied in a timely manner, the Commission may ask the Member State concerned, without prejudice to the procedure of Article 258 TFEU, for clarifications or instruct the EASA to carry out periodical inspections to verify the implementation of remedial actions.249

1.2.1.3.4 Concluding remarksIn the overview just provided, some of the most important powers granted to EU agencies have been listed, together with some illustrations. This (incomplete) over-view shows that agencies do not really function ‘independently’ but are generally involved by the legislator in (composite) procedures whereby agencies need to cooperate with the Commission and the Member States to deliver policy results.250

The overview is also an account of qualitative agencification (cf. section I 1.3). The most far- reaching powers listed have been granted to the most recently estab-lished agencies and when existing agencies’ establishing acts are reviewed, their powers are strengthened. In addition, the overview shows how an instrumental classification of agencies, suggesting one single agency has one main single power, may be deceptive. Instead, most agencies combine many different powers, the most recent agencies combining several significant competences. An agency’s general power is then also influenced by possible synergy effects between these competences.251

247 Administrative Board of the EMSA, Decision on the Policy for Visits to the Member States, 25 June 2004, http:// emsa.europa.eu/ who- are- we/ admin- board/ ab- menu- documents/ download/ 168/ 190/ 23.html.

248 Commission Implementing Regulation (EU) 628/ 2013, OJ 2013 L 179/ 46.249 See also Chamon, n 174, pp 83– 4.250 See inter alia Edoardo Chiti, ‘Les agences, l’administration indirecte et la coadministra-

tion’, in Auby and Dutheil de la Rochère (eds), Droit administratif européen, Bruxelles, Bruylant, 2007, pp 275– 7.

251 See also Chamon, n 174, p 81 and p 84.

http://emsa.europa.eu/who-are-we/admin-board/ab-menu-documents/download/168/190/23.html





1.3 Agencification

Having clarified the concept of EU agency, the process of agencification may be scru-tinized. Concisely put, agencification refers to the process whereby the EU agencies take up an increasingly important role in the EU administration, both in a quantita-tive as well as in a qualitative sense, similar to Bach and Jann’s differentiation between structural and procedural agencification.252 Quantitative agencification refers to the growing number of the EU agencies, the increase in staff and total budgets, etc, and this both in absolute and relative terms. In this regard, Dehousse noted that be-tween 1992 and 2007 one- third of the ‘executive’ positions created at EU level were located in the EU agencies.253 Similarly, a 2006 study by the European Parliament’s budgetary support unit noted that in 2007 the subsidies to the agencies would rise by 11.1 per cent while the general budget itself would only rise by 3.6 per cent.254 In a qualitative sense, agencification relates to the growing importance of the EU agencies in delivering EU policies and the increasingly important powers conferred on them.255 For instance, the establishment of the three ESAs, upgrading them from ‘simple’ committees to EU agencies (cf. section III.1), and the powers with which they have been provided has meant that they fulfil a primary role in the elaboration of the existing regulatory framework for the financial services sector, as well as giving them an important role in further developing this framework.

Figures 1.5256 and 1.6257 give an idea of the quantitative dimension of agen-cification, but the qualitative dimension is much harder to assess. Still, it may be noted that usually, the more recent agencies are granted with more intense powers

252 These authors studied agencification at the national level. In their own words: ‘[p] ar agencifi-cation structurelle, nous entendons les changements opérés dans l’organisation formelle des tâches publiques, comme la création de nouvelles agences […] par agencification procédurale, nous enten-dons un transfert d’autorité des ministères vers les agences.’ See Tobias Bach and Werner Jann, ‘Des animaux dans un zoo administratif: le changement organisationnel et l’autonomie des agences en Allemagne’, (2010) 76 RISA 3, pp 470– 1.

253 Renaud Dehousse, ‘Delegation of Powers in the European Union: The Need for a Multi- principals Model’, (2008) 31 WEP 4, pp 789– 90. See also David Levi- Faur, ‘Regulatory Networks and Regulatory Agencification: Towards a Single European Regulatory Space’, (2011) 18 JEPP 6, p 811.

254 Budgetary Support Unit of the European Parliament, Study on Agencies’ Discharge, 2006, http:// www.europarl.europa.eu/ document/ activities/ cont/ 200712/ 20071213ATT15452/ 20071213ATT15452EN.pdf.

255 This practical description of agencification is preferred over more abstract definitions of the concept as may be found in the work of Levi- Faur, according to whom agencification is ‘the process of formalizing roles and missions in organizations with spatial boundaries and formal identities, either by the devolution of functions from the core organization or the creation of new organiza-tions for performing new functions’. See Levi- Faur, n 253, p 814.

256 This table excludes the CFSP agencies. The figures for the number of staff result from own calculations based on the Court of Auditors’ reports on the annual accounts of the agencies, com-plemented by information from the agencies’ annual activity reports and the establishment plants. See OJ 2013 C 365/ 1; OJ 2012 C 388/ 1; OJ 2011 C 366/ 1; OJ 2010 C 338/ 1; OJ 2009 C 304/ 1; OJ 2008 C 311/ 1; OJ 2007 C 309/ 1; OJ 2006 C 312/ 1; OJ 2005 C 332/ 1; OJ 2004 C 324/ 1; OJ 2003 C 319/ 1; OJ 2002 L 178/ 1; OJ 2001 C 372/ 1.

257 This table excludes the CFSP agencies. The figures for the Budgets result from own calcula-tions based on the Court of Auditors’ reports on the annual accounts of the agencies. See OJ 2013 C 365/ 1; OJ 2012 C 382/ 1; OJ 2011 C 366/ 1; OJ 2010 C 338/ 1; OJ 2009 C 304/ 1; OJ 2008 C 311/ 1; OJ 2007 C 309/ 1; OJ 2006 C 312/ 1; OJ 2005 C 332/ 1; OJ 2004 C 324/ 1; OJ 2003 C 319/ 1; OJ 2002

http://www.europarl.europa.eu/document/activities/cont/200712/20071213ATT15452/20071213ATT15452EN.pdf




Setting the Scene46

under an instrumental classification (cf. section I 1.2.1.2). Whereas in the 1990s agencies only had non- decision- making powers (with the exception of CPVO and OHIM), today the EU legislator no longer seems to establish mere information agencies.

1.4 The EU administration

Since agencification is said to transform the EU administration, it is useful to dedi-cate further attention to the latter concept, even if it may seem rather self- evident

0

5

10

15

20

25

30

35

0

1990

2012

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

1000

2000

3000

4000

5000

6000

7000

8000

Staff Agencies

Figure 1.5 Number of EU agencies and number of staff

1990

2012

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Budgets (million euros) Agencies

0

5

10

15

20

25

30

35

0

200

400

600

800

1000

1200

1400

1600

1800

2000

Figure 1.6 Number of EU agencies and budget size

C 326/ 1; OJ 2001 C 372/ 1; OJ 2000 C 373/ 1; OJ 1999 C 372/ 1; OJ 1998 C 406/ 1; OJ 1997 C 393/ 1; OJ 1997 C 81/ 1; OJ 1995 C 354/ 1. The GSA budget is excluded from the budget figures, since it is affected by the huge Galileo project budget. Including it in the figure would give a distorted image of the evolution of EU agencies’ budgets.

The EU Agencies in the EU Administration 47

at first sight. Traditionally, the two main executive actors at EU level were indeed the Commission and Council. Yet authors in search of a separation of powers at EU level have noted that there are further executive actors involved in the EU administration, notably the comitology committees, the EU agencies, and the Member States themselves.258 Indeed, just as the national judge may be described as the ‘juge communautaire dans l’ordre national’,259 the national administrations may also be seen as EU administrations at national level. This is what Schmidt- Aßmann suggests when observing: ‘[l’a]dministration européenne ne peut se concevoir ni se laisser décrire selon le mode hiérarchique, puisqu’elle se compose d’un ensemble d’administrations nationales et communautaires en interrelations réciproques’,260 rejecting the idea of a centralized Brussels administration. This broad conception may be juxtaposed with the more narrow conception whereby the EU administration is only made up of formal EU actors. Depending on which concept is adopted, the limits to agencification will be different. Agencification of the EU administration sensu stricto focuses on the constitutional and institu-tional issues within the EU legal order. Agencification of the EU administration sensu lato would further focus on the effects of EU agencification on the national administrations.

Although the broad notion of EU administration is not rejected, this study will foremost focus on the first limits. In addition, some of the questions which arise under the broad notion may also be addressed under a narrow conception of the EU administration through the principle of subsidiarity. The choice for a narrow conception of EU administration without rejecting the broad conception is important since the conceptualization of the notion of EU administration is not merely a theoretical discussion. Although this will not be further elaborated, the Lisbon Treaty in its Article 298 TFEU refers to an ‘open, efficient and independent European administration’ and further provides that the Council and Parliament may establish provisions to come to such an administration. Depending on how the reference to the EU administration in Article 298 TFEU is conceived, the EU legislator could then establish provisions relating to national administrations as well as to the EU actors.

To conclude, for the purposes of this study, ‘EU administration’ means the col-lective body of executive bodies at EU level.

2 Situating the EU Agencies in the EU Administration

Before turning to the limits to EU agencification in a subsequent chapter, it is useful to put the phenomenon of agencification in its context. After all, the

258 See Driessen, n 38, p 40 et seq; Renaud Dehousse, ‘Comparing National and EC Law: The Problem of the Level of Analysis’, (1994) 42 American Journal of Comparative Law 4, pp 775– 6.

259 Robert Lecourt, L’Europe des juges, Bruxelles, Bruylant, 1976, p 306.260 Eberhard Schmidt- Assmann, ‘Principes de base d’une réforme du droit administratif ’,

(2008) 24 RFDA 4, p 667.

Setting the Scene48

legislator’s choice to (increasingly) rely on the agency instrument to deliver EU policies is a political one, and understanding this political choice will help in understanding the political limits to this process.

It was noted that the traditional explanations for agencification at the national level do not seem to hold for agencification at the EU level,261 because the prem-ises of the NPM paradigm are difficult to apply to the EU. The reason why the Commission is not a typical bureaucracy and its ‘direct delivery functions’ are so limited, as Hooghe and Kassim observe,262 should be found in the model of executive federalism or Vollzugsföderalismus which was settled for following the early years of integration within the EEC framework, and which has been codi-fied by the Lisbon Treaty in Article 291(1) TFEU.

2.1 Executive federalism in the EU

Although the notion of ‘executive federalism’ has different meanings,263 in the EU context it has been described as follows: ‘While the legislative power in many fields has been transferred to the Union level, the enforcement of EU law rests with the member states.’264 According to Lenaerts, this principle of ‘executive fed-eralism’ found its way to the Community legal order from German constitutional law,265 even if both are not identical.266 It differs from other federal polities, such as the US or Belgium, where the federal level disposes of its own administration.

261 See also Giacinto Della Cananea, ‘Independent Regulatory Agencies in the European Legal Space’, in Della Cananea (ed), European Regulatory Agencies, Paris, Éditions Rive Droite, 2005, pp 157– 8.

262 Liesbet Hooghe and Hussein Kassim, ‘The Commission’s Services’, in Peterson and Shackleton (eds), The Institutions of the European Union, Oxford, OUP, 2012, p 186.

263 The present notion should not be confused with the ‘“executive federalism” in which the re-lations between governments are conducted primarily through the negotiations of political execu-tives’. See Richard Simeon, ‘Criteria for Choice in Federal Systems’, (1982– 1983) 8 Queen’s Law Journal 1 & 2, p 152. Related to this but with a negative connotation, Habermas sees an executive federalism in the EU because the heads of national executives decide on policy at federal (EU) level which is then imposed on the national parliaments. See Jürgen Habermas, The Crisis of the European Union: A Response, Cambridge, Polity, 2012, p viii. Similarly, although Habermas notes Oeter uses the notion ‘in a different sense’, Oeter sees an executive federalism in the EU whereby sensitive political issues are shifted to the EU level outside the reach of national parliaments. See Stefan Oeter, ‘Föderalismus und Demokratie’, in von Bogdandy and Bast (eds), Europäisches Verfassungsrecht: Theoretische und dogmatische Grundzüge, Heidelberg, Springer, 2010, p 104. Franzius sees in this executive federalism the tendency to decide by consensus, which impedes a genuine European public debate on EU pol-itics. See Claudio Franzius, ‘Demokratisierung der Europäischen Union’, (2013) 48 Europarecht 6, pp 666– 7. It should be noted that when these German authors refer to executive federalism, the ori-ginal term which they use is Exekutivföderalismus and not Vollzugsföderalismus.

264 Niels Petersen, ‘Democracy Concept of the European Union: Coherent Constitutional Principle or Prosaic Declaration of Intent’, (2005) 6 German Law Journal 11, p 1520.

265 Koen Lenaerts, ‘Constitutionalism and the Many Faces of Federalism’, (1990) 38 American Journal of Comparative Law 2, p 230.

266 At a fundamental level it should be noted that Article 83 of the German Basic Law provides that ‘[t] he Länder shall execute federal laws in their own right’ (emphasis added). The confirmation of the Member States’ competence in Article 291(1) TFEU on the other hand is less clear. Dubos for instance notes that ‘[l]’exécution du droit de l’Union est […] d’abord une compétence que l’Union confère à l’État.’ See Dubos, n 37, p 295. For a more nuanced view, see Claude Blumann, ‘Le sys-tème normatif de l’Union européenne vingt ans après le traité de Maastricht’, (2012) RAE 2, p 250.


Both options come with their advantages and disadvantages. In the US, for in-stance, federal Congress can rely on a federal administration for a uniform ap-plication of its laws, but Congress could also find it useful from time to time to involve the states in helping to achieve federal policy. Whereas this would be not just merely logical but rather prescribed under a model of executive federalism, the US Supreme Court has developed an anti- commandeering doctrine, whereby commandeering refers to ‘a federal requirement that state officials enact, admin-ister, or enforce a federal regulatory program’.267 This anti- commandeering doc-trine flows from the states’ sovereignty and the limits on federal power secured by the tenth amendment. Interestingly, the ‘executive federalism’ in the EU is in-spired by similar considerations, since the rule of indirect administration is based on the principle of subsidiarity.268

2.1.1 EU agencies between direct and indirect administrationThe distinction in EU administration between the rule of indirect and the excep-tion of direct administration is a traditional one. Both have, following the entry into force of the Treaty of Lisbon, been codified more clearly in primary law in Article 291 TFEU. Hofmann, Rowe, and Türk speak of direct administration if ‘implementation by the institutions and other authorities of the Union takes place […] when there is a clear legal basis conferring administrative functions on them as such’.269 Under this definition the position of the EU agencies is not clear, and it is not clarified by Hofmann et al either.270 For instance, the examples which they give for direct administration only relate to the Commission and the ECB, following legal bases in primary and secondary law. In addition they note that under direct administration the Commission and Council ‘are increasingly supported by EU agencies’,271 raising the question whether acts actually adopted by the agencies also come under direct administration. Secondly, the definition of direct administration in Hofmann et al refers to ‘a clear legal basis’ which for

267 Neil Siegel, ‘Commandeering and Its Alternatives: A Federalism Perspective’, (2006) 59 Vanderbilt Law Review 5, p 1630. Today, the anti- commandeering doctrine is shaped by three rul-ings of the US Supreme Court, namely New York v. United States, Printz v. United States, and Reno v. Condon: see Scott Thompson and Christopher Klimmek, ‘Tenth Amendment Challenges After Bond v. United States’, (2012) 46 University of San Francisco Law Review 4, pp 1014– 15.

268 Jean- Claude Piris, The Lisbon Treaty: A Legal and Political Analysis, Cambridge, CUP, 2010, p 97. In Reno v. Condon the petitioners also remarked on this in their brief, referring to the EU Data Protection Directive: ‘the approach taken by the EU Directive— an instruction to member states to enact statutes or regulations in conformity with the Directive […] is precisely the approach that Congress cannot adopt under this country’s system of dual sovereignty, as explicated in New York v. United States.’ See Brief for the Petitioners, Reno v. Condon, 528 U.S. 141 (2000), No. 98- 1464, pp 43– 4 at footnote 23. It is not completely clear whether the petitioners appreciated that this specific choice does not flow from this specific directive but is rather inspired by the con-stitutional setup of the EU itself.

269 Hofmann, Rowe, and Türk, n 50, p 11.270 Dubos notes the dual nature of EU agencies but still concludes that they are a form of direct

administration since agencies are EU bodies. See Dubos, n 37, p 301.271 Hofmann, Rowe, and Türk, n 50, p 11.

Setting the Scene50

the EU agencies only exists in secondary legislation, without it being clear that that secondary legislation conforms to the Treaties. In addition, although the EU agencies undoubtedly are EU actors, they in any case blur the distinction between direct and indirect administration since their main governing body is (as a rule) dominated by representatives of the Member States.272 For these reasons, Chiti has qualified EU agencies as part of decentralized integration: a grey area between the two ideal types.273

2.2 EU agencies under a functional and procedural perspective

Looking at EU agencification from a functional or procedural perspective then makes more sense. Hoffmann et al’s historical reconstruction of the development of the EU administration in a functional understanding is indeed revealing, since it shows how the EU administration is moving towards an integrated administra-tion. According to the authors, ‘[t] his development began with the need to address the requirements of horizontal cooperation between administrations arising from the creation of the single market’.274 The SEA and the realization of the internal market indeed pose challenges to the national administrations, in the sense that increasingly complex and numerous EU policies need to be implemented by (fol-lowing the successive enlargements) increasingly numerous and (culturally) diver-gent national administrations, all the while maintaining a sufficiently uniform level of implementation across the EU to ensure the effet utile of EU law.

The European networks of national administrative actors and the EU agencies which may be established to manage such networks can then be seen as instru-ments for meeting such a challenge in the broader evolution of the European Administrative Space (EAS).275 The rise of EU agencies may then be situated in a trend towards a more unified EAS, and hence agencification may be seen as a form of administrative integration (see also section III 2.2).

Under a procedural understanding of EU administrative law, Hofmann et al stress that EU law also lays down the procedures through which administrative rule or decision- making is conducted. Although many different procedures exist, these procedures often result in ‘shared administration’,276 which may be defined as EU administration whereby the ‘Commission and the Member States have

272 See also Merijn Chamon, ‘Verzelfstandiging in de nationale en Europese rechtsor-des: nieuwe uitdaging van een meergelaagde administratie’, (2013) TBP 2– 3, p 128.

273 Edoardo Chiti, ‘Decentralisation and Integration into the Community Administrations: A New Perspective on European Agencies’, (2004) 10 ELJ 4, p 410. See also Chapter III, n 56.

274 Hofmann, Rowe, and Türk, n 50, p 8.275 The EAS is a rather elusive concept for which many different definitions have been ad-

vanced: see Johan Olsen, ‘Towards a European Administrative Space?’, (2003) 10 JEPP 4, p 508; Stefan Kadelbach, ‘European Administrative Law and the Law of a Europeanized Administration’, in Joerges and Dehousse (eds), Good Governance in Europe’s Integrated Market, Oxford, OUP, 2001, p 178; Herwig Hofmann, ‘Mapping the European Administrative Space’, (2008) 31 WEP 4, p 671.

276 Hofmann, Rowe, and Türk, n 50, pp 14– 16.


distinct administrative tasks which are inter- dependent and set down in legis-lation and where both the Commission and the national administrations need to discharge their respective tasks for the Community policy to be implemented successfully’.277 Today, the reference to the Commission may be read as referring to the Commission and other bodies at EU level.278 Seen from this perspective, agencification allows for administrative integration and more elaborate forms of shared administration, involving both national administrations, EU agencies, and the Commission,279 and it allows for rationalization and centralization by estab-lishing procedures involving these actors.

2.3 Conclusion

The three perspectives on EU administrative law set out by Hofmann et al help in understanding the (potential) role of EU agencies in the EU administration. Although the position of the EU agencies is not that clear from an organiza-tional perspective, a functional understanding of EU administrative law shows how EU agencies are instruments in achieving a (more) integrated EU admin-istration. Similarly, from a procedural perspective, EU agencies allow for more elaborate procedures in ‘shared administration’. Taking these insights together, the process of agencification seems an atypical form of administrative integration, since the typical form would be further reliance on direct administration (by the Commission or Council).

277 Committee of Independent Experts, Second Report on Reform of the Commission: Analysis of Current Practice and Proposals for Tackling Mismanagement, Irregularities and Fraud, 10 September 1999, Vol. I, p 78.

278 See also Chamon, n 272, p 121.279 To illustrate, Article 36(2) of the EFSA Regulation allows the EFSA to outsource part of its

work to (networks of) national authorities. The resulting scientific opinions will often be presented to the Commission, which uses them to submit drafts to comitology committees composed of na-tional representatives.

IIIn Search of an Agency Model

The Provisions in Agencies’ Establishing Acts

The present chapter will verify to what extent an EU agency model may be iden-tified. The purpose and sense of such a model should be seen in the light of the debate on the EU’s (lack of ) legitimacy. Simply put, agencification without re-course to a model undermines the EU’s legitimacy, while this legitimacy would be bolstered if agencification is subject to a clear framework. Because of the ad hoc way in which agencies have been established in the past, the agencies’ statutes show a great deal of variation on a number of fundamental issues. The most important elements in agencies’ statutes will therefore be discussed and linked back to the institutions’ attempt to come to a horizontal framework for agencies, that is, the introduction of a model, which culminated in the 2012 Common Approach on Decentralised Agencies.

1 Introduction

The general works on agencification following the second wave of agency cre-ation included a presentation of all the existing agencies, their tasks, and their structures.1 Because of the huge increase in agencies in the past fifteen years, including such a presentation here would not be practical, but a study of the existing agencies’ functioning does reveal that no single EU agency model may be identified.2 Still, despite the lack of a model, the EU agencies, as type 2 public sector organizations (cf. section I 1.2), are less heterogeneous than similar organi-zations at national level.

1 See Dorothee Fischer- Appelt, Agenturen der Europäischen Gemeinschaft, Berlin, Duncker & Humblot, 1999, pp 46– 77; Michael Berger, Vertraglich nicht vorgesehene Einrichtungen des Gemeinschaftsrechts mit eigener Rechtspersönlichkeit, Baden- Baden, Nomos, 1999, Vol. 219, pp 33– 57; Vark Helfritz, Verselbständigte Verwaltungseinheiten der Europäischen Union, Berlin, Weissensee Verlag, 2000, pp 11– 52.

2 Deirdre Curtin and Renaud Dehousse, ‘European Union agencies: tipping the bal-ance?’, in Busuioc, Groenleer, and Trondal (eds), The Agency Phenomenon in the European Union, Manchester, Manchester University Press, 2012, p 193. However, Chiti juxtaposes two models: that of the old agencies and that of the agencies such as ACER and the ESAs; see Edoardo Chiti, ‘Existe- t- il un modèle d’Agence de l’Union européenne’, in Molinier (ed), Les agences de l’Union européenne, Bruxelles, Bruylant, 2011, pp 49– 74.

The EU’s Sources of Legitimacy 53

In addition, the agencification process in the EU, unlike in the national states, only seriously started relatively recently. As such, a learning process in the regulations establishing EU agencies may be observed. In general the pro-visions spelling out the functioning and the tasks of the agency and its con-stituent bodies are increasingly more detailed, developments being incremen-tally codified in subsequent legislation. In addition, the growing importance of the Parliament— especially because of the change in the legal basis to establish agencies away from Article 308 EC (Article 352 TFEU)— shows itself in the establishing regulations, by providing an increasingly important role for that institution.

According to Härtel, the lack of a horizontal framework has resulted in the cur-rent heterogeneity resulting from historical and sectorial coincidences.3 As noted, the heterogeneity may also be explained by reference to the learning process in which the EU finds itself. Still, although these explanations may be accepted, they cannot serve as a justification of the heterogeneity.

2 The EU’s Sources of Legitimacy

The problem of the EU’s legitimacy largely falls outside the scope of the present study and will therefore not be elaborated upon here, save for explaining why ra-tionalizing EU agencification is necessary.

The EU’s crisis of legitimacy was identified in the 1990s, following the Single European Act (SEA).4 The increasing competences of the EU eroded both the ‘permissive consensus’ on the EU5 and the idea that European integration was a technocratic affair. As a result, since the 1990s, the EU’s need for proper legit-imacy has been explicitly raised.

2.1 The EU’s democratic deficit

In the ensuing debate the problem of legitimacy has often been narrowed down to a problem of democracy, caused by an alleged democratic deficit at the EU level.6 When, in 1995, Joseph Weiler wrote that ‘the Community and Union of

3 Ines Härtel, ‘Die Europäische Grundrechteagentur: unnötige Burökratie oder gesteigerter Grundrechtsschutz?’, (2008) 42 Europarecht 4, p 492.

4 Gráinne de Búrca, ‘The Quest for Legitimacy in the European Union’, (1996) 59 The Modern Law Review 3, p 350.

5 See Leon Lindberg and Stuart Scheingold, Europe’s Would- be Polity: Patterns of Change in the European Community, Englewood Cliffs, Prentice Hall, 1970, p 41.

6 Nicole Bolleyer and Christine Reh, ‘EU Legitimacy Revisited: The Normative Foundations of a Multilevel Polity’, (2012) 19 JEPP 4, p 473. Arguing against the existence of such a deficit, see Giandomenico Majone, ‘Europe’s “Democratic Deficit”: The Question of Standards’, (1998) 4 ELJ 1; Andrew Moravcsik, ‘In Defence of the ‘Democratic Deficit’: Reassessing Legitimacy in the European Union’, (2002) 40 JCMS 4.

In Search of an Agency Model54

today […] suffer from very serious democratic deficiencies’7 one could indeed have agreed, given the position of the European Parliament at that time. However, the successive Treaty revisions since then have changed this picture. Now, most EU policies depend on the approval of a directly elected body representing the Union citizens. The Treaty of Lisbon has also strengthened the positions of national parliaments and of individual citizens. Although progress should still be made as regards the Parliament’s powers and the way it is elected,8 it is a different matter to say that the EU today suffers from very serious democratic deficiencies. Instead, the greatest democratic challenges today actually seem to flow from deficiencies at the national level,9 and for example from the choice to deal with crises such as the euro crisis outside the framework of the EU. In addition there is the questionable tendency of an increased reliance on the trilogues and the risk of an excessive reli-ance on delegated acts in the absence of a more detailed framework governing the adoption of these acts. But these latter two issues have to do with procedure and do not flow directly from the EU’s constitutional foundations. Lastly, at a funda-mental level one could argue against the existence of a democratic deficit in the EU, since one of the theoretical foundations, that is, the non- existence of a demos, underpinning such a deficit may be questioned.

A demos is allegedly necessary because a group must be able to transfer au-thority as a group and commit to a social contract, and because a shared political identity is necessary for a minority to accept the outcome of majoritarian decision- making.10 This lack of demos further extends beyond the question of democratic legitimacy. On the need for a European constitution, Weiler remarked (in the 1990s) that the EU is foremost in need of a constitutional debate since it would be imprudent to ‘accept a constitution without a constitutional debate. […] What Europe needs, thus, is not a constitution but an ethos and telos to justify, if they can, the constitutionalism it has already embraced’.11

Weiler does not address the question of who should then participate in such a debate, but it is convincing to argue that this would involve the members of the

7 Joseph Weiler, ‘Does Europe Need a Constitution? Demos, Telos and the German Maastricht Decision’, (1995) 1 ELJ 3, p 222.

8 The Report by MEP Duff calling for such a transnational constituency was not adopted by the Parliament. See Draft Second Report on a proposal for a modification of the Act concerning the election of the members of the European Parliament by direct universal suffrage of 20 September 1976, PE 472.030v03- 00, A7- 0027/ 2012. Instead a much watered down Report, putting forward the Spitzenkandidaten, was adopted. See European Parliament resolution of 4 July 2013 (A7- 0219/ 2013) and European Parliament resolution of 22 November 2012 (B7- 0520/ 2012).

9 While national parliaments have been strengthened by virtue of EU law, a persistent problem is the lack of interest among national parliaments in the positions taken by the national gov-ernments in the Council. See eg Koen Lenaerts and Eddy De Smijter, ‘On the Democratic Representation through the European Parliament, the Council, the Committee of the Regions, the Economic and Social Committee and the National Parliaments’, in Winter, Curtin, Kellerman, and de Witte (eds), Reforming the European Union— The Legal Debate, The Hague, Kluwer Law International, 1996, p 185. This has indeed facilitated the type of executive federalism criticized by Habermas: see Chapter I, n 263.

10 Mette Jolly, ‘A Demos for the European Union?’, (2005) 25 Politics 1, p 12.11 Weiler, n 7, p 220.


political system to which that constitution applies and that for this debate to lead to a common understanding these members would presumably need a shared pol-itical identity, that is, form a demos.

2.1.1 The impossibility of a European demosThe fundamental problem for the EU then is that it may only create a legal demos, by creating institutions, an EU citizenship and the concomitant rights, etc. A social demos on the other hand is not easily created, since it consists of an ethno- cultural dimension and a subjective emotional dimension.12 The No- Demos thesis alludes to this lack of a social demos. According to Weiler, the No- Demos thesis finds its origin in the German conception of membership of its polity and was also ap-parent in the Maastricht Urteil of the German Bundesverfassungsgericht.13 In that judgment, however, the emphasis seems to be more on the legal demos, since the Court noted:

The EU sees itself as a Union of the peoples of Europe, a Union based on dynamic devel-opment by democratic states. If it takes on sovereign tasks and exercises sovereign com-petences, it does so by virtue of the democratic legitimation given by the peoples of the Member States through their national parliaments.

From this point of view, the growth in tasks and competences of the Community also necessitates a representation of the peoples of the Member States in a European Parliament, in addition to the participation of these peoples and the democratic legitim-ation given through the national parliaments and which will democratically underpin the EU’s policy.…Already in its current phase of development, the legitimation through the European

Parliament has a supportive function which would be reinforced when a uniform electoral system is established and when the Parliament’s influence on the policy and law- making of the European Communities grows. It is of fundamental importance that the demo-cratic foundations of the Union are elaborated in the same pace as the integration between the Member States itself, so that a lively democracy in the Member States is secured.14

However, two other elements support Weiler’s identification in the Maastricht Urteil of a social demos. The first is the Court’s subsequent reference to the neces-sity of reserving to the Member States ‘important policies, which are subject to a political decision- making process in which the people may articulate a political will formation, through which the people gives a legal expression of its relatively homogeneous spiritual, social and political bond’.15 Secondly, Weiler is right to

12 Jolly, n 10, pp 14– 15.13 Weiler, n 7, p 225.14 Bundesverfassungsgericht, Maastricht Urteil, [1993] 89 BVerfGE 155, paras 96– 100.

Translated from German by the author.15 Ibid, para 101. On this relatively homogeneous bond, the Court refers to Heller, according

to whom the ‘Volk als Vielheit soll sich selbst bewußt zum Volk als Einheit bilden. Ein bestimmtes Maß sozialer Homogenität muß gegeben sein, damit politische Einheitsbildung überhaupt möglich sein soll. Solange an die Existenz solcher Homogenität geglaubt und angenommen wird, es gäbe eine Möglichkeit, durch Diskussion mit dem Gegner zur politischen Einigung zu gelangen, solange kann


note that the juge- rapporteur in the Maastricht Urteil, Paul Kirchhof, puts the emphasis on the social demos in his academic writings. According to Kirchhof, the EU may only evolve into a federation when

the existing Member States have integrated irrevocably in a European Community in which the latter has been given the original law making and - shaping prerogative. These preconditions were fulfilled in 1787 for the USA, in 1848 for Switzerland and in 1866 for the North- German Confederation. In those cases the foundations for statebuilding were present: a people which belongs together and which has expressed its desire to belong to a common state, an economic and cultural homogeneity and, with the exception of Switzerland, a common language.16

It is difficult to imagine the EU could ever meet these preconditions in our day and age, which has led Weiler to make a distinction in the No- Demos thesis between the soft ‘Not Yet’ variant, which may be found in the Court’s Maastricht Urteil, and the hard ‘Never’ variant.17 Obviously the repercussions of the latter are that the EU could never evolve into a new ‘state- like’ entity and could never attain the same ‘democratic qualities’ as ‘nation- states’. As regards the latter, however, the self- evident link between nations and statehood has been criticized by Mancini, who observed that ‘the inability to conceive of statehood in any terms other than nation statehood, or, in a nutshell, to divorce the state from the nation’.18 Mancini is right in his criticism, but so are Weiler and Nergelius when they in turn note that Mancini erroneously assumes that democracy necessitates a state.19 Instead, the notion of state should be dissociated from both the notions of nation and

auf die Unterdrückung durch physische Gewalt verzichtet, solange kann mit dem Gegner parliert werden.’ See Hermann Heller, Gesammelte Schriften, Leiden, Sijthoff, 1971, Vol. 2, p 427. Harada explains that Heller’s notion of social homogeneity was a reaction to the notion of homogeneity which depended on the existence of a nation put forward by Schmitt. See Takeo Harada, Jenseits des Unentscheidbaren: sechs Kapitel zur Frage nach dem Wesen der Zeit in der modernen und postmod-ernen Philosophie Japans, Berlin, Springer, 2002, pp 3– 5. Heller’s notion of social homogeneity then ‘ist immer ein sozial- psychologische Zustand in welchem die stets vorhandenen Gegensätzlichkeiten und Interessenkämpfe gebunden erscheinen durch ein Wirbewußtsein und – gefühl, durch einen sich aktualisierenden Gemeinschaftswillen […] Wodurch dieses Wirbewußtsein erzeugt und zerstört wird, läßt sich nicht allgemeingültig sagen […] In der europäischen Neuzeit […] waren gemeinsame Sprache, gemeinsame Kultur und politische Geschichte die wichtigsten Faktoren der sozialpsychologischen Angleichung gewesen.’ See Heller, this note, p 428 (emphasis added). Interestingly, Heller explicitly refused to lay down a general rule explaining the origin of this social homogeneity. He only remarked that in his day and age (Heller wrote during the Weimar period) the most important factors in such homogeneity were a common culture, language, and political history. Contemporary authors such as Kirchhof on the other hand take these factors as universal and permanent, cf. n 16.

16 Paul Kirchhof, ‘Europäische Einigung und der Verfassungsstaat der Bundesrepublik Deutschland’, in Kirchhof, Isensee, Schäfer, and Tietmeyer (eds), Europa als politische Idee und als rechtliche Form, Berlin, Duncker und Humblot, 1993, p 88. Translated from German by the author.

17 Weiler, n 7, pp 229– 30.18 Federico Mancini, ‘Europe: The Case for Statehood’, (1998) 4 ELJ 1, p 32 (emphasis in

original).19 Joseph Weiler, ‘Europe: The Case Against the Case for Statehood’, (1998) 4 ELJ 1, pp 57– 8;

Joakim Nergelius, The Constitutional Dilemma of the European Union, Groningen, Europa Law Publishing, 2009, p 72. In addition, one can only whole- heartedly agree with Weiler’s rejection of democracy as an end in itself, as proposed by Mancini, rather than the means to an end. See Weiler, this note, pp 61– 2.


democracy. This idea may also be found with Schimmelfennig, who has argued that the EU is democratizing in absence of a demos and differently from the way (national) states have democratized.20

Arguing against the German organic- national conception of demoi, which lies at the heart of the ‘Never’ variant of the No- Demos thesis, Weiler proposes a European demos based on civic and political culture, the exponent of which is the Union citizenship created by the Maastricht Treaty. For Weiler, the national demos and the European demos would then result in multiple demoi,21 legitimizing both national and European polities.

2.1.2 Whether a European demos should be pursuedApart from his own critique on the No- Demos thesis, Weiler further notes that the notions of peoplehood and national identity have a far larger degree of artifi-ciality and social engineering than the organic- national view would have one be-lieve.22 It is actually this observation that undermines both the No- Demos thesis and any argument for an alternative demos. More generally, Anderson explained how nations are ‘imagined communities’ since ‘the members of even the smallest nation will never know most of their fellow- members, meet them, or even hear of them, yet in the minds of each lives the image of their communion’.23

It should be noted this is not a judgemental observation by Anderson, as it does not reject nationalism; yet it shows that ‘nations’ are not original but rather cre-ated. Gellner on the other hand has criticized theories of nationalism for the self- evidentness in which they shroud themselves, to the extent that the proponents of these theories fail to see they are holding a construct at all.24 The assumed natur-alness of the national demos in the writings of academics such as Kirchhof should indeed be stressed. As was pointed out previously, this has its repercussions not only for theoretic debates on democracy, but also for constitutionalism. In his consti-tutional theory, Carl Schmitt observes that the constitution determines the form and type of the political unity which already exists,25 rather than the political unity being the result of a constitutional debate and constitution. Schmitt’s critical dis-tinction between constitution and constitutional law26 stands in stark contrast with an apparent naiveté regarding how this political unity arises when he observes:

The unity of the German Reich does not rest on these 181 articles [of the Weimar Constitution] and their validity, but rather on the political existence of the German people. The will of the German people, therefore something existential, establishes the

20 Frank Schimmelfennig, ‘The Normative Origins of Democracy in the European Union: Toward a Transformationalist Theory of Democratization’, (2010) 2 European Political Science Review 2, pp 211– 33.

21 Weiler, n 7, pp 252– 5. 22 Ibid, p 239.23 Benedict Anderson, Imagined Communities: Reflections on the Origin and Spread of

Nationalism, London, Verso, 1985, p 15.24 Ernest Gellner, Nationalism, London, Phoenix, 1997, p 7.25 Carl Schmitt, Constitutional Theory, Durham, Duke University Press, 2008, p 75.26 Ibid, p 59.


unity in political and public law terms beyond all systematic contradictions, disconnect-edness, and lack of clarity of the individual constitutional laws.27

Thus not only the existence of ‘the German people’ is assumed, but it is also assumed that it is its (collective) will which is the foundation and legitimation of the German state. Such justifications disregard the fact that modern states have been created by elites rather than being natural entities resulting from ‘the will of the people’.28 Furthermore, since the latter does not exist as an objective reality in the same way as the individual does, claiming it can express a single and collective will would appear nonsensical. Although states actually exist prior to nations, traditional nationalist theorists have reversed this relationship by pro-posing that a people may awaken and be conscious of its nationhood.29 Again linking the notions of state and nation, according to Leibholz, the sovereignty of the state depends on a sovereign nation finding its organizational form and expression in the state. For this to materialize, the nation must ‘declare itself to be sovereign when once the people has broadened into a nation— that is to say, when the people has become conscious of its own political and cultural heritage, and acknowledges the feelings of its own existence as an independent concrete entity’.30 In reality, this awakening is not a natural process, but a process of socialization and internalization in which the institutions of mass communi-cation, since the nineteenth century, have played an important role.31 As Laski noted, the legal union of the State was spiritualized into moral pre- eminence,32 by relying on the notion of a general will of a community. But neither this gen-eral will nor such a community exist in a way which may be compared to the unity of will in a human being.33

2.1.3 Overcoming a legitimacy deficit in the absence of a European demosWhat purpose do these metaphysical contemplations serve in the present study? Coming back to the initial starting point of the issue of legitimation, they would argue against the need of a European demos for the EU to be democratic and to legitimize European integration. Since demoi have resulted from states, rather

27 Ibid, p 65.28 In his plea for a European Statehood, Mancini noted: ‘The unification of Italy between 1859

and 1861 was the result of the work of thin political and intellectual élites aided and abetted by two powerful nation states, France and Britain. The claim, sometimes to be found in our primary- school textbooks, that the Risorgimento sprang from popular demand and involved popular participation is a pious untruth.’ See Mancini, n 18, p 36.

29 Gellner, n 24, p 8.30 Gerhard Leibholz, ‘Sovereignty and European Integration: Some Basic Considerations’,

in Brugmans, Cerych, and Lory (eds), Sciences humaines et intégration européenne, Leyde, Sythoff, 1961, p 158.

31 Eric Hobsbawm, Nations and Nationalism since 1780: Programme, Myth, Reality, Cambridge, CUP, 1994, pp 91– 2.

32 Harold Laski, The Foundations of Sovereignty and Other Essays, Union, Lawbook exchange, 2004, p 235.

33 Harold Laski, Studies in Law and Politics, London, Allen and Unwin, 1932, p 256.


than the other way around, there is no reason why the EU’s legitimacy would depend on a demos,34 that is, there is no reason why the EU should necessarily engage in the same fabricated legitimization exercise as national states have; and neither is there a reason why the EU should evolve into a state- like entity.

This is not the same as saying the EU does not need legitimation and popular support; rather, popular support and the EU’s legitimacy could flow from its conception as a problem- solving exercise agreed upon by the Member States.35 Unlike national demoi, national states are objective realities, and so is the eco-nomic interdependence characterizing them.36 Regional integration should then be assessed on its merits in relation to its conception as a problem- solving exercise. This argument is akin to those developed by Ipsen and Majone, who see the EU as a Zweckverband funktioneller Integration and a regulatory state respectively.37 Both stress the problem- solving capacity as the raison d’ être and legitimacy of the EU. But as Joerges points out, both authors also propose EU institutions should be quarantined from political influences and the pol-icies determined by expert knowledge,38 which is not proposed here if only because the dichotomy between the technical and the political realm is not subscribed to.

It should further be noted that the EU as a problem- solving exercise is not to be equated with or confined to the notion of output- legitimacy developed by Scharpf as an alternative to input- legitimacy. According to Scharpf, input- legitimacy pre-supposes a ‘thick’ collective identity, that is, a demos, which the EU is far from acquiring.39 As a result, Scharpf proposes to legitimize EU integration by stressing the output of the integration.40 The difference with Scharpf lies in the fact that he considers output- legitimacy as an alternative to input- legitimacy, because the EU cannot rely on ‘a Gemeinschaftsglaube, arising from pre- existing commonalities of history, language, culture, and ethnicity’.41 Further in contrast with Scharpf,

34 In the same vein, the creation of a demos is only one possible way of legitimizing, post- hoc, the existence of the state.

35 National states themselves are of course also problem- solving exercises in their own right. All forms of regional integration may then be seen as the result of the participating states having been confronted with the limits of their problem- solving capacity. Everson has suggested that also EU agencification in one respect could be seen in this sense as ‘an affirmation that the appro-priate character of the EU is one of a problem- solving technical nature, simply giving “added value” to the Member States’ regulatory regimes’. See M. Everson, ‘Agencies: The “Dark Hour” of the Executive?’, in Hofmann and Türk (eds), Legal Challenges in EU Administrative Law: Towards an Integrated Administration, Cheltenham, Edward Elgar, 2009, p 117.

36 Of course it is not claimed here that this would make the EU an apolitical technocratic pro-ject, based on pure economic considerations.

37 Hans Peter Ipsen, ‘Europäische Verfassung— Nationale Verfassung’, (1987) 22 Europarecht 3, p 203; Giandomenico Majone, ‘The Rise of the Regulatory State in Europe’, (1994) 17 WEP 3.

38 Christian Joerges, ‘Good Governance in the European Internal Market: Two Competing Legal Conceptualisations of European Integration and their Synthesis’, (2001) EUI Working Paper 2001/ 29, p 44.

39 Fritz Scharpf, Governing in Europe: Effective and Democratic?, Oxford, OUP, 1999, pp 8– 9.40 Ibid, p 11. 41 Ibid, p 8.


the legitimacy of the EU here also depends on the nature of the process of reach-ing solutions to common problems. This process needs to be in conformity with democratic principles (which do not depend on the existence of a demos), the rule of law, good governance, etc. This further links with the procedural constitution-alism described by Poiares Maduro as a (partial) alternative for legitimizing global governance,42 which, in contrast to national states, but just like the EU, cannot rely on the existence of a political community with an underlying long- term pol-itical contract.43 Legitimacy is then derived from the quality of the deliberative processes: access thereto and transparency thereof. In this, higher standards than those applicable in the states should be adopted.44 Similarly, this fits well with the argument made by Schimmelfennig on the transformation of democracy in a re-gional organization such as the EU, where ‘all institutional actors share the core values and norms of liberal democracy and accept them as fundamental standards of political legitimacy’.45

Summing up, the EU may derive its legitimacy from its capacity to produce effi-cient solutions to common problems combined with the way in which it produces these solutions. For this no demos is needed, as there are only citizens brought together through the EU, and these citizens should be able to assess the efficiency and effectiveness of the EU in tackling problems, fully conscious of the fact that common challenges require common solutions.46

The subsequent involvement of citizens in this process may well lead to an inter- subjectively shared praxis,47 which according to Habermas is ‘a circulatory process that is generated through the legal institutionalisation of citizens’ com-munication’,48 resulting in an ethical– political self- understanding that makes democratic will- formation possible.49 Yet if the present section has been clear enough, it should be equally clear that such a specific outcome of that praxis, as foreseen by Habermas, is not viewed here as necessary for legitimating the EU, since that would again depend on whether such democratic praxis indeed creates a demos.50

42 Miguel Poiares Maduro, ‘From Constitutions to Constitutionalism: A Constitutional Approach for Global Governance’, in Lewis (ed), Global Governance and the Quest for Justice: Volume I: International and Regional Organisations, Oxford, Hart Publishing, 2006, p 244.

43 Ibid, p 239.44 Ibid, p 244. As Popelier points out, such considerations also seem to have inspired the

Commission White Paper on Governance. See Patricia Popelier, ‘Governance and Better Regulation: Dealing with the Legitimacy Paradox’, (2011) 17 EPL 3, pp 558– 9.

45 Schimmelfennig, n 20, p 221.46 Similarly, Kranz rejects the idea of a democratic deficit: see Jerzy Kranz, ‘Gibt es ein

Demokratiedefizit in der Europäischen Union?’, (2013) 51 AVR 4, p 413.47 Jürgen Habermas, ‘Citizenship and National Identity: Some Reflections on the Future

of Europe’, in Beiner (ed), Theorizing Citizenship, Albany, State University of New York Press, 1995, p 262.

48 Jürgen Habermas, ‘Reply to Grimm’, in Gowan and Anderson (eds), The Question of Europe, London, Verso, 1997, p 264.

49 Ibid, p 264.50 See also Simon Hix, ‘The Study of the European Union II: The “New Governance” Agenda

and Its Rival’, (1998) 5 JEPP 1, p 53.


2.2 Legitimacy of the EU administration

Applying these insights specifically to the EU administration, the latter’s legit-imacy will be bolstered if it succeeds in delivering policy and if administrative pro-cedure meets a sufficient standard. These two sources are also invoked by Craig, who refers to both the effectiveness and accountability of the EU administration.51

According to Craig, the formal legitimacy of the EU’s administration is un-problematic. This indeed seems clear for bodies such as the Commission but Craig extends this conclusion to the agencies, noting that ‘the relevant administrative powers and duties are enshrined in formal Community legislation’.52 It should be clear that this conclusion is not subscribed to, since the questions which are cen-tral to the present study partly result from the finding that the agencies’ formal legitimacy cannot be compared to that of the Commission. Just like the comitol-ogy system up until the entry into force of the SEA, EU agencies find themselves on constitutional terra incognita, even following the Short- selling case in which the Court sanctioned the agency instrument (cf. Chapter IV).

Seen under this perspective, the problem of an unguided agencification within the EU and the heterogeneity in the EU agencies’ statutes becomes clear. Rationalizing the creation and functioning of agencies then adds to greater legit-imacy both for the agencies, the process of agencification, and the EU itself, be-cause it allows the latter to achieve its objectives more effectively and in a more efficient and transparent manner. While such a rationalization cannot affect what Craig has called the formal legitimacy of the EU agencies, which is a question of constitutional nature, it still is inspired by what Bomhoff has called European law’s ‘perfection- seeking’ dynamics, whereby in the European legal order ‘[e] xpansive “general principles” on a high level of abstraction should, and do, inform all ac-tions by legal actors’.53 Perhaps unfortunately, the underlying drive which makes the European legal order ‘the best it can be’54 has so far not descended to and permeated the process of agencification.

The search for greater transparency, homogeneity, and rationality in the stat-utes of agencies is not new but has so far only culminated in the 2012 Common Approach on Decentralised Agencies. The quest for greater transparency and rationality was originally sparked by the second wave of agency creation in the 1990s.55 As early as the beginning of the second wave, Parliament noted that the heterogeneity characterizing the agencies could not be wholly justified.56 Later, in

51 Paul Craig, ‘Legitimacy in Administrative Law: European Union’, in Ruffert (ed), Legitimacy in European Administrative Law: Reform and Reconstruction, Groningen, Europa Law Publishing, 2011, pp 201– 8.

52 Ibid, p 200.53 Jacco Bomhoff, ‘Perfectionism in European Law’, (2012) 14 CYELS, p 99.54 Ibid, p 76.55 In relation to the French agencies, Autin also called for a rationalization, noting: ‘Les créations

se sont multipliées au fil des ans, sans idée directrice ni plan d’ensemble, ainsi qu’il a été fréquem-ment souligné.’ See Jean- Louis Autin, ‘Le devenir des autorités administratives indépendantes’, (2010) 26 RFDA 5, p 875.

56 As early as 1992, for instance, the Parliament adopted a resolution in which it noted: ‘the setting- up, legal status and operation of these bodies and agencies is far from standardized […], with


its Herman Report of 1999, it summed up a number of key issues which today remain problematic after the adoption of the Common Approach.57 Some of the issues are considered in the Common Approach whereas others, especially those related to the position of agencies in the EU’s institutional architecture, are still untouched. Following the adoption of the Herman Report by the Parliament, the Commission partially followed up on Parliament’s invitation in its White Paper on Governance, announcing a framework for the agencies.58

The Parliament kept the pressure on the Commission by adopting a resolution in 2004 in which it repeated its criticism of the proliferation of agencies.59 The interinstitutional agreement was eventually proposed in 2005,60 but the Council objected. Officially, it did not want to start negotiations because the framework would bind future legislators, an interinstitutional agreement being an inappro-priate instrument for this purpose.61 In its opinion, the Council’s legal service (CLS) recognized that an interinstitutional agreement could have binding effect, following Commission v. Council,62 but that it could not contain legislative or supra- legislative rules.63 However, Hummer, in his study on interinstitutional agreements, has argued that the proposed interinstitutional agreement could ‘be binding insofar as its content shows that the three institutions intend to enter into a commitment towards each other’.64

Principally, it seems obvious that the institutions cannot resort to interinstitutional agreements to adopt arrangements going against primary law or sec-ondary legislation. Such amendments should instead be adopted following the appropriate procedures. In his study on interinstitutional conventions, Driessen noted that what interinstitutional agreements ‘can do is to build upon primary law’.65 The same would a fortiori be the case as regards secondary legislation. According to Driessen this could even result in allowing for greater participation

the result that there is a wide spectrum of different types of body […][and stressing] the importance of standard provisions for the implementation and operation of specialist bodies and agencies, with a view to improving their efficiency, facilitating parliamentary control and setting up a transparent, rational system.’ See points 4 and 10 of the Resolution of the European Parliament on the setting up and operation of specialized bodies and agencies, OJ 1993 C 42/ 63.

57 See points 21 to 29 of the Resolution of the European Parliament, OJ 1999 C 219/ 427.58 European Commission, COM (2001) 428 final, p 24.59 Resolution of the European Parliament, OJ 2004 C 92E/ 119.60 European Commission, COM (2005) 59 final.61 Françoise Comte, ‘Agences européennes: relance d’une réflexion interinstitutionelle euro-

péenne?’, (2008) RDUE 3, p 487. See also Maroš Šefčovič, ‘Ein Ordnungsrahmen für die dezen-tralisierten Agenturen der EU’, (2012) EuZW 21, p 802.

62 Case C- 25/ 94, Commission v. Council, [1996] ECR I- 1469.63 Legal Service of the Council of the European Union, Opinion on the Draft Interinstitutional

Agreement on the operating framework for the European regulatory agencies, 6 April 2005, Doc. 7861/ 05. Similarly see the CLS’s resistance, invoking the institutional balance, against the framework agreement between the Commission and Parliament: Council of the European Union, 18 October 2010, Doc. 15018/ 10.

64 Waldemar Hummer, ‘From “Interinstitutional Agreements” to “Interinstitutional Agencies/ Offices”?’, (2007) 13 ELJ 1, p 67.

65 Bart Driessen, Interinstitutional Convention as Checks and Balances in EU Law, Leuven, KUL, 2006, PhD Thesis, p 89 (emphasis in original).


of the Parliament, although such commitments would not be enforceable before the Court.66 Of course this is different for those issues on which neither primary nor secondary law provide any rules, as was the case in Commission v. Council.

It is important to note here that the arrangement at issue in Commission v. Council, although binding, only contained a principle rather than a clear rule. The Commission’s draft interinstitutional agreement on the other hand did con-tain clear rules,67 which under a binding agreement would have restricted the legislator’s discretion. The CLS therefore found that the legislator itself should lay down these rules and explored the possibilities for such a framework. The first option identified by the CLS was Article 308 EC, but that legal basis only provided for the Parliament’s consultation, meaning that it would not be able to decide on the horizontal rules which would be binding on it when sector- specific legal bases were later used to establish agencies. The second was a combination of all possible sector- specific legal bases which may be used for establishing an agency, but this would result in incompatible procedures. The CLS concluded that an interinstitutional agreement was not suitable for laying down binding horizontal measures. While such measures could, in theory, be adopted through legislation, the Treaties did not contain an adequate legal basis.68

Even if this legal analysis was interesting, the true reasons for the Council’s ob-jections should be searched for in the political sphere, since the agreement would have limited the flexibility left to the legislator. The dual boards in the European Supervisory Authorities (ESAs) and the Agency for the Cooperation of Energy Regulators (ACER) (cf. section II 3.1) would for instance not have been possible under the rules of the interinstitutional agreement. This, much more than an interinstitutional agreement not being suitable to adopt binding rules, explains the Council’s staunch position. Otherwise, it would not make much sense to reject an interinstitutional agreement on the grounds that it is non- binding,69 only to adopt a non- binding Common Approach seven years later.

What is more, in the course of those seven years, the Treaty of Lisbon also entered into force and brought a number of changes which merit special consid-eration in the light of the pre- Lisbon opinion of the CLS. For one, the old Article 218(1) EC was finally updated to the new institutional reality which had evolved since the SEA.70 Whereas it referred only to the Council and Commission settling

66 Ibid.67 However, it also contained a number of principles, such as that on the legal basis. The

Commission proposed that an establishing act ‘must be built on the provision of the EC Treaty which forms the specific legal basis of the policy in question’.

68 See Legal Service of the Council of the European Union, n 63.69 After all, it should be recalled that the Court has ruled that the intention of the parties deter-

mines whether an interinstitutional agreement is binding: see n 62. See also Fabien Le Bot, ‘La soft law et les procédures d’adoption des actes de l’Union européenne’, (2014) RUE 576, p 139. The Council could therefore easily have concluded the agreement with an explicit reservation as regards its non- binding nature.

70 Up until the entry into force of the Lisbon Treaty, this provision had remained unchanged. See Anne- Marie Tournepiche, Les accords interinstitutionnels dans l’Union européenne, Bruxelles, Bruylant, 2011, p 311.


their ‘methods of cooperation’, Article 295 TFEU now refers to the three institu-tions who may ‘conclude interinstitutional agreements of binding nature’. The new Article thus recognizes the Parliament’s position,71 and allows for binding arrangements to be concluded. However, the possibility to adopt the proposed interinstitutional agreement pursuant to this new Article was never explored. In addition, the Lisbon Treaty also altered the procedure under Article 352 TFEU introducing the requirement of Parliament’s consent. This of course undermines the CLS’s initial rejection of Article 308 EC as a legal basis. While the Parliament would still not have a formal role in the negotiations, it would have to be involved in practice, since the power of consent in the end comes down to a veto power.72

Because of the Council’s refusal to negotiate, the Commission withdrew its draft and called for an interinstitutional discussion.73 The interinstitutional working group started its work in 200974 and produced the non- binding Common Approach on Decentralised Agencies in 2012, finalizing an eleven- year process that had started with the White Paper on Governance of 2001.75

3 The EU Agencies’ Statutes

An analysis of the establishing acts of the agencies reveals that at the time of the adoption of the Common Approach there was still much scope in rational-izing their functioning, since only in three areas, which will not be further dis-cussed, has there been a genuine horizontal approach. This is so for the agencies’ budgets, access to documents, and the processing of personal data. In the follow-ing sections, the cornerstones on which an agency is built76 will be highlighted. The agencies’ establishing acts are discussed as they stood before the adoption of the Common Approach. When changes have been effectuated following the Common Approach this is indicated.77

71 In 2006, Driessen found Article 218(1) EC was an imperfect legal basis for interinstitutional agreements concluded between institutions other than the Council or the Commission. See Driessen, n 65, p 72.

72 Alan Dashwood, ‘Community Legislative Procedures in the Era of the Treaty on European Union’, (1994) 19 ELRev 4, p 364.

73 See European Commission, COM (2008) 135 final. For a view from the inside on the reform process following the 2008 communication, see Comte, n 61. See also European Commission Withdrawal of Obsolete Proposals, OJ 2009 C 71/ 17.

74 Rapid Press Release, 18 March 2009, ‘EU starts discussions on European Agencies’, IP/ 09/ 413.

75 For a number of horizontal studies conducted on the EU agencies during that period, see European Commission, SEC (2001) 340; Commission Budget Directorate General, Meta- Evaluation on the Community Agency System, 15 September 2003; Euréval and Rambøll- Management, Meta- study on Decentralised Agencies: Cross- cutting Analysis of Evaluation Findings, September 2008; see also the four- volume Rambøll report presented in December 2009.

76 See Ronald van Ooik, ‘The Growing Importance of Agencies in the EU: Shifting Governance and the Institutional Balance’, in Curtin and Wessel (eds), Good Governance and the European Union: Reflections on Concepts, Institutions and Substance, Antwerpen, Intersentia, 2005, p 132.

77 In the footnotes reference will be made to the old (pre- Common Approach) and new (post- Common Approach) establishing acts. If no distinction is made, the new act did not alter the

The EU Agencies’ Statutes 65

3.1 The Boards

Because an EU agency is an autonomous body, it cannot rely on the Commission or another institution to set out its policy, although the regulations generally provide that the agencies’ work programmes need to be consistent with the Commission’s priorities or the EU’s policies. As a result, every EU agency has its own organ de-termining its policy. In addition, that body will normally also be responsible for taking the necessary administrative measures, allowing the agency to function.

3.1.1 Tasks, nomenclature, and independence of the BoardsWhile the Common Approach refers to the Boards as Management Boards, these bodies do not have a unified name.78 In the cases of the ACER and the ESAs the two functions, that is, the policy and administrative functions, have also been split up and divided between two different bodies, the policy function being entrusted to the Board of Supervisors (ESAs) or the Board of Regulators (ACER). These boards bring together the national regulatory au-thorities (NRAs), with the Commission only having a non- voting member. The administrative tasks then reside with the classic Management Board (ESAs) or the Administrative Board (ACER). While the Common Approach does not impose a uniform name, the Commission has promoted a uniformization in its recent legislative proposals.79

The independence of the EU agencies (and therefore that of the Boards as well) in administrative and policy matters is restricted by the agency’s dependence on the subsidies from the EU budget and the obligation to fulfil its tasks within the framework of the EU’s policies (cf. section II 3.3). It is within the contours of these restrictions that the agencies enjoy their autonomy.80 This is different for the Common Foreign and Security Policy (CFSP) agencies which operate under the authority of the HR and to which the Council, in the case of the EDA, may even address specific guidelines.81 The EDA’s Board is also composed of representatives at ministerial level.

old one. This is relevant to the ENISA, EMSA, and GSA. The ENISA and EMSA Regulations were recast in 2013; that of the GSA in 2014. See Regulation (EU) 100/ 2013, OJ 2013 L 39/ 30; Regulation (EU) 526/ 2013, OJ 2013 L 165/ 41, and Regulation (EU) 512/ 2014, OJ 2014 L 150/ 72.

78 For Management Boards see Articles 4 CdT, 33 EASA, 25 EASO, 14 ECDC, 78 ECHA, 8 EEA, 10 EIGE, 25 EFSA, 65 EMA, 9 EMCDDA, 6 old ENISA (Article 5 new ENISA), 45 ESAs, 12 eu- LISA, 37 Europol, 12 FRA, and Article 20 Frontex. For Administrative Boards, see Articles 12 ACER, 23 EFCA, 10 EMSA, 25 ERA, 5(1) GSA, and Article 126 OHIM. For Governing Boards, see Articles 4(1) a Cedefop, 10 CEPOL, 4(1) a EIT, 7 ETF, 8 EU- OSHA, and Article 5(a) Eurofound.

79 See for instance the proposal by the Commission under the fourth railway package, COM (2013) 27 final.

80 Further, Article 110 of the EU Staff Regulations provides that the implementing rules which the Commission adopts for its staff will also apply to the agencies and that only exception-ally different implementing rules may be adopted by the (Boards of the) agencies subject to the Commission’s consent.

81 See Article 4(1) EDA.


3.1.2 Composition of the BoardsThe composition of the Boards varies but a general rule is that every Member State has its own representative. Six agencies are an exception to this rule.

3.1.2.1 Boards in which not every Member State is directly representedThe exceptional European Food Safety Authority (EFSA) Board may be ex-plained by the specific context in which the agency was established, as noted by the Commission in its original proposal.82 Following the food scares in the 1990s, the Commission proposed a strong and independent European food regu-lator with a Board composed of sixteen members; the three institutions each ap-pointing four members, with another four representing consumers and industry.83 However, in the final regulation, the Council appoints the fourteen members in consultation with the Parliament.

The second exception is the European Institute for Gender Equality (EIGE). In its original proposal, the Commission again cited the requirement of inde-pendence to introduce a parity between the Commission and the Council, with the Commission appointing three more members representing NGOs.84 The EIGE proved an interesting opportunity for the Commission to promote its idea of an equal representation of the EU’s two executive branches in the Boards of agencies (cf. section II 3.1.2.2). However, the Parliament first proposed a solution analogous to the one adopted for the EFSA.85 The Commission accepted these amendments in its revised proposal,86 subject to the introduction of a new rule prescribing an equal say between Commission and Council on matters for which the Commission is also responsible (ie the budget and the work programme). In the end the Board is composed of eighteen representatives of the Member States appointed by the Council following a rule of rotation and one Commission representative.87

The third exception is probably the most important one, as it may be a char-acteristic of an agency model in economic regulation. This exception is found in the ACER and ESAs’ regulations. In the ACER, the Administrative Board is composed of nine members, five appointed by the Council and the others divided between the Commission and Parliament, despite the Commission’s original pro-posal for a Commission– Council parity.88 In the original proposal for the ESAs

82 European Commission, COM (2000) 716 final.83 See Article 24 of the original proposal cited n 82.84 See Article 10 in European Commission, COM (2005) 81 final.85 Legislative resolution of the European Parliament, OJ 2006 C 291E/ 91.86 European Commission, COM (2006) 209 final.87 Following the adoption of the regulation, the three institutions adopted an addendum ex-

plaining that the composition of the EIGE’s Board would not be a precedent for future agencies: see Addendum to Regulation (EC) 1922/ 2006 of the European Parliament and of the Council, OJ 2007 L 54/ 3.

88 In the original proposal, the Commission again pushed for a parity between itself and the Council (six members each). See Article 9(1) of European Commission, COM (2007) 530 final. The Parliament in first reading changed this to a Commission- Council- Parliament parity (two members each). See Article 13(1) of the Legislative Resolution of the European Parliament, OJ 2009 C 286E/ 149.


the Commission no longer insisted upon parity, instead proposing a Board com-posed of the Chairperson of the ESA, a Commission representative, and four members of the Board of Supervisors. However, the Board is composed of the Chairperson and members of the Board of Supervisors, the Commission represen-tative participating in the meetings of the Board but without a right to vote, just like the Executive Director.

It should be noted that had the Body of European Regulators of Electronic Communications (BEREC) been an agency, it would squarely fall under this exception as well, since in the Commission’s original proposal for a European Electronic Communications Market Authority (ECMA) there was also the idea of two Boards, characteristic of both the ACER and ESAs.89 However, since the BEREC itself is not an agency, the Management Committee of the BEREC Office simply duplicates the composition of the BEREC Board of Regulators, although the Commission representative does have the right to vote in the Management Committee, unlike in the Board of Regulators.

Lastly, the composition of the Boards of the agencies which operate in a policy field which is subject to ‘variable geometry’ will also depend on which Member States cooperate in this policy field. So far there are four such agencies in the Board of which (at least) one Member State is not represented because it does not partici-pate in the policy concerned. This is firstly the case for the EDA and Denmark. The Management Board of Frontex is composed of the EU Member States which are signatories to the Schengen acquis, although the United Kingdom and Ireland are invited to participate in the Board’s meetings.90 The case of the eu- LISA is more complicated, since it manages the Visa Information System (VIS) and the Schengen Information System (SIS II) for parties to Schengen and EURODAC in the EU’s asylum policy. The UK and Ireland participate in the latter, but only partially in VIS and not in SIS II. Denmark participates in EURODAC but only partially in VIS and SIS II. Lastly, since the Single Resolution Board (SRB) is established as part of the package on the Banking Union, only the Member States participating therein will be represented on the Board.

3.1.2.2 Commission representation in the BoardsAs a rule, the Commission always has at least one representative and may have up to six representatives with the right to vote on the Board.91 Interestingly,

The Council suggested a Board of only six members, one for the Commission and five for the Council (see Article 11(1) of the Common Position (EC) 10/ 2009, OJ 2009 C 75E/ 1), before the institutions found a compromise on second reading.

89 See Article 25(1) of European Commission, COM (2007) 699 final.90 The United Kingdom and Ireland were not allowed to take part in the adoption of the Frontex

Regulation, pursuant to their opt- out from the Schengen acquis. The UK subsequently unsuccess-fully challenged the validity of the Regulation on this account before the Court, see Case C- 77/ 05, UK v. Council, [2007] ECR I- 11459.

91 The Commission has one representative on the Board pursuant to Articles 7(1) BEREC Office, 34(1) EASA, 25(1) EFSA, 10(1)b EIGE, 37(1) CPVO, 37(1) Europol, and 127(1) OHIM; two representatives on the Board pursuant to Articles 12(1) ACER, 4(1)c CdT, 25(1) EASO, 8(1) EEA, 65(1) EMA, 9(1) EMCDDA, 6(1) new ENISA, 13(1) eu- LISA, 12(1)c FRA, and 21(1)


according to Blin, the presence of the Commission in the Board follows from the Meroni judgment (cf. section IV 5), since a Board composed solely of Member State representatives would effectively substitute the EU and its institutions, and this in violation of the division of competences and the institutional balance.92 While Blin alludes to the notion of control, the Commission does not necessarily require a representation on the Board to control an agency and its representation in itself is also insufficient to guarantee ‘control’.

While the Commission is represented on the Boards of the Community Plant Variety Office (CPVO) and Office for Harmonisation in the Internal Market (OHIM), it does not have voting rights. For evident reasons, the CFSP agencies are also an exception to the rule. In the Board of CEPOL the Commission also only has observer status, and while it is fully associated with Eurojust’s work, it is not represented in the Eurojust College.93 Yet, following the entry into force of the Lisbon Treaty, few objective reasons stand in the way of enlarging the Boards of Justice and Home Affairs (JHA) agencies with an extra full member represent-ing the Commission.

Although in some cases there is some logic to reserving only one or up to six seats for the Commission,94 in general there is not much sense in so much het-erogeneity. This was also recognized by the Commission, who proposed slimmed- down Boards with Council– Commission parity in 2002.95 The Commission pre-sented this under the guise of increased efficiency, but evidently it would also have strengthened its own position considerably and would more than make up for its loss of influence on the agencies’ Boards in consequence of successive enlarge-ments (see section V 4.2.2.2).

3.1.2.3 European Parliament representation on the BoardsIn a number of Boards, the Parliament is also represented, or it may appoint inde-pendent scientific personalities.96 These are then full members of the Board with

Frontex; three representatives on the Board pursuant to Articles 4(2)d Cedefop, 6(1)d Eurofound, 8(1)d EU- OSHA, 7(1) ETF, 14(1) ECDC, 6(1) old ENISA, and 79(1) ECHA; four representatives pursuant to Articles 11(1) (old and new) EMSA and 26(1) ERA, pursuant to Article 5(2) new GSA; five representatives pursuant to Article 5(2) old GSA; finally, pursuant to Article 24(1) EFCA, the Commission has six representatives.

92 Olivier Blin, ‘Les Structures des Agences de l’Union européenne’, in Couzinet (ed), Les Agences de l’Union européenne, Toulouse, PUSST, 2002, p 83.

93 See Articles 9(3) CEPOL and 11 Eurojust.94 Since the Boards of the EFSA and EIGE have a reduced number of members, more than one

representative for the Commission would seem disproportionate. The large number of Commission representatives in the EFCA’s Board does not seem excessive considering the Commission’s inspec-tion role in the Common Fisheries Policy.

95 European Commission, COM (2002) 718 final, p 9. Further see European Commission, COM (2005) 59 final, p 14.

96 See Articles 12(1) ACER, 14(1) ECDC, 79(1) ECHA, 8(1) EEA, 65(1) EMA, 9(1) EMCDDA, 7(1) ETF, 5(2) GSA. The members appointed may be representatives (ACER, ECDC, EMA) or scientific personalities (EEA, EMCDDA, ETF, ECHA, GSA); the members usually have voting rights, but not in the case of the GSA, EMCDDA and ETF; the Parliament will always appoint two members, but three in the case of the ETF and one in the case of the GSA.


the right to vote, except in the case of the European Training Foundation (ETF) and European GNSS Agency (GSA) where the representatives are non- voting members. Seeing as agencies are executive bodies, the Parliament’s representation on these Boards seems an anomaly, unless these members represent general or scientific interests.

The Commission’s original proposals for these agencies did not foresee a role for the Parliament in the Board, but the Parliament tabled amendments to this effect.97 Granted, this technique may allow the Parliament to fulfil some of its control function, which should also apply to agencies (cf. section V 4.2.2.3), but involving it in the actual functioning of an agency confuses accountability. Instead, the Parliament could insist on having traditional control mechanisms at its dis-posal vis- à- vis agencies (cf. section V 4.2.2.3).

3.1.2.4 Stakeholder participation on the BoardsSome Boards also have members representing specific sectorial interests.98 As a rule, these representatives do not have voting rights, the exceptions being the Cedefop, Eurofound, and EU- OSHA.

The further six agencies then have five different procedures for the appointment of non- voting stakeholder representatives and there does not seem to be a clear reason why in some cases these representatives are nominated by the Commission itself, whereas in other cases the Commission must cooperate with the Council and the Parliament to appoint stakeholder representatives. It is also unclear why the appointment of non- voting expert members to the Board should be subject to a politicized procedure.

3.1.2.5 Third countries, international organizations, and other agencies on the Boards

The Boards of some agencies are also open to participation by third countries (cf. section II 3.4). These are either members of the European Economic Area or countries with which the EU has concluded agreements providing for the partici-pation of that country in the agency. These third countries will be able to partici-pate fully in the agency, although they will not be represented by a voting member on the Board (see however section II 3.4).

97 For the EMA, see Article 54 in European Commission, COM (90) 283 final. Amendments 168 and 88 in the Parliaments resolution proposed to add two representatives of the Parliament on the Board: see Resolution of the European Parliament, OJ 1991 C 183/ 145. For the ECDC, the Commission originally proposed a Board of fifteen members, with a parity between the Commission and the Council and three members nominated by the Commission representing stakeholders. See Article 14 in European Commission, COM (2003) 441 final. The European Parliament, however, insisted on having two representatives as well: see amendment 22 in Legislative Resolution of the European Parliament, OJ 2004 C 97E/ 93. The positions of the different institutions on the composition of the Board of the ACER were already dealt with pre-viously: see text at n 88.

98 See Articles 4(2)b&c Cedefop, 34(3) EASA, 79(1) ECHA, 65(1) EMA, 11(1) EMSA, 6(1) old ENISA, 26(1)a- f ERA, 7(1) ETF, 8(1)b&c EU- OSHA, 6(1) b&c Eurofound. Under the 2013 recast ENISA Regulation, the stakeholders are no longer represented on the Board.


Because of the specific field in which they are active, the establishing regula-tions of the Fundamental Rights Agency (FRA), European Asylum Support Office (EASO), and GSA provide for the participation of representatives of other international organizations on the Boards. On the Board of the FRA, the Council of Europe has a representative with limited voting rights;99 on the Board of the EASO the United Nations High Commissioner for Refugees is represented by a non- voting member, similarly to the European Space Agency in the GSA Board.

A last specific category of observers are the EU agencies themselves. Since there is a tendency to fragment policy fields by establishing several agencies in over-lapping areas, rather than establishing ‘super- agencies’, EU agencies’ Boards will also welcome observers from other EU agencies to streamline the inter- agency cooperation, similar to the multiple working arrangements which are concluded between agencies. As a result, some regulations explicitly provide that another agency may send observers to the Board’s meetings while most of the time other agencies’ representatives will be involved in an ad hoc manner in meetings.

Indeed, apart from the express provisions in the establishing regulations them-selves, it should be noted that agencies’ Boards may invite experts to attend the meetings, allowing them to invite representatives of international organizations or third countries on an ad hoc basis.

3.1.2.6 Concluding remarksThe existing heterogeneity in the composition cannot be justified in all cases, and in some it is detrimental to the effectiveness of the Board. In the Common Approach the institutions have agreed on the following template: one rep-resentative for each Member State, two for the Commission, one for the Parliament, and where appropriate ‘a fairly limited number of stakeholders’ representatives’. The compromise on this point was a serious defeat for the Commission in light of its earlier objective of Council– Commission parity. At the same time it is a minor victory for the Parliament, which now has a legitimate claim to be represented on all Boards even if the merits of such a representation are doubtful.

The effects of this template should be gauged in light of the voting rules with which the Common Approach also deals (cf. section II 3.1.6). Bernard especially criticizes the fact that each Member State retains its representative, thus giving precedence to state interests over efficient decision- making in technical or sci-entific matters.100 Although this seems a bit exaggerated, since the Board is not concerned with the actual technical and scientific work, Bernard’s fundamental critique is correct. Applying the Common Approach still results in bloated Boards of at least thirty members. The fact that the Boards remain bloated under the

99 See Article 12(8) of the FRA Regulation.100 Elsa Bernard, ‘Accord sur les agences européennes: la montagne accouche d’une souris’,

(2012) RDUE 3, p 418.


new rules proposed by the Common Approach is also implicitly acknowledged by the institutions, since the Common Approach further suggests that Executive Boards should be established (cf. section II 3.1.8). Bernard especially takes issue with the right of non- concerned Member States to be represented on the Board of an agency,101 such as the land locked EU Member States being represented in the European Fisheries Control Agency (EFCA) and European Maritime Safety Agency (EMSA). Lastly, Bernard also criticizes the Parliament’s representation on the Board as symbolic and rightly notes that its representation obfuscates its role as ‘controlling body’.102

3.1.3 Board members’ qualitiesOnly the establishing regulations of the recent agencies elaborate on the qual-ities a Board member should possess. Because of the Boards’ dual function a member should have both administrative and policy expertise. Both types of ex-pertise are explicitly mentioned in a number of regulations,103 whereas in other regulations Board members should only have expertise in the specific policy field concerned;104 still other regulations generally refer to ‘relevant experience and expertise’.105

In addition to these two elements, some regulations also insist on a broad geographic distribution of the Board members,106 while others prescribe a bal-ance between men and women on the Board.107 The Common Approach only refers to the basic two types of expertise, giving priority to the technical expertise as it provides that the members should be appointed based on their knowledge of the agency’s core business, taking account of relevant managerial, adminis-trative, and budgetary skills.108 Remarkably, following the 2013 review of the EMSA Regulation, EMSA Board members still only require relevant technical expertise.109

101 Ibid, p 419. As regards land locked states’ representation in the EFCA and EMSA, this may of course be justified if they are important flag states.

102 Ibid.103 See Articles 34(1) EASA, 79(2) ECHA, 25(1) EFSA, 65(1) EMA, 6(3) new ENISA, and

12(1) FRA.104 See Articles 25(3) EASO, 10(2) EIGE, 24(2) EFCA, 11(1) EMSA, 6(2) old ENISA, 13(3)

eu- LISA, and 21(2) Frontex.105 See Articles 14(2) ECDC, 26(1) ERA, and 5(2) new GSA.106 See Articles 25(1) EFSA and 65(1) EMA. This requirement seems understandable in the case

of the EFSA, since the Member States are not represented on its Board, but the same does not hold for the EMA.

107 See Articles 4(9) Cedefop, 10(2) EIGE, 7(2) ETF, 6(2) Eurofound, and 8(2) EU- OSHA, and recital 22 of the preamble to the FRA Regulation. In the case of Cedefop the Board is also called upon to ensure a balance between men and women when appointing the Chair and Vice- Chairs of the Board. Even if the Common Approach does not contain a provision to this end, following the 2013 review of the EMSA Regulation, this requirement is also provided for the EMSA Board. See Article 1(5) of Regulation (EU) 100/ 2013, OJ 2013 L 39/ 30.

108 See point 10 of the Common Approach.109 Article 11(1) EMSA has remained unaltered on this point.


3.1.4 Tenure of Board membersUsually, the establishing regulations prescribe the period for which Board mem-bers are appointed, but for older agencies no pre- defined term is established.110 When the Board is composed of the heads of the NRAs a term is not provided either.111 A third category for which no term is provided are the former third pillar112 and second pillar agencies.

The regulations which do provide for terms are not uniform on this point. For the ESAs it is two and half years;113 in other agencies it is three years;114 for others four years;115 for still others five years;116 and the members of the EIT Board are appointed for six years.117

Another major point on which the establishing regulations differ is the issue of renewability of the mandate. Obviously, in those cases where no term is originally provided, there is no mention of the possibility to renew the term. In other cases, the mandate of a Board member is usually118 renewable,119 although often only once.120

The Common Approach proposes a four- year renewable mandate, without lim-iting the number of times the mandate may be renewed.

3.1.5 Chairs of the BoardsThe Chairs and possible Vice- Chairs are usually elected by the members of the Board from among its voting members, but a number of exceptions to this rule exist.

A first exception are the JHA agencies, the Chair of the CEPOL correspond-ing to the rotating presidency of the Council, whereas the Chairperson of the

110 This is so for the EEA, EMCDDA, ENISA, CPVO, and OHIM. It should be noted that the ENISA is an agency of the third wave, but because it was originally established for only five years, no term of membership was provided for in the original regulation. In the recast regulation of 2013 a term of membership of four years is provided for: see n 115.

111 This is the case for the ACER Board of Regulators, the BEREC Office, and the ESAs’ Boards of Supervisors.

112 This is the case for CEPOL and Europol.113 See Article 45(1) ESAs; however, this is only for the ESAs’ Management Boards, see n 111.114 See Articles 5(1) CdT, 4(3) Cedefop, 25(5) EASO, 10(3) EIGE, 65(4) EMA, 6(3) Eurofound,

and 8(3) EU- OSHA.115 See Articles 12(1) ACER, 14(2) ECDC, 79(3) ECHA, 25(2) EFSA, 11(3) EMSA, 13(4) eu-

LISA, 9(1) Eurojust, 21(1) Frontex, and 5(2) new GSA. Regarding Eurojust, its College is composed of the ‘national members’ provided for in Article 9 of the establishing decision. That Article provides they are appointed for ‘at least’ four years.

116 See Articles 34(1) EASA, 24(3) EFCA, 11(3) old EMSA, 26(3) ERA, 7(3) ETF, 12(3) FRA, and 5(2) old GSA.

117 See Article 1(2) of the Annex to the EIT Regulation.118 The terms of the Board members in the following agencies are non- renewable: EIGE, EIT,

and FRA.119 See Articles 5(2) CdT, 4(3) Cedefop, 34(1) EASA, 25(5) EASO, 24(3) EFCA, 65(4) EMA,

11(3) new EMSA, 6(3) Eurofound, and 8(3) EU- OSHA.120 See Articles 12(1) ACER, 14(2) ECDC, 79(3) ECHA, 25(2) EFSA, 11(3) old EMSA, 26(3)

ERA, 45(1) ESAs, 7(3) ETF, 13(4) eu- LISA, 21(1) Frontex, and 5(2) GSA. The regulations of the ECDC, ESAs, and Frontex speak of extendable mandates.


Europol Board is designated by and from the three Member States which have prepared the eighteen- month programme. For Eurojust, the College elects its President from among its members subject to the approval by the Council by qualified majority voting (QMV). A second exception relates to the Chairpersons of the ESAs, which are nominated by their Boards of Supervisors (without being voting members themselves), but the Parliament can block their appointment. A third exception are the agencies the Boards of which are chaired by the Commission.121 A final specific situation may be found at the eu- LISA, which works in a field characterized by variable geometry. As a result the Chair may only be a representative of a Member State bound by all relevant EU legislation.

The terms of office of the Chair and Vice- Chairs of these bodies also vary between one and five years.122 Sometimes the regulation expressly provides the mandate is renewable.123 Unfortunately, the Common Approach is silent on this issue.

3.1.6 Decision- making on the BoardsThe voting procedures are not uniform across the Boards, but the more recent regulations become increasingly detailed and fine- tuned on this issue. Some regu-lations provide for an absolute majority.124 In other regulations the threshold is a simple majority of the votes cast,125 a simple majority,126 a simple majority of the members,127 or a majority of the members.128 A two- thirds majority is also

121 See Articles 4(3) CdT, 25(1) EFCA, 7(4) ETF.122 For Chairs of a one- year term, see Articles 4(4) BEREC Office, 8(4) EU- OSHA, 6(4)

Eurofound. For Chairs of a two- year term, see Articles 12(2) ACER, 14(3) ECDC, 80(2) ECHA, 25(4) EFSA, 14(2) eu- LISA, 22(2) Frontex, and 5(4) new GSA. For Chairs of a two- and- a- half- year term, see Articles 6(3) old ENISA, 12(5) FRA, and 5(4) old GSA. For Chairs of a three- year term, see Articles 35(2) EASA, 26(2) EASO, 8(2) EEA, 25(2) EFCA, 10(4) EIGE, 2(4) EIT, 65(5) EMA, 9(2) EMCDDA, 12(2) EMSA, 7(1) new ENISA, 38(2) CPVO, 27(2) ERA, 128(2) OHIM. The Chairpersons of the ESAs are a special case in that they chair both the Board of Supervisors and the Management Board. Given their importance, they are appointed for a period of five years. No specific term is provided in the establishing acts of other agencies: see Articles 4(3) CdT, 7(4) ETF.

123 This issue is normally dealt with in the agencies’ establishing acts. For inter alia the EEA this is not the case, but the Rules of Procedure of the EEA Board provide in Article 1(1) that the Chair may be re- elected. The establishing acts of other agencies provide for the renewability explicitly: see Articles 14(3) ECDC, 25(4) EFSA, 6(3) old ENISA, 7(1) new ENISA, 38(2) CPVO, 128(2) OHIM. Other establishing acts specify the term may only be renewed once: see Articles 12(2) ACER, 35(2) EASA, 26(2) EASO, 80(2) ECHA, 25(2) EFCA, 2(4) EIT, 65(5) EMA, 9(2) EMCDDA, 12(2) EMSA, 27(2) ERA, 14(2) eu- LISA, 22(2) Frontex, 12(5) FRA, and 5(4) GSA.

124 See Articles 28(1) EASO, 4(7) Cedefop, 27(1) EFCA, 9(1) new ENISA, 8(6) EU- OSHA, 6(6) Eurofound, 20(5) Frontex, and 5(6) new GSA.

125 See Article 12(8) FRA.126 See Article 129(5) OHIM. Here the Commission’s representative does not count towards

the majority.127 See Articles 15(1) ECDC and 41(1) CPVO. For the CPVO only the Member States’ repre-

sentatives have voting rights.128 See Articles 25(5) EFSA, 10(8) EIGE, 6(4) old ENISA, 16(1) eu- LISA.


a common threshold: either two- thirds of the members,129 or of the members present.130

Usually every regular member (thus excluding the observers or members without voting rights) has one vote, but in the case of the ETF and Translation Centre for the Body of the European Union (CdT) the Commission only has one vote. In the OHIM and CPVO the Commission members do not have voting rights.131 Lastly, the Member States’ votes have equal weight, since QMV is not prescribed.

In the more recent regulations the applicable threshold is determined by the nature of the decision to be adopted. Thus, if the default voting rule does not pre-scribe a two- thirds majority, some regulations provide for such a higher threshold to adopt the budget,132 or the internal rules of the agency.133 The annual work programme is sometimes adopted with a two- thirds134 or even three- fourths ma-jority.135 The appointment and/ or dismissal of the Director is sometimes taken by a special majority of two- thirds,136 three- fourths,137 or even four- fifths.138 Similarly, a special provision is often included on the Rules of Procedure of the Board, to be adopted by a two- thirds139 or a three- fourths majority.140 If the Board has been granted the power to adopt linguistic arrangements for the agency, this is done by unanimity.141

While a differentiation in the voting rules depending on the nature of the deci-sion to be taken makes sense, the existing heterogeneity is again hard to account for. The Common Approach codified the trend whereby an absolute majority ap-plies for ‘current business matters’ and a two- thirds majority for the important decisions.142 While the recast ENISA Regulation has taken over these new voting rules,143 the 2013 review of the EMSA Regulation has not.

A combined reading of the provisions of the Common Approach on the Board’s composition and its voting rules shows a weakened role for the Commission and

129 See Articles 7(1) iuncto 4(9) BEREC Office, 6(2) CdT, 10(7) CEPOL, 37(1) EASA, 82 ECHA, 8(3) EEA, 65(6) EMA, 9(1) EMCDDA, 14(1) EMSA, 29 ERA, 8(1) ETF, 37(8) Europol, 5(6) old GSA.

130 See Article 12(4) ACER. 131 See nn 126 and 127.132 See Articles 15(1) ECDC, 10(8) iuncto 10(6)d EIGE, 6(4) old ENISA, 9(2) new ENISA,

12(8) iuncto 12(6)d FRA. In the case of EU- OSHA this decision requires a general absolute ma-jority and a majority of the government group: see Article 8(6) EU- OSHA.

133 See Articles 15(1) ECDC, 6(4) ENISA, 12(8) iuncto 12(6)g FRA.134 See Articles 15(1) ECDC, 10(8) iuncto 10(6)a EIGE, 6(4) old ENISA, 9(2) new ENISA,

12(1) j eu- LISA, 12(8) iuncto 12(6)a FRA, 5(6) new GSA. In the case of EU- OSHA this decision requires a general absolute majority and a majority of the government group: see Article 8(6) EU- OSHA.

135 See Articles 29(1)f EASO and 20(2) c Frontex.136 See Articles 15(1) ECDC, 30(3) EFCA, 10(8) iuncto 10(6)c EIGE, 6(4) old ENISA, 9(2) new

ENISA, 18(1) eu- LISA, 12(8) iuncto 12(6)c FRA, 26(2) Frontex, and 5(6) new GSA.137 See Articles 39(1) EASA and 7(2) old GSA. 138 See Article 16(1) EMSA.139 See Articles 15(1) ECDC and 6(4) old ENISA, 9(2) new ENISA.140 See Article 29(1)a EASO.141 See Articles 37(1) EASA, 14(5)f ECDC, 35(1) ERA, 8(2) ETF, 47(2) Europol, 25(2) iuncto

12(8) FRA.142 See point 13 of the Common Approach. 143 See Article 9(2) new ENISA.


the relativity of the Parliament’s representative. First, the Commission did not succeed in getting a Council– Commission parity, and there is no special weigh-ing mechanism as an alternative. The effects of successive enlargements of the EU are therefore not countered even if each new EU Member State dilutes the power of the old representatives. This is so for both the existing Member States and the Commission, but the Member States collectively gain from an accession (cf. sec-tion V 4.2.2.2).

Lastly, it should be noted that as regards decision- making the Common Approach does contain an important new element, namely the establishment of an alert or warning system which may be activated by the Commission if it has serious concerns that the Board is about to take a decision outside the agen-cy’s mandate or in violation of EU law or its policy objectives. In such cases the Commission may ask the Board to refrain from taking a decision and may for-mally inform the Parliament and Council.144 However, the Common Approach does not provide that such a mechanism must be written in the establishing acts of agencies or in the Rules of Procedure of the Boards and unless this is done, the mechanism remains an informal arrangement, with the Board under no obliga-tion to refrain from taking action when the Commission so requests.

3.1.7 Meetings of the BoardsUsually the establishing act prescribes that the Board should meet a number of times a year,145 and if not the Board’s Rules of Procedure will provide for a minimum.146 Apart from these ‘ordinary’ meetings, some establishing acts also provide the Chair may convene a meeting on his own instance,147 or if one- third of the members so requests,148 but in the case of the ETF a single majority is required.149 Sometimes the Commission may also have a meeting convened.150

144 See point 59 of the Common Approach.145 For Boards meeting at least once a year, see Articles 4(6) Cedefop, 26(3) EFCA, 10(10) EIGE,

9(3) EMCDDA, 8(2) new ENISA, 39(3) CPVO, 8(3) ETF, 8(5) EU- OSHA, 6(5) Eurofound, 129(3) OHIM. For Boards meeting twice a year, see Articles 12(3) ACER, 6(1) CdT, 10(6) CEPOL, 36(3) EASA, 27(2) EASO, 14(3) ECDC, 13(3) EMSA, 6(5) old ENISA, 28(2) ERA, 15(1) eu- LISA, 37(7) Europol, 12(9) FRA, 23(3) Frontex, 5(5) GSA. The EIT Board meets thrice a year: see Article 3(3) EIT Annex. The BEREC Office Board meets four times a year: see Article 7(1) iuncto 4(6) BEREC Office. For the ESAs this is five times a year: see Articles 45(3) ESAs. In the following agen-cies’ establishing acts no minimum is provided for: ECHA, EEA, EFSA, EMA, Eurojust.

146 For the EEA, ECHA, and EMA this is twice a year. For EFSA it is four times a year. For Eurojust it is once a week.

147 Pursuant to the Articles cited in n 145, this is so for the ACER, BEREC Office, EASA, EASO, ECHA, EFCA, EIGE, EMSA, ENISA, CPVO, ERA, ESAs, eu- LISA, FRA, Frontex, GSA, and OHIM.

148 Pursuant to the Articles cited in n 145, this is so for the ACER, BEREC Office, Cedefop, EASA, EASO, ECDC, ECHA, EIGE, EMSA, ENISA, CPVO, ESAs, eu- LISA, EU- OSHA, FRA, Frontex, GSA, and OHIM.

149 The situation for the ERA is different as a majority of the members or one third of the Member States’ representatives may have the Board convene. This second possibility is also provided for in the EFCA Regulation.

150 Pursuant to the Articles cited in n 145, this is so for the ACER, BEREC Office, EFCA, EMSA, CPVO, ERA, eu- LISA, and OHIM.


In the case of the eu- LISA, the Executive Director may also have the Board convene.

The variety of the frequency of ordinary meetings, ranging from once a week to once a year, is one of those differences which are indeed wholly justifiable, since the policy field of some agencies indeed requires constant monitoring and consultation. However, the differences in the possibility to convene extraordinary meetings are a different matter.

3.1.8 Executive BoardsIt was noted previously that the sheer size of most Boards of EU agencies im-pedes efficient and effective decision- making. For this reason some establishing regulations also (allow to) establish miniature Boards.151 These miniature Boards are called Bureaus,152 Executive Bureaus,153 Executive Boards,154 or Executive Committees.155

Usually the Executive Boards are composed of the Chair and Vice- Chairs of the Board in addition to one or two representatives of the Commission and possible stakeholders.156 The function of the Executive Board is to prepare the decisions of the Management Board. Although formally the Management Board stays in charge it is clear that the Executive Board will become more important, the Management Board becoming a forum in which the Member States may exercise a droit de regard.

The Common Approach on Decentralised Agencies also mentions the pos-sibility of introducing an Executive Board if this promises more efficiency. As regards the composition of the Executive Board, the Common Approach only mentions the presence of a Commission representative.157 While the Commission thus failed to promote the principle of a Council– Commission parity in the Management Boards of agencies (or such a parity when deciding on budgetary and administrative matters), its ensured membership in the Executive Boards and the Common Approach’s endorsement and promotion of this body is an interest-ing (but far from equivalent) alternative.

3.2 The Directors

3.2.1 Tasks and nomenclature of the DirectorsThe second major actor within an agency is the Director, who represents the agency externally. He is also responsible for the day- to- day administration of the

151 The establishment is not mandatory under the Frontex and EASO regulations: see Articles 29(2) EASO and 20(7) Frontex.

152 See Articles 4(8) Cedefop, 8(2) EEA, 8(8) EU- OSHA, 6(8) Eurofound.153 See Article 20(7) Frontex. 154 See Article 13 FRA.155 See Articles 29(2) EASO and 10 EMCDDA.156 The employers’ and employees’ organizations in the Cedefop, EU- OSHA, and Eurofound,

and the CoE representative in the FRA, also sit on the Executive Board. The EEA regulation only provides that the Board shall elect a Bureau, without further details.

157 See point 10 of the Common Approach.


agency and assists the Board by drafting the agency’s key documents. Just like the Board, the Directors do not have a unified name and also sometimes go by the names of Executive Director, Administrative Director, Administrative Manager, or President in the establishing acts.158 Although the Common Approach recog-nizes that Directors all perform the same or similar tasks, it does not address the issue of these differences in title.

3.2.2 Appointing the DirectorMore important differences may be noted in the appointment procedures. The basic procedure is that the Directors are appointed by the agencies’ Boards following a pro-posal by the Commission.159 Even to this basic rule there are some exceptions, since for the oldest agencies (Cedefop and Eurofound) the basic scenario is reversed.160 The first two decision- making agencies also come under a different regime, since the Directors are appointed by the Council following either a proposal by the Board (OHIM) or the Commission having received the opinion of the Board (CPVO).161 The Europol Director is also appointed by the Council following a proposal by the Board.162 The CEPOL decision provides that the Board will appoint the Director following a proposal by a selection committee rather than by the Commission.163 The situation of the FRA is again different in that the Commission will adopt a list of candidates who are invited before the Parliament and Council. The latter two institutions then communicate their preferences to the FRA Board.164 Although only the CEPOL and Eurojust decisions mention the involvement of a selection committee, it should be noted that such a committee will be established for every procedure, pursuant to Article 30 of the Staff Regulations.165

These differences are hard to justify and become problematic if one takes into account that the Directors are generally accountable before the agency’s Board,166

158 For the agencies headed by a Director, see Articles 3(c) ACER, 9 CdT, 4(1)c Cedefop, 11 CEPOL, 13(b) ECDC, 9(c) EIGE, 4(1)c EIT, 11 EMCDDA, 10 ETF, 7a(c) EU- OSHA, 5(c) Eurofound, 36(b) Europol, 6 EUISS, 8 EUSC, 11(d) FRA. For the agencies headed by an Executive Director, see Articles 38 EASA, 24(b) EASO, 76(1)b ECHA, 9 EEA, 29 EFCA, 24(b) EFSA, 56(1)g EMA, 15 EMSA, 5(b) old ENISA, 4(b) new ENISA, 30 ERA, 51(1) ESAs, 11(1)b eu- LISA, 25 Frontex, and 3 GSA. The BEREC Office (see Article 8) has an Administrative Manager, Eurojust (see Article 29) has an Administrative Director and the OHIM and CPVO both have Presidents (see Articles 42 CPVO and 124 OHIM).

159 For Eurojust, the College appoints the director based on a list composed by the selection board, on which the Commission is represented.

160 See Articles 6(1) Cedefop and 8(1) Eurofound.161 See Articles 43(1) CPVO and 125(1) OHIM. 162 See Article 38(1) Europol.163 See Article 11(1) CEPOL. 164 See Article 15(2) FRA.165 See Règlement 259/ 68 (CEE) du Conseil, JO 1968 L 56/ 1.166 See Articles 8(1) BEREC Office, 9(3) CdT, 7(5) Cedefop, 11(5) CEPOL, 31(1) EASO, 12(4)

EIGE, 9(2) EEA, 11(4) EMCDDA, 10(5) ETF, 11(3) EU- OSHA, 9(4) Eurofound, 15(5) FRA, 25(4) Frontex, 5(2) EIT, 18(6) eu- LISA, 38(5) Europol. For a number of agencies accountability before the Board is only foreseen by a provision which clarifies that the Board exercises disciplinary authority over the Director: see Articles 13(9) ACER, 33(2)h EASA, 14(5)a ECDC, 78 ECHA, 23(2)f EFCA, 10(2)i EMSA, 25(2)g ERA, 43(8) ESAs, 6e old GSA, 6(5) new GSA, whereas in Eurojust the Director works under the authority of the College and its President: see Article 29(4)


the meetings of which he usually also attends.167 Ultimately the Directors should be appointed by and be accountable before the same body.

3.2.2.1 Refining the appointment procedureFor some, although not all, of the recent agencies, the appointment procedure has also been elaborated.168 As a result, the Commission is often required to propose more than one candidate.169 How many candidates should be on the Commission’s list is not always provided for, meaning two candidates would be enough.170 In other cases, a minimum of three is provided.171

Again a trend may be noted whereby the establishing acts of the more recent agencies also provide for a greater role for the Parliament. The candidate selected by the Board will then be invited to the Parliament to make a statement and answer questions.172 For the most recent agencies such as the EASO and the eu- LISA it is provided that the Parliament may adopt an opinion on the Director- designate and the Board will then inform the Parliament of the manner in which it has taken into account that opinion. Although the establishing regulations of the agencies at the most provide for a hearing of the Director- designate, the Commission and Parliament have agreed in their interinstitutional framework

Eurojust. There is no clear provision on the accountability of the Director in the establishing acts of the EFSA, EMA, or ENISA. Articles 43(4) CPVO and 125(4) OHIM provide that the Council exercises disciplinary authority over the President.

167 The Director attends the meetings according to the establishing acts of 9(2) BEREC Office, 12(3) ACER, 10(5) CEPOL, 36(2) EASA, 27(1) EASO, 14(6) ECDC, 81(2) ECHA, 26(2) EFCA, 25(10) EFSA, 12(4) EIGE, 9(3) EMCDDA, 13(2) EMSA, 6(5) ENISA, 8(3) new ENISA Regulation, 28(1) ERA, 40(7) & 45(2) ESAs, 15(2) eu- LISA, 9(3) Eurofound, 37(5) Europol, 15(5) FRA, 23(2) Frontex, 5(5) GSA, and 129(2) OHIM. The EMA Regulation provides in Article 61(4) that the Director attends all meetings of the agency or its committees. The ACER, GSA, and OHIM regulations thereby clarify that the Director or President attends the meetings of the Board unless the Board/ Chair decides otherwise. As a result, the establishing acts of the CdT, Cedefop, EEA, EIT, CPVO, ETF, and Eurojust do not explicitly provide the Director may attend the meetings, but the Boards of those agencies have still provided for this possibility in their Rules of Procedure.

168 This was not the case for CEPOL, EASA, EEA, EMCDDA, EMSA, ERA, or Europol. The EMSA regulation merely provided that the Commission ‘may’ propose one or several candidates. Following the 2013 review of the EMSA Regulation, the Director’s appointment procedure has indeed been elaborated. See Article 3(8) of Regulation (EU) 100/ 2013, OJ 2013 L 39/ 30.

169 However, Articles 39(1) EASA and 11(1) EMCDDA for instance only speak of the Director being appointed by the Board on a proposal from the Commission and Articles 16(1) EMSA and 31(1) ERA provide that the Commission ‘may propose a candidate or candidates.’ Following the 2013 review of the EMSA Regulation the Commission must propose at least three candidates.

170 See Articles 30(1) EASO, 17(1) ECDC, 84(1) ECHA, 12(1) EIGE, 30(1) EFCA, 26(1) EFSA, 7(2) ENISA, 24(2) new ENISA, 18(1) eu- LISA, 15(2) FRA, 26(1) Frontex, and 15b(2) new GSA. The EFCA regulation indeed explicitly states that the list should be made up of ‘at least’ two candidates.

171 See Articles 16(2) ACER, 16(1) new EMSA, 10(1) ETF, 38(1) Europol, and 7(2) old GSA.172 See Articles 16(2) ACER, 30(1) EASO, 17(2) ECDC, 84(1) ECHA, 26(1) EFSA, 12(1) EIGE,

64(1) EMA, 11(2) EMCDDA, 7(2) ENISA, 24(2) new ENISA, 10(1) ETF, and 18(2) eu- LISA. Although these are also recent agencies, a similar procedure is not provided for in the case of Frontex or the EFCA, although in practice an exchange of views with the nominee may still be organized. The EMCDDA Regulation explicitly provides that this procedure only applies for the first appoint-ment of the Director.


agreement to have all the nominees for the post of Director appear before the rele-vant Parliamentary committee.173

The situation is different for the ESAs. Similarly to its role in appointing the Chairperson of the Board, the Parliament will have to confirm the candi-date selected by the Board of Supervisors.174 The Parliament therefore has real power in appointing the Director. Unfortunately the Common Approach has not codified these recent trends on the Parliament’s involvement.175 Instead, it merely prescribes that the Board should be the appointing authority, choosing from a list drawn up by the Commission. That the Common Approach was only a partial codification is clear from the 2013 review of the EMSA Regulation, which lays down a more elaborate appointment procedure for the Director than the one provided for in the Common Approach,176 and this in line with the original Commission proposal which predated the Common Approach.177

3.2.2.2 The profile of the DirectorWhat is the profile of an agency Director? Given that he is the external represen-tative and the head of an independent agency, except in the cases of the ESAs, one would expect that the founding acts also stress his independence. However, this is not always the case, especially with the older agencies. The more recent agencies indeed stress his independence,178 sometimes adding that he shall not seek nor follow any instructions from other bodies;179 for some agencies only the latter requirement is prescribed.180 Secondly, one would expect that the Director should also have administrative and policy expertise. Again, the older agencies do not prescribe anything of the sort, while the more recent agencies do insist on expert knowledge,181 and some on both expert knowledge and management or

173 See point 32 of the Framework Agreement of the European Parliament and the Commission, OJ 2010 L 304/ 47.

174 See Article 51(2) ESAs.175 Following the conclusion of the Common Approach, the Committee on the Environment,

Public Health and Food Safety of the European Parliament suggested that the EEA Regulation should be updated as regards the procedure for the appointment of the Director. The Parliament in plenary agreed to this and requested the Commission to adopt a legislative proposal to this end. See European Parliament resolution of 11 September 2013 with recommendations to the Commission on Parliament’s rights in the appointment procedure of future Executive Directors of the European Environment Agency.

176 See n 175.177 See European Commission, COM (2010) 611 final.178 See Articles 16(1) ECDC, 83(1) ECHA, 15(1) EMSA, 7(1) ENISA, 11(1) new ENISA, 30(1)

ERA, 11 ETF, 8(2) Eurofound, and 15(5) FRA. The GSA Director’s independence was only stressed in preamble but now also in Article 7 new GSA.

179 See Articles 38(1) EASA, 31(1) and (2) EASO, 52 ESAs, 17(2) eu- LISA, and 25(1) Frontex.180 See Articles 8(1) BEREC Office, 16(1) ACER, 29(1) EFCA. However, in Article 21 of the

regulation on ‘declaration of interests’ a reference is made to the BEREC Office Administrative Manager’s independence.

181 See Articles 8(2) BEREC Office, 16(2) ACER, 39(1) EASA, 30(1) EFCA, 5(1) EIT Annex. Article 8(2) Eurofound simply provides the Director is chosen on the grounds of his competence.


administration skills.182 This means that for a number of older agencies, but also those active in JHA, neither technical expertise, nor management skills, nor inde-pendence are formally required to fill the post of Director.183

The Common Approach does not fully recognize this serious shortcoming. It does provide that the selection procedure for appointing the Director should guarantee a rigorous evaluation of candidates and a high level of independ-ence, but it does not prescribe that the ‘rigorous evaluation’ should test both the candidate’s expertise and his experience or skills in leading a bureaucratic organization.

3.2.3 Tenure of the DirectorThe term of office of a Director is usually five years,184 and is always renewable.185 Again, for some recent agencies the procedure for the renewal of the mandate has been elaborated. Those establishing acts explicitly prescribe that an evaluation should precede the renewal and sometimes further provide that the second term is not a full one, but only lasts three years.186 In the acts establishing the recent agencies, it is furthermore provided that the Parliament may express its view on a proposed extended mandate or even invite the Director for a hearing.187

The Common Approach codifies some of these trends but unfortunately not all. It confirms that a positive evaluation of the Director’s first mandate should pre-cede the decision to extend the mandate and it confirms that the second mandate should not be preceded by a fresh appointment procedure.188 Still, the Common Approach does not refer to the reduced term of office for the second mandate. Evidently, since the Common Approach did not recognize Parliament’s role in

182 See Articles 30(1) EASO, 84(1) ECHA, 16(1) EMSA, 7(2) old ENISA, 31(1) ERA, 51(2) ESAs, 10(2) ETF, 18(1) eu- LISA, 15(1) FRA, 26(2) Frontex, 7(2) old GSA, 15b(2) new GSA. In the 2013 recast ENISA Regulation, this requirement has been moved to the preamble to the Regulation: see recital 45.

183 This the case for the CdT, Cedefop, CEPOL, EEA, EIGE, EFSA, EMA, EMCDDA, CPVO, EU- OSHA, Eurojust, Europol, and OHIM.

184 For a small number of agencies this is four years: see 11(1) CEPOL, 5(1) EIT Annex, Article 38(1) Europol. For the BEREC Office, it is three years: see Article 8(3).

185 No limit on the number of renewals is provided for in acts establishing CdT, Cedefop, EEA, EFSA, CPVO, EU- OSHA, Eurofound, and OHIM. For the other agencies, it is provided the term may be renewed or extended only once: see Articles 16(4) ACER, 8(4) BEREC Office, 11(1) CEPOL, 39(4) EASA, 30(2) EASO, 17(1) ECDC, 84(2) ECHA, 30(4) EFCA, 12(2) EIGE, 5(1) EIT, 64(1) EMA, 11(2) EMCDDA, 16(2) EMSA, 31(3) ERA, 51(3) ESAs, 10(1) ETF, 18(4) eu- LISA, 29(2) Eurojust, 38(1) Europol, 15(3) FRA, 26(5) Frontex and 7(1) old GSA, and 15b(3) new GSA.

186 See Articles 16(4) ACER, 30(2) EASO, 10(1) ETF, 18(4) eu- LISA, and 15(3) FRA. Following the review of the EMSA and GSA Regulations, these also provide for a shorter, albeit four- year, second term: see Article 16(2) new EMSA, 15b(3) new GSA.

187 See Articles 16(5) ACER, 8(4) BEREC Office, 30(3) EASO, 24(4) and (5) new ENISA, 10(1) ETF, 18(5) eu- LISA, 15(3) FRA, and 15b(2) new GSA.

188 For this issue, see European Commission, COM (2005) 190 final. See also Madalina Busuioc and Martijn Groenleer, ‘Wielders of Supranational Power? The Administrative Behaviour of the Heads of European Union Agencies’, in Busuioc, Groenleer, and Trondal (eds), The Agency Phenomenon in the European Union, Manchester, Manchester University Press, 2012, p 136.


the initial appointment procedure, it is also silent on Parliament’s role in the pro-cedure extending a Director’s mandate.

3.2.4 Dismissing the DirectorFor a long time the establishing acts of the agencies were completely silent on the issue of removal. As the Director’s contract is governed by the Staff Regulations, which allow the compulsory resignation by the appointing authority, a Director could still be dismissed. Most recent regulations do provide a specific removal pro-cedure.189 Typically ‘the same procedure’ as that for the appointment will apply.190 These provisions do not excel in clarity, however. Other agencies’ establishing acts describe the procedure more fully, providing that the Commission will make a pro-posal for dismissal to the Board,191 or that both one- third of the Board members or the Commission may make such a proposal,192 or that the Board takes this dec-ision independently.193 Lastly, it should be noted that the Directors are removable at will, given the silence of the establishing regulations on the reasons for dismissal.

Here again these different procedures may be rationalized. Principally, the pro-cedure for dismissal should mimic the appointment procedure as much as pos-sible. Starting from this basic scenario, different variations could be envisaged depending on how agencies are conceptualized in the EU institutional order: if the Commission is viewed as the EU’s prime executive actor, it would make sense to strengthen the Commission in this procedure. Similarly, if the Parliament is to exercise control over the EU’s executive function, it could be granted the power to request the Board to dismiss the Director (cf. section V 4.2.3.1.2). The Common Approach only provides that a procedure for dismissal should be foreseen and that it should ‘mirror’ the appointment procedure.194

3.3 Work programmes and annual reports

The work programme and annual report are, in a sense, two sides of the same coin. In the programme, the agency will set out its policies and objectives for the coming year. In the report, it reflects on its activities in the preceding year.

189 Articles 29(1)b EASO and 2(2)j EIT Annex provide that the Board may appoint and dismiss the Director but no specific procedure is provided for.

190 See Articles 14(5)a ECDC, 84(1) ECHA, 10(6)c EIGE, 16(1) EMSA, 31(1) ERA, and 26(4) Frontex. Effectively the same procedure is provided by Articles 16(7) ACER, 43(1) CPVO, 29(4) Eurojust, 38(7) Europol, and 125(1) OHIM. The Europol decision provides further rules should be adopted for the appointment and dismissal procedures, and the Europol Board has adopted a deci-sion to this end. See Decision of the Europol Management Board, OJ 2009 L 348/ 3.

191 See Articles 39(1) EASA, 64(1) EMA, 10(5) ETF.192 See Articles 30 EFCA, 15(7) FRA, 15b(4) new GSA.193 See Articles 26(1) EFSA, 7(2) ENISA, 24(7) new ENISA, 18(7) eu- LISA, 7(2) old GSA.

Pursuant to Article 51(5) ESAs the Board of Supervisors rather than the Board takes this decision and it should be noted the dismissal procedure is not parallel to the appointment procedure, since no role is provided for the Parliament.

194 See point 19 of the Common Approach.


3.3.1 The work programmeAs a result, the programme is one of the most important documents for the agency on all accounts, as it sets out which activities the agency will develop in the coming period. From an autonomy perspective it is highly relevant to ascertain which procedure applies to the adoption of the work programme and which actors are involved.195 From an accountability perspective the im-portance of the work programme is sometimes underestimated,196 although it should be evident that the work programme will function as a benchmark against which the actual activities of the agency over the period concerned should be assessed.

Virtually all establishing acts provide that the agency should have a work pro-gramme, and even if there is no such obligation the agencies concerned do so of their own accord.197 The basic procedure provides that the agency’s Director sub-mits a draft to the Board which adopts the final work programme.198 Sometimes this basic procedure is not elaborated,199 but usually the Commission must also give an opinion on the work programme.200 For the Cedefop, the Commission must agree to the work programme, while for the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) a negative opinion of the

195 Martijn Groenleer, The Autonomy of European Agencies: A Comparative Study of Institutional Development, Delft, Eburon, 2009, p 120.

196 Busuioc for instance notes that the ‘work programme is not an element of accountability per se given its ex ante dimension. It can, however, play a role in ex post accountability’ (em-phasis added). See Madalina Busuioc, The Accountability of European Agencies: Legal Provisions and Ongoing Practices, Delft, Eburon, 2010, p 100.

197 These agencies are the CPVO, Eurojust, and OHIM. Yet all three adopt work pro-grammes: see Inter- institutional Working Group on regulatory agencies, Annual Work Programme, 2010, Analytical Fiche Nr° 13, p 1.

198 Because of their specific structures, this is slightly different for the ESAs, where the Director prepares the programme for the Board which proposes the programme to the Board of Supervisors. For the ACER the programme is prepared by the Director and adopted by the Board after approval by the Board of Regulators.

199 See Articles 9(3) BEREC Office, 8(1) CdT, 15(5)d and 16(2)b ECDC, 83 ECHA, 25(8) and 26(2)b EFSA, 64(3) and 65(9) EMA, 2(2)a old EIT Annex. For the EIT, Article 18(1) of the regulation only prescribed the involvement of the Commission, Parliament, and Council for the adoption of the very first (triennial) work programme, the three institutions having three months’ time to address the Board with their opinions on its preliminary decision. In function of the Horizon 2020 programme the EIT Regulation was amended in 2013 and now inter alia provides that the Commission will deliver an opinion on the preliminary programme within three months and that the EIT must justify its position in case there is a disagreement. See Article 15 EIT. Under the 2013 recast ENISA Regulation, the Commission does not need to send an opinion to the Board any more, but the Director still needs to consult it when he prepares a draft.

200 See Articles 17(6) ACER, 8(1) Cedefop, 10(9)d and 11(4)c CEPOL, 33(2)c and 38(3)k EASA, 29(f ) EASO, 8(5) EEA, 23(2)c and 29(3)a EFCA, 10(6)a EIGE, 9(5) EMCDDA, 10(2)d and 15(2)a EMSA, 10(2)d and 15(2)ab new EMSA, 6(8) and 7(4)b ENISA, 25(2)c and 30(2)a ERA, 9(c) ETF, 12(1)j and 17(6)a eu- LISA, 10(1) EU- OSHA, 12 EUROFOUND, 37(10)b and 38(4)f Europol, 12(6)a FRA, 20(2)c and 25(3)a Frontex, and 6(b) GSA. For the 2013 recast ENISA Regulation, see n 199. As regards CEPOL, it should be noted that the programme is adopted by the Council rather than the Board; for Europol, the Council also needs to endorse the programme.


Commission results in a higher majority threshold for adopting the programme. Similarly in the EMSA and ERA regulations, the Commission has fifteen days to express its disagreement, in which case the Board must re- examine the pro-gramme, adopting it with a two- thirds majority (if the Commission representa-tives are in the majority) or unanimously by the Member States’ representatives. Something similar may be found in the EFCA Regulation, although the period for the Commission to express its disagreement is thirty days and no special maj-ority rules apply.

In some agencies the internal scientific body will also give its opinion,201 whereas in other agencies the body which brings together the stakeholders has this prerogative.202 For Europol’s and CEPOL’s work programme the Council needs to give its approval.

The establishing acts of the older agencies in particular do not contain fur-ther provisions on the work programme, even though it is linked to the agency’s budget. After all, the activities which the agency will be able to deploy will also depend on the subsidy from the EU budget. Because of this, some establishing acts prescribe that the adoption of the work programme will not prejudice the budgetary procedure or should be in accordance with the budget.203 Although this may further seem self- evident, only some agencies explicitly provide that the work programme should be in line with the EU’s own legislative programme in the field concerned or in line with the activities already deployed by the EU.204 Similarly, although the agency’s work programme is linked to its budget there are sometimes no clear deadlines for the Board to adopt its work programme,205 and if a deadline is provided it varies between 31 January and the end of the year n- 1,206 which is not adapted to the budgetary procedure since an estimate of revenue and expend-iture typically needs to be sent to the Commission by 15 March of the year n- 1. The establishing acts of most of the recent agencies then also prescribe that the adopted work programme should be forwarded to the Commission, Parliament,

201 See Articles 8(5) EEA, 9(5) EMCDDA, 12(6)a FRA.202 See Articles 8(1) and 10(4)a EASO, 7(4)b old ENISA, 12(5) new ENISA, 17(6)a eu- LISA. In

the case of the EU- OSHA the Advisory Committee on Safety and Health at Work, which is not an agency body, may give its opinion.

203 See Articles 13(5) ACER, 33(2)c EASA, 8(4) EEA, 23(2)c EFCA, 10(6)a EIGE, 10(2)d EMSA, 25(2)c ERA, 43(4) ESAs, 12(6)a FRA, and 20(2)c Frontex. The following provisions lay down that the programme should be in accordance with the annual budgetary procedure: Articles 29(1)f EASO and 12(1)j eu- LISA.

204 See Articles 8(1) Cedefop, 33(2)c EASA, 29(f) EASO, 14(5) ECDC, 8(3) EEA, 25(8) EFSA, 4(2) EIGE, 15(a) ENISA, 6(8) ENISA, 13(4) new ENISA, 12(1)j eu- LISA, 20(2)c Frontex.

205 See Articles 8 CdT, 8(1) Cedefop, 10(6)a EIGE, 65(9) EMA, 10(1) EU- OSHA, 12 Eurofound, 37(10)b Europol, 12(6)a FRA.

206 Before 31 January for the ECDC and EFSA. Before 31 March for the CEPOL, EMCDDA, and Europol (in the case of CEPOL and Europol this concerns a draft work programme). Before 15 June for the EEA. Before 30 September for ACER, BEREC Office, EASO, ESAs, eu- LISA, Frontex. Before 31 October for the ECHA and EFCA. Before 15 November for the GSA. Before 30 November for the EASA, EMSA, ENISA, ERA and ETF (in the case of the ETF this concerns a draft work programme). Before the end of the year is the deadline for BEREC, which has a different work programme from its Office.


and Council,207 and sometimes also to the Member States,208 and sometimes that it should be published.209

3.3.1.1 The multiannual work programmeLastly there is the question of a multiannual work programme. Although an agency evidently should not plan its functioning year by year, it is remarkable that not all agencies are required to adopt multiannual work programmes.210 The period for which they are adopted varies between three and five years, al-though most of the time no specific period is provided for.211 Still, these differ-ences are not so much a problem since the period will foremost be determined by the state of play of the policy field concerned and the challenges which have to be met.

The many differences which may be noted in agencies’ establishing acts on the issue of the work programmes are overwhelming and the Common Approach does indeed contain a great number of suggestions to come to some rational-ization. It thus proposes that a template should be elaborated so that work pro-grammes may be compared. As regards procedure, the Common Approach only mentions that the Commission should give formal advice and that the Parliament should be consulted before the multiannual programmes are adopted, explicitly stressing that such a consultation cannot be binding.

3.3.2 The annual reportThe second key document of the agency is its annual report, which mirrors the annual work programme, since the agency will report on its activities during the previous year and should link this back to the objectives it set out in the work pro-gramme. Although this might seem self- evident, the Court of Auditors in 2008

207 See Articles 13(5) ACER, 5(4) BEREC (this does not apply to the programme of the BEREC Office), 10(12) CEPOL, 33(2)c EASA, 29(1)f EASO, 83(3) ECHA, 23(2)c EFCA, 26(3)b EFSA, 65(9) EMA, 9(4) EMCDDA, 10(2)d EMSA, 10(2)d new EMSA, 7(6) ENISA, 13(6) new ENISA, 25(2)c ERA, 43(4) ESAs, 12(1)j eu- LISA, 12(6)a FRA, 20(2)c Frontex. The Europol decision pro-vides the Council will forward the programme to the Parliament.

208 See Articles 33(2)c EASA, 83(3) ECHA, 23(2)c EFCA, 26(3) EFSA, 65(9) EMA, 10(2)d (old and new) EMSA, 7(6) ENISA, 13(6) new ENISA, 25(2)c ERA.

209 See Articles 10(12) CEPOL, 83(3) ECHA, 26(3)b EFSA, 13(3) EIT, 7(6) ENISA, 13(6) new ENISA, 43(6) ESAs, 12(1)j eu- LISA. The publication requirement for the EIT and ESAs is only imposed for the multiannual programme, not for the annual programme. The Frontex, EASA, and EASO Regulations similarly provide that the programme should be produced in all languages: see Articles 32(1)c EASA, 41(2) EASO, and 27(2) Frontex. For EASA this provision may be found under the heading ‘Publication of documents’.

210 For those that must, see Articles 13(6) ACER, 14(5)d ECDC, 78(d) ECHA, 8(4) EEA, 17f EFCA, 25(8) EFSA, 10(6)a EIGE, 15(a) EIT, 9(4) EMCDDA, 43(6) ESAs, 12(2) ETF, 12(1)h eu- LISA, 12(1) Eurofound, 5(1) FRA, 20(2)i Frontex, 8a new GSA. It should be noted that the multian-nual programme of the FRA is adopted by the Council following the procedure of Article 308 EC.

211 No period is prescribed for the multiannual programmes of the ACER, ECDC, ECHA, EEA, EFSA, ESAs, eu- LISA, Frontex, and GSA; a three- year period is prescribed for the EIGE, EIT, and EMCDDA; a four- year period is prescribed for the ETF and Eurofound; a five- year period is prescribed for the EFCA and FRA.


remarked that the agencies’ annual reports were often merely descriptive and that the ‘extent to which objectives had been met and the impact of the agency’s work were questions that were seldom covered’.212

In the establishing acts these documents are usually called annual reports,213 but sometimes they are called general reports,214 annual activity reports,215 man-agement reports,216 or annual general reports.217 The establishing acts usually do not provide much guidance as to the content of the annual report. For the agen-cies of the former first pillar it is provided that upon adoption by the Board, the annual reports should be forwarded to the Commission, Parliament, Council, and Court of Auditors.218 Sometimes the report should also be forwarded to the Member States,219 the Economic and Social Committee,220 the Committee of the Regions,221 or specific partners.222 Normally the Board will submit this report to

212 Special Report 5/ 2008 of the Court of Auditors The European Union’s Agencies: Getting Results, OJ 2008 C 238/ 10, para 31. See also Busuioc, n 196, p 63.

213 See Articles 13(12) ACER, 5(5) BEREC, 8(3) CdT, 12b(1) Cedefop, 10(9)e CEPOL, 16(4) ECDC, 8(6) EEA, 3(2) EIGE, 15(b) EIT, 65(10) EMA, 9(7) EMCDDA, 10(2)b EMSA, 5(3) new ENISA, 43(5) ESAs, 10(2) EU- OSHA, 13(1) Eurofound, 32(1) Eurojust, 4(1)g FRA, 6(g) old GSA, 6(h) new GSA. The EFSA Regulation also refers once to the agency’s annual report (see Article 38(1)f ).

214 See Articles 78(a) ECHA, 23(2)b EFCA, 25(8) EFSA, 6(9) old ENISA, 25(2)b ERA, 37(10)c Europol, 20(2)b Frontex. Although the EMCDDA Regulation provides that the Board will adopt the agency’s annual report (see Article 9(7)) it equally provides that the Director will submit the general report of the agency’s activities to the Parliament (see Article 12). Similarly, the preamble to the EMSA Regulation (see recital 10) speaks of draft general reports prepared by the Director whereas the Regulation itself (see Article 10(2)b) refers to the annual report adopted by the Board. The ECDC Regulation also refers to both the annual report (see Article 16(4)) and the general report (see Article 16(3)a).

215 See Articles 13 ETF, 12(1)k eu- LISA. The EIT Regulation refers both to the annual activity report (see Article 13(3)) and the annual report (see Article 15).

216 See Articles 42(2)b CPVO and 124(2)d OHIM.217 See Articles 57 EASA, 29(1)c EASO. In the heading to its Article 10, the EU- OSHA

Regulation refers to the annual general report but in the Article itself simply to the annual report.218 See Articles 13(12) ACER, 5(5) BEREC, 8(3) CdT, 12b(1) Cedefop, 33(2)b EASA, 29(1)c

EASO, 16(4) ECDC, 83(3) ECHA, 8(6) EEA, 23(2)b EFCA, 26(3) EFSA, 10(6)b EIGE, 65(10) EMA, 9(7) EMCDDA, 10(2)b EMSA, 7(7) ENISA, 5(3) new ENISA, 25(2)b ERA, 43(5) ESAs, 13(4) ETF, 12(1)k eu- LISA, 10(2) EU- OSHA, 13(1) Eurofound, 12(2)b FRA, 20(2)b Frontex, 6(g) old GSA, 6(h) new GSA. It should be noted that the annual report of the ERA should not be sent to the Court of Auditors. General notable exceptions (apart from the CFSP and JHA agencies) are the OHIM and CPVO, where the Director adopts the management report and submits it to the Board and Commission (in the case of OHIM also to the Parliament) without further details, and the EIT, where it is only provided that the Board adopts the report before 30 June.

219 See Articles 33(2)b EASA, 83(3) ECHA, 8(6) EEA, 23(2)b EFCA, 65(10) EMA, 9(7) EMCDDA, 10(2)b EMSA, 25(2)b ERA, 13(4) ETF, 10(2) EU- OSHA, 6(g) old GSA.

220 See Articles 13(12) ACER, 5(5) BEREC, 12b(1) Cedefop, 16(4) ECDC, 83(3) ECHA, 26(3) EFSA, 10(6)b EIGE, 65(10) EMA, 7(7) old ENISA, 43(5) ESAs, 13(4) ETF, 10(2) EU- OSHA, 13(1) Eurofound, 12(2)b FRA, 20(2)b Frontex, 6(g) old GSA. It should be noted that the 2013 recast ENISA Regulation no longer imposes this requirement.

221 See Articles 13(12) ACER, 5(5) BEREC, 16(4) ECDC, 26(3) EFSA, 10(6)b EIGE, 7(7) old ENISA, 12(2)b FRA. It should be noted that the 2013 recast ENISA Regulation no longer imposes this requirement.

222 This is the case for the ETF, where the report is also sent to the partner countries; the CdT, where the report is also sent to the other agencies who buy services from the CdT; and the EU- OSHA, where the report is sent to the Advisory Committee on Safety, Hygiene and Health Protection at Work.


these institutions or bodies,223 and this before 15 June.224 The establishing acts of the more recent agencies further provide that the report should be published or be made public.225

The Common Approach also devotes special attention to the annual report and explicitly links the annual report to the work programme. It also provides that the annual report should include feedback on the recommendations which the agency received following the discharge by the budgetary authority and the recommenda-tions following audits. Similarly to the work programme, the Common Approach prescribes that the annual report should consist of a number of common elements, so as to ease comparison. As regards procedure, the Common Approach provides that the Director should draft the report and submit it to the Board. The final report is then transmitted to the Commission, Council, Parliament, and Court of Auditors by either the Director or the Board, by 1 July. The Common Approach thus proposes a certain rationalization and uniformization and introduced a clear link between the work programme and the report. Again, however, the Common Approach is also only a partial codification, since no guidance is given on whether the report should also be forwarded to other actors (EU bodies, Member States) and whether it should be made public.

3.4 External relations

As regards the agencies’ external relations, it should be noted that virtually every establishing act recognizes the international context in which EU agencies operate and most establishing acts also task the agency concerned to be actively involved in the international sphere.

The types of external action which are foreseen do vary, however. As Ott et al point out, three main types of external action may be discerned. EU agencies can (i) be a forum open to third countries or even international organizations, (ii) assist

223 As was already noted, in the cases of the CPVO and OHIM it is the Director who adopts the report and therefore also forwards it. In the cases of the ECDC, ECHA, EFSA, and (old) ENISA the Director forwards the report even if the Board adopts it: see Articles 16(4) ECDC, 83(3) ECHA, 26(3) EFSA, 7(7) old ENISA. The 2013 recast ENISA Regulation has changed this to the default situation. In the case of Eurojust the College prepares the report and the President forwards it to the Council who in turn informs the Parliament. In the case of the ETF, the Board adopts and for-wards the report but the Director then presents it to the preparatory bodies of the Council and the Parliament’s committees.

224 15 June is the deadline for forwarding the report pursuant to Articles 13(12) ACER, 5(5) BEREC, 8(3) CdT, 12b(1) Cedefop, 33(2)b EASA, 29(1)c EASO, 16(4) ECDC, 8(6) EEA, 26(3) EFSA, 10(6)b EIGE, 65(10) EMA, 9(7) EMCDDA, 10(2)b EMSA, 43(5) ESAs, 13(4) ETF, 10(2) EU- OSHA, 13(1) Eurofound, 12(1)k eu- LISA, 12(2)b FRA, 20(2)b Frontex. The first of July is the deadline pursuant to Articles 5(3) new ENISA, 6(g) old GSA, and 6(h) new GSA. For other agencies there is no deadline for forwarding the report, only for adopting it: 31 March pursuant to Article 6(9) ENISA; 30 April pursuant to Articles 78(a) ECHA, 23(2)b EFCA, and 25(2)b ERA; 30 June pursuant to Article 15(b) EIT. The JHA agencies and OHIM and CPVO do not have a deadline for adopting or forwarding their report.

225 See Articles 13(12) ACER, 10(12) CEPOL, 32(1)c EASA, 29(1)c EASO, 83(3) ECHA, 23(2)b EFCA, 38(1)f EFSA, 10(6)b EIGE, 13(3) EIT, 7(7) old ENISA, 5(3) new ENISA, 43(5) ESAs, 12(1)k eu- LISA, 20(2)b Frontex.


the institutions in their external policy, and (iii) engage in relations with third countries or international organizations in their own right.226 In the last category a distinction can then be made between those agencies which are merely called upon to develop appropriate contacts with counterparts and those agencies whose mandates explicitly provide that they may enter into (administrative) agreements.

3.4.1 EU agencies open to third countriesAlthough it cannot be a priori excluded that third countries might participate in an EU agency in absence of an explicit provision to this end in the agency’s estab-lishing act, some such acts set out the basic requirements for third countries to be able to participate on a structural basis in an EU agency.227

The first such provision was introduced in the EEA Regulation in 1990, which read, in Article 19:

The Agency is open to countries which are not members of the European Communities but which share the concern of the Communities and the Member States for the objec-tives of the Agency under agreements concluded between them and the Community fol-lowing the procedure in Article 228 of the Treaty.

An identical provision was also included in the EMCDDA Regulation.228 In the ETF Regulation this provision was elaborated and it was added that the ‘agree-ments shall, inter alia, specify the nature and extent of and the detailed rules for the participation of these countries in the work of the Foundation, including provisions on financial contributions and staff’.229 In further establishing acts the following three elements may then be identified if a provision on the structural participation of third countries is foreseen: 230 (i) the EU needs to have concluded an agreement with the third country in question, (ii) the third country needs to apply relevant EU legislation in the field,231 and (iii) the agreement must lay down the specific rules for the third country’s participation in the agency.232 As a result, despite the EU agencies’ independence, it will be the political institutions that decide whether and which third countries may participate structurally in

226 Andrea Ott, Ellen Vos, and Florin Coman- Kund, ‘EU Agencies and Their International Mandate: A New Category of Global Actors?’, (2013) Cleer Working Papers 7, p 8.

227 It should be noted that as early as the 1991 IGC, the Commission proposed amendments to current Article 218 TFEU (then Article 228 EEC) inter alia to the effect that international agree-ments making ‘provision for participation by non- member countries or international organizations in bodies set up by Community law’ be concluded by the Council using QMV with the Parliament’s assent. See European Commission, SEC (91) 500, p 21.

228 See Article 21 EMCDDA. 229 See Article 23 ETF.230 See Articles 31 ACER, 66 EASA, 30 ECDC, 49 EFSA, 19 EIGE, 17 EMSA, 24 ENISA, 36

ERA, 75 ESAs.231 Rapoport notes that this is in fact a formidable requirement which de facto excludes participa-

tion in an agency for a lot of third countries. See Cécile Rapoport, Les partenariats entre l’Union européenne et les états tiers européens: étude de la contribution de l’Union européenne à la structuration juridique de l’espace européen, Bruxelles, Bruylant, 2011, pp 503– 7.

232 An important aspect here is the financial contribution which third countries are required to make in order to cover the costs of their participation. In more detail, see ibid, pp 543– 50.


an agency. This also holds for one exceptional case where a specific agreement is not required for a third country to participate. Article 28 of the FRA Regulation provides that countries which have concluded a stabilization and association agreement can participate in the FRA as observers, pursuant to a decision of the Association Council setting out the practical arrangements which are otherwise found in the specific bilateral agreements.233 Other establishing acts explicitly allow for ad hoc participation of relevant third countries or international organi-zations in the work of the agency.234

When a third country or international organization participates in an agency on a case- by- case basis or even under a structural participation, it does not have voting rights. The only exception to this is the participation of Iceland, Norway, Switzerland, and Liechtenstein in the EU’s Schengen acquis. In the agreement on Iceland’s and Norway’s participation in Frontex, for in-stance, voting rights for these countries are provided for but limited to specific matters relevant to these countries and to training activities.235 The same has been provided in the agreement with Switzerland and Liechtenstein.236 On the other hand, while Norway and Iceland are also involved in the Dublin system, the agreements detailing the arrangements for their participation in the EASO do not provide for any voting rights.237 In the establishing act of the eu- LISA the possibility for non- EU states to exercise voting rights on the Board of an EU agency was further recognized for the first time in an estab-lishing act itself.238

Norway, Iceland, and Liechtenstein are of course also a special category of third countries because of their membership of the EEA. Through this membership, they have the possibility to send observers to the Boards of many agencies.239 This is made possible through the decisions of the EEA Joint Committee amending the Annexes or Protocols to the EEA Agreement. The decisions typically provide that the European Free Trade Area (EFTA) States shall participate fully in the Board of the agency concerned and that they shall have the same rights and obli-gations within that Board as the EU Member States, apart from the right to vote. This allows the EEA EFTA states to participate in those agencies without having

233 The only Association Council adopting such a decision has been that established by the EU- Croatia SAA. Evidently, the decision is not in effect anymore. See Decision 1/ 2010 of the EU- Croatia Stabilisation and Association Council, OJ 2010 L 279/ 68.

234 Of course, the Board will generally be able to invite ‘other persons’ to its meeting whose par-ticipation might be interesting. For specific provisions, see Articles 106 and 107 ECHA, 77 EMA, 9 EU- OSHA and 23(2) ETF.

235 See Article 1(2) of the Arrangement between the European Community and the Republic of Iceland and the Kingdom of Norway, OJ 2007 L 188/ 17.

236 See Article 1(2) and (3) of Arrangement between the European Community and the Swiss Confederation and the Principality of Liechtenstein, OJ 2010 L 243/ 4.

237 See the Arrangement between the European Union and the Principality of Liechtenstein, OJ 2014 L 170/ 50. For that with Norway, see OJ 2014 L 109/ 3.

238 See Article 37 eu- LISA. So far, no agreement has been concluded with third countries grant-ing them voting rights in the eu- LISA Board.

239 See the list at http:// www.efta.int/ eea/ eu- agencies. On this issue in further detail, see Rapoport, n 231, pp 508– 11.

http://www.efta.int/eea/eu-agencies

http://www.efta.int/eea/eu-agencies


to conclude an agreement to this end with the EU. Despite this advantage over normal third countries, the Commission has still complained about the ‘lengthy negotiations between the EU and the EEA EFTA states […] before incorporating new legal acts, whenever the status or level of participation of these states in EU agencies needs to be addressed’.240 As an alternative, the Commission has sug-gested a horizontal agreement specifically aimed at the EEA EFTA states’ partici-pation in EU agencies.241

3.4.1.1 EU agencies in pre- accession and the ENPThe possibility for third countries to participate fully, with the exception of exer-cising voting rights, in EU agencies is an interesting way to familiarize candidate countries with the day- to- day functioning of the EU.242 The Europe Agreements (EAs) concluded with the Central and Eastern European Countries (CEECs) gen-erally invited the latter to cooperate with the EU in a number of fields, but only a participation in the EEA was explicitly foreseen.243 In its Agenda 2000 the Commission emphasized the potential benefits of participation in EU agencies,244 with the Luxembourg 1997 European Council concluding that the EU agencies open to participation should be determined on a case- by- case basis.245 In 1999, then, the Commission identified the EEA and the EMCDDA as the agencies for which participation was possible in the near term.246

As was noted previously, the agreements need to be concluded ad hoc for each agency and third country separately. This is different from the participation of (potential) candidate countries and European Neighbourhood Policy (ENP) Partners in EU programmes, which is enabled generally through the conclusion of a Framework Agreement with the country in question.247 Ultimately, the only agency for which a series of agreements with the different CEECs was concluded was the EEA. The only candidate countries with which an agreement was con-cluded on their participation in the EMCDDA were Croatia and Turkey.248 Apart from the requirement that a country applies the relevant EU acquis, which for

240 European Commission, SWD (2012) 425 final, p 10. 241 Ibid.242 See Marc Maresceau, ‘Pre- accession’, in Cremona (ed), The Enlargement of the European

Union, Oxford, OUP, 2003, p 24.243 The EAs with Hungary and Poland even foreseeing this before the EEA had started function-

ing: see eg Article 80(3) of the EA with Poland.244 See inter alia European Commission, Agenda 2000, Bulletin of the European Union,

Supplement 5/ 97, p 54.245 See Presidency Conclusion of the Luxembourg European Council, 12 and 13 December

1997, para 21.246 European Commission, COM (99) 710 final.247 Those framework agreements then stipulate that the Commission will conclude separate

Memoranda of Understanding with the third countries for each programme in which the country will participate.

248 See Agreement between the European Community and the Republic of Turkey, OJ 2007 L 323/ 24; Agreement between the European Union and the Republic of Croatia, OJ 2011 L 334/ 7. The Commission had negotiated an agreement with Bulgaria and Romania but these countries acceded to the EU before the Council could formally conclude the agreement. See Withdrawal of obsolete Commission proposals, OJ 2010 C 252/ 7. Further, an agreement with Norway has also been concluded for that country’s participation in the EMCDDA.


certain agencies is a considerable feat,249 the main reason for the lack of further agreements on structural participation is the rather cumbersome and lengthy pro-cedure.250 Croatia for instance had formally expressed its interest in participating in the EMCDDA in 2005, but negotiations only started in 2006 and contin-ued for three years,251 the agreement itself being finally concluded in December 2011. As a result, Croatia only participated for a good eighteen months in the EMCDDA before it became a Member State proper.

The lack of agreements on third country participation in EU agencies may then raise the impression that this instrument is not used to its full potential in the pre- accession strategy. However, because of the possibility of ad hoc participation in EU agencies, third countries may already be familiarized with the EU agencies.252 As the Commission further explained, participation takes different forms,253 and the conclusion of an agreement for participation in an EU agency is only the final step.254 Part of the funding under the Instrument for pre- Accession Assistance may then be used by a third country to cover the costs of participating in an EU agency.255 A similar arrangement may be found for the ENP countries in the European Neighbourhood and Partnership Instrument (ENPI).256 However, the possibility of ENP countries participating in EU agencies has remained a theoret-ical possibility, even if the Commission in 2006 stressed the importance of such a participation by adopting a general approach for such participation.257

Although the issue of third country participation in EU agencies is of a prac-tical nature itself, it touches upon a more fundamental issue, viz the agencies’ independence. The third countries’ participation shows that EU agencies, despite being autonomous bodies, actually work under the guidance of the institutions. This is very clear in the EU’s relations under the pre- accession process and the ENP. The pace and the extent of the development of these relations are necessarily determined by the political institutions.

3.4.2 Assistance functionWhere the provisions on the structural participation of third countries in EU agencies seem to have taken a definitive form, this is different for the provisions

249 See Constance Chevallier- Govers, ‘La participation des états tiers aux agences et aux programmes de l’union européenne’, in Bosse- Platière and Rapoport (eds), L’État tiers en droit de l’Union européenne, Bruxelles, Bruylant, 2014, p 199; supra n 231.

250 Because of this, the Commission in May 2003 ruled out the possibility of concluding negoti-ated agreements with the CEECs that would join in 2004. See European Commission, 13 May 2003, ‘Guide on Participation of candidate countries in Community programmes, agencies and committees over 2003 and 2004’, p 6.

251 European Commission, COM (2010) 21 final.252 Further on this participation simplifiée, see Rapoport, n 231, pp 537– 43.253 European Commission, COM (99) 710 final, p 9.254 European Commission, COM (2003) 748 final, p 10.255 See Article 71 of Regulation (EC) 718/ 2007 of the European, OJ 2007 L 170/ 1.256 See Article 2(2)x Regulation (EC) 1638/ 2006 of the European Parliament and of the

Council, OJ 2006 L 310/ 1.257 European Commission, COM (2006) 724 final, pp 4– 7.


prescribing that the agencies may assist the institutions when they carry out the EU’s external policies. The rationale behind this assistance function is clear. While the EU is represented in international forums by the Commission, the latter usu-ally lacks the expertise to participate in these forums on a technical level. In order for the EU to participate on that level as well, the EU agency may be involved in the EU delegation. Still, an explicit provision to this end is not required in the agency’s mandate, as it may normally be subsumed under most agencies’ general task of providing the Commission with expertise and information.

Where a provision is dedicated to this assistance in the external sphere, the wording in some establishing acts is remarkable, since it is not provided that the agency will assist the Commission (as the EU’s representative) but rather that it will assist the Community or the Union.258 From an international perspective this appears logical, since it is the EU and not the EU Commission which participates in international forums. However, by providing in an EU regulation that an EU agency will assist the EU, the impression is given that the EU is an international legal subject distinct from the EU agency. This in turn might imply that the EU agency may also be an international legal subject distinct from the EU. Put differently, this raises the question of how far the legal personality of EU agen-cies reaches. However, in the EMSA Regulation, it is provided that the agency will facilitate the cooperation between the Member States and the Commission (rather than the EU) by providing the necessary assistance for the Member States and the Commission to contribute to the work of inter alia the International Maritime Organisation.259 Apart from this issue, the agencies’ task of assisting the Commission in the external sphere is and should be uncontroversial: it forms part of the agencies’ general assistance or information task and thus forms part of their very raison d’ être.

3.4.3 EU agencies as international actorsApart from being open to third countries and assisting the EU institutions, EU agencies may also actively reach out in the international sphere, developing con-tacts with (authorities in) third countries or international organizations. As noted previously, under this type of external action agencies are mostly called upon to develop appropriate contacts with relevant (international) counterparts.260 Since EU agencies generally do not operate in fields where the EU is isolated from an international context, such contacts are necessary and cannot be controversial. However, a number of establishing acts also (or further) provide that such inter-national cooperation may or should be formalized in the form of working ar-rangements or agreements. This raises certain questions on the procedure and the nature of these agreements, as well as the nature of EU agencies in the international

258 See Articles 27(1) EASA, 4 EFCA, 3(j) ENISA. 259 See Article 2(4)(h) EMSA.260 See inter alia Articles 11 ECDC, 15(2) and (3) EEA, 33 EFSA, 8(1) EIGE, 28c(1) and 52

EMA, 8 EUROFOUND, 26a(1) Eurojust, 23(1) Europol, 8(2) FRA, 9(2) Frontex and the 12th recital of the preamble to the EU- OSHA regulation.


sphere. The only EU agency which has been empowered in the past to conclude genuine international agreements is Europol, by virtue of being originally estab-lished as an international organization and thus having clear international legal personality.261

A lot of establishing acts provide that the agency concerned must conclude administrative agreements with other relevant bodies.262 However, for the ESAs this is just a possibility.263 As regards procedure, the EFCA Regulation provides that the Board needs to give its approval and that the Director will keep the Commission and Member States informed on any negotiations.264 In the case of the EASA the Regulation provides that the Commission must give its prior ap-proval before such agreements may be concluded.265 In the cases of the FRA and EMCDDA the Board can only conclude the agreement when the Commission has delivered an opinion. That opinion is not binding but, should it be negative, the Board can only conclude the agreement with a special majority.266 For the EASO and Frontex, the procedure is (completely) left to the agency, but it is also explicitly provided that the agencies’ external relations need to be in accordance with the TFEU.267 For Eurojust and Europol, it is the Council which has to ap-prove the conclusion of agreements.268 For the EIGE the situation is different still. Although its establishing act allows it to develop international contacts, it also provides that the EU (and not the EIGE itself) will enter into agreements with international organizations or third countries in the interest of the EIGE.269 Lastly, in the ESA Regulations it is provided that the agencies’ administrative ‘arrangements shall not create legal obligations in respect of the Union and its Member States nor shall they prevent Member States and their competent au-thorities from concluding bilateral or multilateral arrangements with those third countries’.270

As a result, there is quite some variety in the provisions on these procedural aspects. Still, a recurring theme seems to be the timorousness of allowing the EU agencies much autonomy in their external relations. From this perspective, the provisions in the FRA and EMCDDA Regulations are remarkable since, despite the Commission opinion which is required when they conclude agreements, they explicitly acknowledge that these two agencies might conclude agreements against the Commission’s opinion, as long as a supermajority may be found in the Board.

The paragraphs in the Common Approach dedicated to the agencies’ inter-national relations confirm this. It is thus provided that the agencies’ external

261 See Andrea Ott, ‘EU Regulatory Agencies in EU External Relations’, (2008) 13 European Foreign Affairs Review 4, p 10; Ott, Vos and Coman- Kund, n 226, p 27.

262 See Article 17g EFCA. A similar provision may be found in Articles 27(2) EASA, 52 EASO, 11(2) ECDC, 20 EMCDDA, 8 FRA.

263 See Articles 33(1) ESAs. Similarly, see Articles 26a(2) Eurojust, 14(2) Frontex.264 See Article 17g EFCA. 265 See Article 27(2) EASA.266 This is a three- fourths majority in the case of the EMCDDA and a two- thirds majority for

the FRA. See Articles 20 EMCDDA, 8(3) FRA.267 See Articles 52 EASO, 14(2) Frontex.268 See Articles 26a(2) Eurojust, 23(2) Europol. 269 See Article 8(2) EIGE.270 See Articles 33 ESAs.


policy should be embedded in their (multi- )annual work programmes. Through the procedures which apply to the work programmes (cf. section II 3.3), the institutions are therefore kept informed. In addition, the Common Approach provides that appropriate working arrangements between an agency and its parent DG should be worked out and that the external policy and specific initiatives should be approved by the agency’s Board. Lastly, the Common Approach stresses that an early exchange of information in this area between the agencies, the Commission, and the EU delegations should be ensured. The combined effect of these provisions is clear: the EU institutions will never be presented with a fait accompli or see the EU’s external action being pre- empted by the external action of an EU agency. To this end the Common Approach explicitly provides that it should be ensured that EU agencies ‘are not seen as representing the EU position to an outside audience or as committing the EU to international obligations’ and that the consistency of EU policy needs to be assured.

Again, however, the lack of detail in the Common Approach should be noted, since it does not attempt to come to a harmonization of the procedure applicable when agencies conclude agreements. In addition, the provisions on agencies’ ex-ternal relations are affected by other provisions in the Common Approach since they refer back to the work programme and decisions of the agencies’ Boards, without however harmonizing these elements (cf. sections II 3.1.6 and II 3.3).

3.5 Seat Agreement

From the purposes of a host agreement as identified by Muller,271 it may be questioned whether the EU agencies should also conclude Seat Agreements when there is already a Protocol annexed to the TEU, TFEU and Euratom Treaty dealing with the Privileges and Immunities of the European Union.272 According to Schusterschitz this is indeed the case, since ‘for a large number of independent community bodies […] the general Protocol and the country- specific modalities [as agreed between the Member State concerned and the Commission], were not sufficient for determining their exact status in their seat state’.273 As a preliminary observation it should be noted that it is not clear from the Protocol itself that it applies to agencies.274 As a result, all the establishing

271 Alexander Muller, International Organizations and Their Host States: Aspects of Their Legal Relationship, The Hague, Kluwer Law International, 1995, p 21.

272 See Protocol No. 7, OJ 2010 C 83/ 266.273 Gregor Schusterschitz, ‘European Agencies as Subjects of International Law’, (2004) 1

International Organizations Law Review 1, p 172.274 Thus, Articles 1 to 4 of the Protocol refer to privileges and immunities granted to the

Union. This raises the question whether because of their separate legal personality, these privi-leges and immunities are not extended to the agencies. Chapter V of the Protocol on the other hand deals with the Officials and other servants of the EU. Given that the Staff Regulations also apply to the agencies’ staff, there is no doubt that this part of the Protocol also governs the privi-leges and immunities of agencies’ staff. However, Article 5 of the Protocol refers to the commu-nications of the EU’s institutions rather than those of all the EU’s bodies. In addition Article 10


acts of all the EU agencies, except Eurojust, explicitly provide the Protocol ex-tends to the agency.275

Remarkably, in the light of Schusterschitz’ observation, the establishing acts of most agencies do not mention the possibility, let alone the obligation, to conclude Seat Agreements.276 Recently institutional practice changed on this point, since the agencies of the fourth wave do contain provisions to this end and, when the Frontex Regulation was revised in 2011, such a provision was also included. These Regulations therefore emphasize the duty of the Host State to enter into such a legal commitment.277 Only some of the other agencies have succeeded in concluding a seat agreement with their host state. This is the case for the Eurofound,278 Cedefop,279 EEA,280 ETF,281 EFSA,282

refers to the customary privileges and immunities of the representatives of the Member States, participating in the work of the institutions and the advisory bodies of the EU. Finally Articles 21 and 22 (which date from before the Lisbon Treaty) provide that the Protocol also applies to the ECB and the EIB, which were not EU institutions proper before Lisbon, the EIB still not being one. When the EMCF was established Ehlerman and Louis disagreed whether the Protocol could automatically apply to it. See Claus- Dieter Ehlermann, ‘Die Errichtung des Europäischen Fonds für Währungspolitische Zusammenarbeit’, (1973) 8 Europarecht 3, pp 204– 5; Jean- Victor Louis, ‘Le Fonds Européen de Cooperation Monetaire’, (1973) 9 CDE 3, p 281.

275 Another exception are the CFSP agencies. The establishing acts of the EUISS and the EUSC are silent on whether the Protocol applies to them. Instead, the Member States’ representatives adopted a decision on the matter. This decision takes the form of an international agreement. See Decision of the Representatives of the Governments of the Member States of the European Union meeting within the Council on the privileges and immunities granted to the European Union Institute for Security Studies and the European Union Satellite Centre, and to their bodies and staff members, Council document 11639/ 01. For the EDA the matter is slightly more complex since Article 25 of the establishing decision provides both the Protocol and a separate decision, again an international agreement lay down the privileges and immunities of both the agency and its staff. See Decision of the representatives of the Governments of the Member States meeting within the Council on the privileges and immunities granted to the European Defence Agency and to its staff members, Council document 11502/ 1/ 04.

276 The exception being the CEPOL decision which as early as 2005 contained a provision man-dating the Director to negotiate a Headquarters Agreement.

277 See Articles 53 EASO, 74 ESAs, 22 eu- LISA, and 15a Frontex.278 According to the interinstitutional working group on agencies, Eurofound has a Seat

Agreement: see Inter- institutional Working Group on regulatory agencies, Agencies’ Seat and Role of the Host Country, 2010, Analytical Fiche Nr° 3, p 1. However, a proper Seat Agreement has not (yet) been concluded between Eurofound and Ireland and the legal framework is instead governed by the Protocol and a letter from the Irish Ministry of Foreign Affairs to the Commission on the application of the protocol. See T.P. Corcoran Chief of Protocol of the Department of Foreign Affairs, Exchange of Letters concerning the Protocol on the Privileges and Immunities of the European Communities, 19 October 1981, (on file with the author).

279 Seat Agreement between the Government of the Hellenic Republic and the European Centre for the Development of Vocational Training, Official Journal of the Hellenic Republic, 12 September 1995, Issue A, Nr 190, 5529– 31.

280 Headquarters Agreement between the European Environment Agency and the Government of Denmark, I- 32150, UNTS, Vol. 1889, 311.

281 Accordo relativo alla sede tra la Fondazione Europea per la Formazione Professionale e il Governo della Repubblica Italiana, Supplemento ordinario alla Gazetta Ufficiale, n. 100, 2 maggio 1997, 7– 16.

282 Accordo di Sede tra La Republica Italiana e l’Autorità Europea per la Sicurezza Alimentare, Gazetta Ufficiale delle Republica Italiana, n. 21, 26 gennaio 2006, 5– 19.


EASA,283 ECDC,284 EFCA,285 ENISA,286 Eurojust,287 Europol,288 CEPOL,289 EMCDDA,290 EMSA,291 ECHA,292 ACER,293 EASO,294 and OHIM.295

Other agencies do not have Seat Agreements and this has been criticized in the past by both the European Parliament296 and the Court of Auditors,297 although

283 Just like in the case of Eurofound, the interinstitutional working group found the EASA had a seat agreement, but the legal framework is in fact governed by the Protocol and a letter from the German government to the executive director of EASA. See Bundesministerium für Verkehr Bau- und Wohnungswesen, Ansiedlung der Europäischen Agentur für Flugsicherheit (EASA) in Köln, 2 November 2004, (on file with the author).

284 Värdlandsavtal för Europeiska centrumet för förebyggande och kontroll av sjukdomar/ Seat agreement between the European Centre for Disease Prevention and Control and the Government of Sweden, January 2011, (on file with the author).

285 For the Memorandum of Understanding between Spain and the Community Fisheries Control Agency on the Establishment of its Headquarters on Vigo and the Seat Agreement between the Kingdom of Spain and the Community Fisheries Control Agency, see Annex III of European Fisheries Control Agency, Annual Report 2008, 2009, pp 85– 102.

286 Seat Agreement between the Government of the Hellenic Republic and the European Network and Information Security Agency, Official Journal of the Hellenic Republic, 8 June 2007, Issue A, Nr. 125, 2874– 87.

287 Exchange of notes between the Kingdom of the Netherlands and the College of Eurojust constituting an interim agreement regarding the status of Eurojust in the Netherlands, I- 41082, UNTS, Vol. 2304, 293; Exchange of notes constituting an agreement between the Kingdom of the Netherlands and Eurojust concerning the privileges and immunities of the staff of Eurojust and their family members, I- 43326, UNTS, Vol. 2400, 99.

288 Exchange of notes constituting an agreement between the Kingdom of the Netherlands and the European Police Office in the light of the Netherlands Government policy framework on at-tracting and hosting international organisations, I- 44967, UNTS, Vol. 2517, 273; Agreement be-tween the Kingdom of the Netherlands and the European Police Office (Europol) concerning the headquarters of Europol, I- 41633, UNTS, Vol. 2323, 439.

289 Headquarters Agreement between the Government of the United Kingdom of Great Britain and Northern Ireland and the European Police College, I- 41636, UNTS, Vol. 2324, 3.

290 Memorandum of Understanding between the Portuguese Government, the European Maritime Safety Agency and the European Monitoring Centre for Drugs and Drug Addiction on the Establishment of the Headquarters of these two European Entities in Lisbon, Diário da República I Série- A, No 224— 22de Setembro de 2004, 6081.

291 Ibid; Protocol between the government of the Portuguese Republic and the European mari-time safety agency, Diário da República I Série- A, No 224— 22de Setembro de 2004, 6073.

292 Seat Agreement between the Government of the Republic of Finland and the European Chemicals Agency, Suomen Säädöskokoelman Sopimussarja, 25 January 2008 N:o 11, 90.

293 Seat Agreement between the Government of the Republic of Slovenia and the Agency for the Cooperation of Energy Regulators, Uradni list Republike Slovenije, Št. 22/ 30. 12. 2010, 1999.

294 Seat Agreement between the Government of Malta and the European Asylum Support Office, available at: http:// ec.europa.eu/ dgs/ home- affairs/ what- we- do/ policies/ pdf/ 20110418_ easo_ seat_ agreement_ en_ and_ mt_ en.pdf.

295 Acuerdo de Sede entre El Reino de España y la Unión Europea (Oficina de Armonización del Mercado Interior— Marcas, Dibujos y Modelos— OAMI), Boletín Oficial Del Estado, Núm. 254 Viernes 21 de octubre de 2011 Sec. I., 110030.

296 See for instance the Parliament’s resolution of 2010 on the ECDC budget, stressing this issue. In its 2012 resolution Parliament welcomed the signing of a seat agreement after five years of nego-tiation. See Resolution of 5 May 2010 of the European Parliament, OJ 2010 L 152/ 142; Resolution of 10 May 2012 of the European Parliament, OJ 2012 L 286/ 184.

297 See points 16– 17 of Report of the Court of Auditors, OJ 2009 C 304/ 17. In its 2011 Report on EU- OSHA, the Court noted that the agency did not have a seat agreement, unlike the other two EU agencies established in Spain. The EU- OSHA replied that it would finalize negotiations with the Kingdom of Spain in 2012. See Report of the Court of Auditors, OJ 2012 C 388/ 219.

http://ec.europa.eu/dgs/home-affairs/what-we-do/policies/pdf/20110418_easo_seat_agreement_en_and_mt_en.pdf





not in a general and consistent way.298 On the other hand, even if Seat Agreements have been concluded, this does not mean that all issues are settled. In its general resolution on the 2008 discharge for the EU agencies, the Parliament for instance noted that when Seat Agreements are in place they are often unsatisfactory, and called upon the interinstitutional working group to develop common standards for the Seat Agreements.299

The section on the agencies’ seat and the role of the host country is one of the longest in the Common Approach, taking up a full page of the twelve- page docu-ment. The Common Approach explicitly states that its provisions are without prejudice

to the political decision on an agency’s seat taken by common agreement between the rep-resentatives of the Member States meeting at Head of State or government level or by the Council, to the desirability of geographical spread, and to the objective set in December 2003 by the representatives of the Member States.300

The objective referred to is the commitment to give priority to the new Member States in the distribution of the seats of new agencies.301

The difficulties of coming to an agreement on the location of the agencies during the European Council summit of Laeken302 will have inspired the draft-ers of the Common Approach to prescribe that the decision on an agency’s seat has to be taken before the end of the legislative process. Before, these decisions were sometimes postponed and the agency was provisionally located in Brussels, which hampered its functioning since it could not provide certainty to its (pot-ential) functionaries on their actual place of work.303 The Common Approach partially codifies existing practices and includes a general obligation to conclude such agreements before a (new) agency starts its operational phase; it extends this obligation to the existing agencies without such agreements (even if no such pro-vision was included in the new GSA regulation). The Common Approach also stresses the accessibility of the agency’s location.

That criterion is not just important for attracting staff but also affects the efficiency of the agency when it fulfils its network function. Bringing together experts and organizing workshops or conferences is not straightforward if the agency is located outside a city and/ or distant from important transport links

298 Rapporteurs Haug and Elles, citing the problems Frontex has in negotiating a Seat Agreement, did note in 2006 that ‘the governments of those Member States who have an agency seat, must make a contribution in the form of buildings, building sites or infrastructure […] If Member States insist on having “their” agencies they should also assume their financial responsibilities.’ See point 33 of Rapporteurs Jutta D. Haug and James Elles, Working Document on a Meeting with the Decentralised Agencies on the PDB for 2007, 29 June 2006, Vol. PE 367.332v02- 00.

299 See point 6 of Resolution of the European Parliament, OJ 2010 L 252/ 241.300 Common Approach, point 6.301 Council of the European Union, Doc 5381/ 04, p 27.302 See Ellen Vos, ‘Agencies and the European Union’, in Zwart and Verhey (eds), Agencies

in European and Comparative Perspective, Antwerpen, Intersentia, 2003, pp 126– 7 at footnote 61.303 Groenleer, n 195, p 107. See also the delay in deciding on a seat for the EEA, Chapter

III, n 152.


such as international airports.304 This is a problem which a number of existing agencies face, such as the EFSA and ENISA; most exemplary is the ERA, which has a dual location (Lille and Valenciennes) because its official location is rather remote. 305

3.6 The Common Approach’s contribution to an agency model

From the observations on the Common Approach’s provisions on each of the sections discussed previously, it may already be clear that the Approach has not introduced an agency model.

The Common Approach is non- binding and therefore only contains soft commitments on the part of the institutions. In addition, a lot of its provi-sions further explicitly provide that exceptions may be made in justified cases or simply at the legislator’s discretion. Lastly, most of the suggestions contained in the Approach are (only) a (partial) codification of existing practices.306 If the Common Approach is seen as the culmination of the institutions’ attempts, since the Parliament’s resolution of 1993,307 to come to a rationalization of agencifica-tion, it should be clear that it is a disappointment. Still it may be hoped that the Common Approach will help the Commission to draft proposals and will make it more difficult for the legislator to deviate from the now (partially) codified practice.

Scholten noted other shortcomings,308 such as the Common Approach only applying to future agencies, the issue of the democratic accountability of self- financed agencies not being addressed, and lastly the agencies’ position within the institutional architecture not being clarified. Of course, the Common Approach could not have settled the last issue even if it had been binding, since it can only be settled through a Treaty revision (cf. section VI 2).

If most elements of the Common Approach were a codification, then what were the genuinely innovative elements? Some of them have already been mentioned, leaving only one major novelty to comment upon.

304 See Euréval and Rambøll- Management, n 75, p 60.305 That agencies’ seats are regarded as political trophies for domestic politics was exempli-

fied when the original draft opinion of the Parliamentary Committee on Transport and Tourism on the budget execution for the year 2008 was prepared. The draft contained a paragraph criti-cizing the dual location and the extra costs this entails, but this paragraph was deleted following an amendment by the then mayor of Valenciennes (acting as MEP). See Amendments 1– 2 to the Draft Opinion on the discharge in respect of the budget of the European Railway Agency for the financial year 2008, TRAN_ AM(2010)438470.

306 These three issues have also been noted by Comte: see Françoise Comte, ‘Agences décen-tralisées: vers un statut unifié? Approche commune du Parlement européen, du Conseil de l’Union européenne et de la Commission européenne sur les agences décentralisées’, in Govaere and Hanf (eds), Scrutinizing Internal and External Dimensions of European Law— Les dimensions internes et externes du droit européen à l’ épreuve, Bruxelles, Peter Lang, 2013, Vol. I, pp 147– 8 and 155.

307 See n 56.308 Miroslava Scholten, ‘The Newly Released “Common Approach” on EU Agencies: Going

Forward or Standing Still?’, (2012) CJEUL Online, http:// www.cjel.net/ online/ 19_ 2- scholten, pp 3– 4.

http://www.cjel.net/online/19_2-scholten

http://www.cjel.net/online/19_2-scholten


3.6.1 Abolishing or merging agenciesThe Common Approach provided that acts establishing agencies should contain review or even sunset clauses. Whereas it is rather common to find periodical review provisions in EU legislative acts, sunset clauses for EU agencies do not exist since EU agencies are permanent bodies. Yet the Common Approach provided that closing down an agency or merging it with another agency should be real options.

As already noted, the ENISA was established with a sunset clause, but instead of closing down the agency following the period provided for, the legislator has prolonged the validity of the original ENISA Regulation thrice.309 In an attempt to solve this issue differently, the Commission proposed to merge the ENISA with the to- be- established ECMA, also in order to solve the problems in the ENISA’s functioning.310 The Commission’s proposal met with obvious resistance,311 and the Member States in Council quickly agreed not to merge the ENISA.312 Just like the Council, the Parliament was not in favour of abolishing the ENISA.313 In the end the ECMA was not established as an agency,314 and the ENISA was not abolished.

The second serious initiative by the Commission was in 2013 when it suggested, in its proposal to strengthen Europol and align it to the new legal reality under the Treaty of Lisbon, that Europol could also take over the activities of CEPOL, thereby abolishing the latter.315 Since its establishment, CEPOL had indeed under-performed and its small size meant that it never acquired the critical mass necessary to operate effectively as an EU agency. In addition, CEPOL was housed on the same estate as the UK National Policing Improvement Agency in Bramshill, but the UK government decided to close down the Bramshill site,316 resulting in loca-tion problems for CEPOL. As will be noted (see section II 4.1), the Parliament and Council in the end resisted a merger and kept CEPOL as a self- standing agency. Whether a rationalization of agencification through a merger between agencies will ever succeed is thus doubtful if even an undersized, malfunctioning agency such as CEPOL proves to be unabolishable.

4 Practice Following the Common Approach

The proof of the pudding is in the eating, and this is no different for the Common Approach, since its effects will depend on how the institutions follow up on their

309 See Chapter I, n 58.310 European Commission, COM (2007) 699 final, p 3. Some of these identified weaknesses

were ENISA’s remote location, affecting recruiting and networking; high turnover of staff; lack of critical mass of operational staff; etc. See European Commission, COM (2007) 285 final.

311 See Parliamentary Question (E- 841/ 08) of MEP Georgios Georgiou on the dissolution of the European Network and Information Security Agency (ENISA), 21 February 2008.

312 See Rapid Press Release, 12 June 2008, ‘2877th Council meeting Transport, Telecommunications and Energy Luxembourg, 12- 13 June 2008’, 10410/ 08 (Presse 165).

313 See Legislative resolution of the European Parliament, OJ 2010 C 8E/ 337.314 See Chapter III, at n 30.315 See European Commission, COM (2013) 173 final.316 See Report of the Court of Auditors, OJ 2012 C 388/ 23.

Practice Following the Common Approach 99

commitments and in how far the schedule of the Commission’s roadmap imple-menting the Common Approach is respected.317

At the time of finalizing this study a number of proposals had been adopted by the Commission, with the legislative negotiations pending. As a result, the proposals will only be briefly commented upon.

4.1 Review of agencies’ acts following the Common Approach

The Commission’s proposals for the fourth railway package evidently included a re-vision of the ERA Regulation.318 The influence of the Common Approach can be seen, since the Administrative Board would be renamed the Management Board, an executive Board would be created, a multiannual work programme would have to be adopted, etc. Especially relevant for the ERA would be the provision on the Headquarters Agreement. Unlike the Common Approach, the proposal suggested a clear time limit to conclude the agreement and also emphasized proper operating conditions ‘including multilingual, European- oriented schooling and appropriate transport connections’. Remarkably, the proposal no longer requires the ERA Director to have administrative skills or knowledge of the railway sector. On the other hand, the appointment procedure for the Director has been elaborated from the original regulation and goes beyond the provisions of the Common Approach.

In March 2013 the Commission adopted a proposal to modernize the OHIM Regulation, inter alia providing for more efficient procedures and a manda-tory cooperation between national trade mark offices and the OHIM.319 The Commission’s proposal confirms a lot of what has been noted previously. While some elements are clearly inspired by the Common Approach (eg the multiannual programme; renaming of the Board, the President, and the agency itself), a lot of the proposed changes also clearly go beyond the principles set out in the Common Approach (eg the appointment and removal procedures for the Director). Again this casts doubt on the real added value of the Common Approach. Interestingly, the proposal does not foresee changes to the budgetary procedure, and neither does it contain a clause on a Headquarters Agreement. Although OHIM already has such an agreement, it would still be advisable to include this obligation in the regulation. After all, it is perfectly imaginable that the existing agreement might require updating in the future, while there would not be a clear obligation for Spain to renegotiate the existing agreement.320 Lastly, it may be noted that the

317 See European Commission, 19 December 2012, ‘Roadmap on the follow- up to the Common Approach on EU decentralised agencies’. See also European Commission, ‘Progress report on the implementation of the Common Approach on EU decentralised Agencies’, COM (2015) 179 final.

318 European Commission, COM (2013) 27 final.319 European Commission, COM (2013) 161 final. Following the finalization of the present

manuscript, the proposal was adopted as Regulation (EU) 2015/ 2424 of the European Parliament and of the Council, OJ 2015 L 341/ 21.

320 The existing agreement, cf. n 295, does not include a provision on the review of the agree-ment either.


Commission further proposed to change the voting rules in the Management Board in line with the Common Approach to also do away with the anomaly that its representatives do not have voting rights.321

In March 2013 the Commission submitted a key proposal on Europol and CEPOL,322 two agencies with origins under the third pillar which the Lisbon Treaty has greatly affected.323 The proposal was also of critical importance be-cause the Commission suggested merging both agencies (as previously discussed). Ultimately this was rejected by both the Parliament and Council,324 and CEPOL was moved from the UK to Budapest.325 The Commission also proposed to align the Europol regulation with the other principles set out in the Common Approach, thereby significantly improving the Parliament’s position. At the same time the Commission proposed to do away with Europol’s special external competences, as it would no longer be empowered to conclude international agreements.326

In May 2013 the Commission proposed an overhaul of the EU legislation on the production and commercialization of plant reproductive material.327 Only a small part of the Commission’s proposal dealt with the CPVO, the Commission proposing to change the agency’s name to European Agency on Plant Varieties (EAPV) as suggested by the Common Approach. Remarkably enough the Commission did not take up this opportunity to propose other changes in line with the Common Approach.

5 Conclusion

The brief overview of the provisions on the agencies’ governing bodies and func-tioning has shown there is no agency model in the EU, even if there are three main areas which have largely been harmonized prior to the adoption of the Common Approach. These are the rules on the agencies’ budgets, access to documents, and the processing of personal data by EU bodies (including the agencies). Rules on other issues still form the subject of a learning exercise on the part of the three main institutions, whereby the establishing acts are drafted in an increasingly more detailed fashion.

However, even if more attention is dedicated to these issues, there does not seem to be a general understanding on EU agencies which forms the inspiration of these (more detailed) provisions. For instance, the novelty introduced in the

321 See n 131.322 European Commission, COM (2013) 173 final.323 On this in general see Steve Peers, ‘Mission Accomplished?: EU Justice and Home Affairs

Law after the Treaty of Lisbon’, (2011) 48 CMLRev 3, pp 661– 93.324 European Commission, COM (2014) 465 final, p 1.325 Regulation 543/ 2014 of the European Parliament and of the Council, OJ 2014 L 163/ 5.326 European Commission, SWD (2013) 98 final, p 3.327 European Commission, COM (2013) 262 final. This proposal was revoked in line with the

Commission’s 2015 Work Programme; see OJ 2015 C 80/ 17.

Conclusion 101

ESAs’ Regulations providing greater power to the Parliament seems inspired by securing greater democratic accountability of these agencies, but the fundamental question of the Parliament’s relations with subsidiary organs such as the EU agen-cies is not addressed. Unfortunately, the Common Approach does not change much in this, as it is not the fruit of a genuine common understanding shared between the institutions.

Since the Common Approach is not an act of primary law, it obviously could not settle all these questions; but it has not provided new impulses to this debate either, since it is merely a partial codification. In addition it is a non- compulsory declaration of intent, in line with the wishes of the Council. This would have been different if the institutions had succeeded in adopting a binding interinstitutional agreement, even if the draft agreement suffered from some of the same deficien-cies as the Common Approach.328

The institutional practice preceding and following the Common Approach is illustrative of the above. Some legislative revisions post- Common Approach for which the proposals were made before 2012 already go beyond the provisions of the Common Approach.

328 For instance, by starting from the already obsolete idea that agencies fulfil purely tech-nical tasks without a margin of discretion. See Merijn Chamon, ‘EU Agencies: Does the Meroni Doctrine Make Sense?’, (2010) 17 MJECL 3, pp 303– 4.

IIIThe Political Limits to Agencification

In the present chapter, the political limits to agencification will be explored. As has already been noted, agencification seems foremost driven by the (political) interests of the competent actors, rather than a concern for ensuring the legal soundness of this institutional development under current primary law. To better understand this political driving force, agencification’s context is first elaborated. This will be done (i) by looking at the forerunners of, and often alternatives to, EU agencies (section III 1); (ii) by presenting the dynamics driving agencification (section III 2); and (iii) by looking at the effects of agencification (section III 3).

1 The Origins of Agencies

To understand why EU agencies are established, it is interesting to first look at their origins. As Kreher notes, many EU agencies have institutional or procedural forerunners at EU level.1 At the same time these precursors are, to an extent, also alternatives to agency creation.

The European Medicines Agency (EMA), for instance, was established as a permanent structure around already long established comitology committees.2 Other agencies may be identified as ‘upgraded’ committees as well, although not necessarily upgraded comitology committees. The European Supervisory Authorities (ESAs) for instance find their origin in the level three committees of the Lamfalussy process for financial regulation,3 which involves four levels. On the first level legislation is adopted, which is then implemented through comitol-ogy on the second level. The coordination among national regulators is situated at the third level and monitoring of national implementation by the Commission

1 Alexander Kreher, ‘Agencies in the European Community— A Step towards Administrative Integration in Europe’, (1997) 4 JEPP 2, pp 232– 3.

2 These Committees were established by two Directives dating from 1975 and 1981: see Directive (EEC) 75/ 319 of the Council, OJ 1975 L 147/ 13; Directive (EEC) 81/ 851 of the Council, OJ 1981 L 317/ 1. Under the current EMA Regulation they are referred to as the Committee for Medicinal Products for Human Use and the Committee for Medicinal Products for Veterinary Use: see Article 56 of the EMA Regulation.

3 See The Committee of Wise Men, Final Report on the Regulation of European Securities Markets, 15 February 2001, http:// ec.europa.eu/ internal_ market/ securities/ docs/ lamfalussy/ wisemen/ final- report- wise- men_ en.pdf.

http://ec.europa.eu/internal_market/securities/docs/lamfalussy/wisemen/final-report-wise-men_en.pdf




The Origins of Agencies 103

at the fourth level.4 To coordinate at the third level, three committees, distinct from the comitology committees, were established. These committees were the European Banking Supervisors (CEBS), the Committee of European Insurance and Occupational Pensions Supervisors (CEIOPS), and the Committee of European Securities Regulators (CESR),5 the latter in itself being a formalized structure for the Forum of European Securities Commissions, a national regula-tors’ bottom- up initiative dating from the 1990s.6 The tasks of these three com-mittees have now been taken over by the ESAs. By establishing the ESAs, the legislator adopted the solution of creating genuine agencies, an idea which had already been floated at the time of the committees’ establishment.7

In establishing the EEA, the EU legislator created a permanent structure for the substantive work already carried out under the Corine programme (cf. section IV 1.1.4).8 The origins of Eurojust are similar since it has been established as a more formalized and permanent structure of existing cooperation.9 The principal deci-sion to create Eurojust was taken at the Tampere European Council in 1999 and (provisionally) concluded a process kicked off by the Maastricht Treaty. Using the new justice and home affairs (JHA) powers, in 1996 the Council adopted a joint action establishing a framework for the exchange of liaison officers.10 In 1998 the Council further built on this framework to establish a European Judicial Network (EJN).11 In 2002 the Eurojust Decision further strengthened the means of cooperation and coordination, stressing the link between the EJN and Eurojust by providing the EJN secretariat would be housed in the Eurojust secretariat.12

The origins of Frontex may be traced to a Council working group under Article 74 TFEU which provides that the Council may adopt measures to ensure adminis-trative cooperation between the Member States in the Area of Freedom, Security and Justice (AFSJ). Following the entry into force of the Treaty of Amsterdam, the Council established a Strategic Committee on Immigration, Frontiers and Asylum (SCIFA).13 For matters related to border control and border operations,

4 See Elliot Posner, ‘The Lamfalussy Process: Polyarchic Origins of Networked Financial Rule- Making in the EU’, in Sabel and Zeitlin (eds), Experimentalist Governance in the European Union, Oxford, OUP, 2010, p 47.

5 Decision (EC) 2001/ 527 of the Commission, OJ 2001 L 191/ 43; Decision (EC) 2004/ 5 of the Commission, OJ 2004 L 3/ 28; Decision (EC) 2004/ 6 of the Commission, OJ 2004 L 3/ 30.

6 On this and the transformation from the CESR to the ESMA, see Pierre- Henri Conac and Vincent Caillat, ‘De CESR à l’ESMA: le Rubicon est franchi’, (2010) novembre Bulletin Joly Bourse 6, pp 500– 10.

7 See Beatrice Vaccari, ‘Le processus Lamfalussy: une réussite pour la comitologie et un exemple de “bonne gouvernance europeenne”’, (2005) RDUE 4, p 807.

8 See also the account of the EEA’s origins in Anthony Zito, ‘European Agencies as Agents of Governance and EU Learning’, (2009) 16 JEPP 8, pp 1229– 31.

9 For an account of the establishment of Eurojust, see Michel Mangenot, ‘Un organe judici-aire pour l’Union européenne: Eurojust (1999– 2004)’, (2005) EIPA Scope 1, pp 27– 31.

10 Joint Action (JHA) 96/ 277 of the Council, OJ 1996 L 105/ 1.11 Joint Action (JHA) 98/ 428 of the Council, OJ 1998 L 191/ 4.12 The new EJN Decision aims to ensure a better cooperation between the EJN and Eurojust.

See Decision (JHA) 2008/ 976 of the Council, OJ 2008 L 348/ 130.13 See Council of the European Union, 16 March 1999, Doc 6166/ 2/ 99, p 3.

The Political Limits to Agencification104

the SCIFA group was enlarged with the heads of the different national border agencies, resulting in SCIFA+. In its plan for the management of external bor-ders of 2002 the Council took it upon itself to better structure this cooperation by setting up a common unit of external borders practitioners.14 However, the Commission found the SCIFA+ setting inadequate for the actual coordination in the field.15 To overcome the limits inherent to a Council working party, the Commission proposed to establish a new agency, resulting in Frontex.16

The origins of the European Chemicals Agency (ECHA) may be traced back to the Joint Research Centre (JRC), a Commission DG housing the Commission’s own scientific service,17 itself dating back to the 1950s. In the 1990s the Commission established a European Chemicals Bureau within the JRC to help support the EU’s activities in the field of chemicals.18 The justifications for est ablishing the Bureau, which was not an agency, indeed fit well with some of the functional ac-counts of agency creation today (cf. section III 2.1). When the Commission devel-oped its plans for the REACH programme in the 2000s, the initial plan was also to expand the existing Bureau,19 but following further study, the Commission ultimately proposed establishing a genuine agency.20 Incidentally, enlarging the Bureau was also proposed by the Danish presidency in 1993 as an alternative to the creation of the European Agency for Safety and Health at Work (EU- OSHA) when the negotiations on that dossier had momentarily stalled.21

The Agency for the Cooperation of Energy Regulators (ACER) was established as part of the third energy package and replaced the European Regulatory Group for Electricity and Gas (ERGEG) as the platform for regulatory cooperation and coordination.22 The ERGEG had been established by the Commission under the second energy package,23 to formalize the already ongoing cooperation between national regulators within the Florence (electricity) and Madrid (gas) Forums,24

14 Council of the European Union, 14 June 2002, Doc 10019/ 02, pp 12– 13.15 European Commission, COM (2003) 323 final, p 7.16 Following the entry into force of the Lisbon Treaty, Frontex also took over the tasks of

CIREFI, another Council working party established in the 1990s. See Council of the European Union, 22 February 2010, Doc 6504/ 10. For the decision establishing CIREFI, see Conclusions of the Council, OJ 1996 C 274/ 50.

17 A more elaborate account of the establishment of the ECHA may be found in Maria Martens, ‘Executive Power in the Making: The Establishment of the European Chemicals Agency’, in Busuioc, Groenleer, and Trondal (eds), The Agency Phenomenon in the European Union, Manchester, Manchester University Press, 2012, pp 42– 62.

18 See Communication of the European Commission, OJ 1993 C 1/ 3.19 European Commission, COM (2001) 88 final, p 25.20 See European Commission, COM (2003) 644 final. Interestingly, the Commission still sug-

gested establishing the agency in Ispra to ensure continuity. However, since Finland lost out against Italy for the seat of the EFSA, the ECHA seat was ‘given’ to Finland.

21 The other alternative suggested by the presidency was the creation of a task force within the Commission. The Commission’s legal service rightly noted that these alternative solutions squarely fell within the Commission’s power of internal organization and could not in any event be ‘ordered’ by the Council. See Commission européenne, 24 mars 1993, ‘Note de la Service Juridique à l’attention de M.S. Crespo Valera’, JUR (93) 2606.

22 Decision (EU) 2011/ 280 of the Commission, OJ 2011 L 129/ 14.23 Decision (EC) 2003/ 796 of the Commission, OJ 2003 L 296/ 34.24 See recital 5 of the preamble to the ERGEG Decision.

The Origins of Agencies 105

which had been set up under the first energy package by DG Energy.25 By estab-lishing the ERGEG, the Commission also recuperated the Council of European Energy Regulators (CEER), a national regulators’ bottom- up initiative doing much the same work as ERGEG,26 which now continues to work closely with ACER.

Apart from the ERGEG, other European Regulators Groups can or could be found at the EU level.27 The European Regulators Group for Electronic Communications Networks and Services (ERGECNS), for instance,28 was sup-posed to be abolished and replaced by the European Electronic Communications Agency (ECMA),29 but in the end the Body of European Regulators for Electronic Communications (BEREC) replaced the ERGECNS.30

Prior to the establishment of the European Asylum Support Office (EASO), practical cooperation in the context of the Common European Asylum System (CEAS) was undertaken in a number of informal forums. One of them was the Commission’s European Union Network for Asylum Practitioners (EURASIL) expert group established in 2002 to take over the work of the Centre for Information, Discussion and Exchange on Asylum (CIREA) established by the Council in 1992.31 In the field of asylum there were also bottom- up initiatives by the Member States, such as development of a European Asylum Curriculum. The Commission observed that these informal instruments lacked efficiency,32 and proposed the EASO as a solution to the challenges in this field.

Tracing back the origins of a selection of EU agencies shows that they are rarely established out of the blue. Usually agencies have their (more modest) precursors within the Commission, the Council, or even outside the EU. Indeed, it is not exceptional to see the Commission recuperating bottom- up initiatives organized by national administrations.

The preceding shows how the process of agencification is incremental, building on or elaborating existing structures. This insight is fundamental in light of the political limits to agencification since it shows how the establishment of an agency is part of a broader process of administrative capacity- building at EU level and coordination between national authorities. The origins of agencies should also be remembered when confronted with the alleged reasons for establishing an agency. After all, there are always alternatives to agency creation, one of which (simply) is the status quo.

25 See Marco Zinzani, Market Integration Through ‘Network Governance’: The Role of European Agencies and Networks of Regulators, Antwerp, Intersentia, 2012, p 107.

26 Ibid, pp 114– 17.27 For the European Regulators Group for Postal Services, see Decision of the European

Commission, OJ 2010 C 217/ 7.28 Decision (EC) 2002/ 627 of the Commission, OJ 2002 L 200/ 38.29 See the original proposal, European Commission, COM (2007) 699 final.30 Decision (EU) 2010/ 299 of the Commission, OJ 2010 L 127/ 18.31 For the decision setting up CIREA, see Tony Bunyan, Key Texts on Justice and Home Affairs

in the European Union, London, Statewatch, 1997, pp 68– 9. On EURASIL and the transfer of tasks from CIREA to the Commission, see Council of the European Union, 30 April 2002, Doc 8101/ 02, p 1; European Commission, SEC (2009) 153, p 9.

32 See the Commission Staff Working Document, SEC (2009) 153, p 9.


2 Explaining Agencification

While the position of agencies and agencification in the EU administration was explained previously (cf. section I 1), a fundamental question remains why the EU legislator has increasingly resorted to this instrument instead of relying on the options provided for by the Treaties. This question also puzzled the early com-mentators of agencification. As Kelemen and Tarrant note, ‘[m] ost of the literature on EU […] agencies and networks explains their creation in functional terms’.33 Indeed, one often reads that agencies deal with information asymmetries,34 can take blame for principals,35 increase credible commitment,36 isolate decisions from political pressure,37 help pool expertise,38 enable social partners and third coun-tries to be involved in policy- making,39 help relieve the Commission from tech-nical tasks,40 help coordinate administrative activities,41 etc. The same functional

33 Daniel Kelemen and Andrew Tarrant, ‘The Political Foundations of the Eurocracy’, (2011) 34 WEP 5, p 923.

34 See inter alia Damien Geradin and Nicolas Petit, ‘The Development of Agencies at EU and National Levels: Conceptual Analysis and Proposals for Reform’, (2005) 23 YEL, p 170; Zinzani, n 25, p 34. Antun Bilić, Tatjana Josipović, and Siniša Petrović, ‘Independent Regulators in the Network Industries’, in Bodiroga- Vukobrat, Gerald, and Barić (eds), Regulierungsagenturen im Spannungsfeld von Recht und Ökonomie, Hamburg, Kovac, 2012, p 244.

35 See inter alia Renaud Dehousse, ‘Regulation by Networks in the European Community: The Role of European Agencies’, (1997) 4 JEPP 2, p 258; Nada Bodiroga- Vukobrat and Adrijana Martinovic, ‘European Decentralised or Regulatory Agencies— Quest for a Common Approach’, in Bodiroga- Vukobrat, Sander, and Barić (eds), Regulierungsagenturen im Spannungsfeld von Recht und Ökonomie, Hamburg, Kovac, 2012, pp 75– 6.

36 See inter alia Giandomenico Majone, ‘The Credibility Crisis of Community Regulation’, (2000) 38 JCMS 2, p 289; Dorothee Fischer- Appelt, Agenturen der Europäischen Gemeinschaft, Berlin, Duncker & Humblot, 1999, pp 200– 1.

37 See inter alia Robert Uerpmann, ‘Mittelbare Gemeinschaftsverwaltung durch gemein-schaftsgeschaffene juristische Personen des öffentliches Rechts’, (2000) 125 AöR 4, p 562. Martijn Groenleer, ‘The European Commission and Agencies’, in Spence (ed), The European Commission, London, John Harper, 2006, p 163.

38 See inter alia Giandomenico Majone, ‘The Agency Model: The Growth of Regulation and Regulatory Institutions in the European Union’, (1997) 6 EIPA Scope 3; Jacqueline Dutheil de la Rochère, ‘EU Regulatory Agencies: What Future do They Have?’, in Bulterman, Hancher, McDonnell, and Sevenster (eds), Views of European Law from the Mountain, Alphen aan den Rijn, Kluwer Law International, 2009, p 355.

39 Isabelle Muller- Quoy, ‘L’apparition et le développement des Agences de l’Union europée-nne’, in Couzinet (ed), Les Agences de l’Union européenne, Toulouse, PUSST, 2002, p 17.

40 See inter alia Michaela Wittinger, ‘Europäische Satelliten: Anmerkungen zum Europäischen Agentur(un)wesen und zur Vereinbarkeit Europäischer Agenturen mit dem Gemeinschaftsrecht’, (2008) 42 Europarecht 5, p 618; Muller- Quoy, n 39, p 17; Groenleer, n 37, p 163; Loïc Grard, ‘Le Contrôle des Actes des Agences de Régulation: Analyse Comparée’, in Peraldi Leneuf and Normand (eds), La légistique dans le système de l’Union européenne: quelle nouvelle approche, Bruxelles, Bruylant, 2012, p 138; Michael Berger, Vertraglich nicht vorgese-hene Einrichtungen des Gemeinschaftsrechts mit eigener Rechtspersönlichkeit, Baden- Baden, Nomos, 1999, Vol 219, pp 89– 90; Ronald van Ooik and Tom Eijsbouts, ‘De wonderbaarlijke ver-menigvuldiging van Europese agentschappen. Verklaring, analyse, perspectief ’, (2006) 54 SEW 3, p 103.

41 Uerpmann, n 37, p 562.

Explaining Agencification 107

reasons spurring national agencification are thereby often cited but, as Halberstam notes, national agencification should also be seen as the result of specific constitu-tional and interinstitutional dynamics.42

2.1 The functional accounts of agencification

Closely scrutinizing these functional reasons reveals they merely make the case for an independent body to which certain tasks are entrusted. After all, most of these advantages can equally be attained by empowering the Commission. Sometimes EU agencies do have a functional ‘bonus’. In the EU risk regulatory regime, for instance, risk assessment and management are separated, which is translated in the organizational separation of the European Medicines Agency (EMA) and European Food Safety Authority (EFSA) from the Commission. EU agencies may also bring more visibility to the EU’s policies compared to the Commission.43

While agencies fulfil certain functional needs, this does not mean that they are also indispensable to address these needs. The functional accounts do not readily explain why the alternative solutions suggested by the Treaty itself (granting executive powers to the Member States, Commission or Council) are not selected. In addition, Thatcher observes that the functionalist accounts of agency creation have not adequately predicted EU agencification, since EU agencies have not (firstly) been set up in those areas where there were the strong-est functional reasons to do so.44 Perhaps this is why these functional accounts have not been integrated in a neo- functionalist narrative. In fact, Shapiro is one of the few who notes that ‘a kind of neofunctionalism’ may be observed in agencification:

If currently direct routes to further political integration of the Union are blocked […] further growth can be achieved indirectly through the proliferation of small, limited jurisdictions, allegedly ‘technical agencies’ that will appear politically innocuous.45

42 Daniel Halberstam, ‘The Promise of Comparative Administrative Law: A Constitutional Perspective on Independent Agencies’, in Rose- Ackerman and Lindseth (eds), Comparative Administrative Law, Cheltenham, Edward Elgar, 2010, pp 193– 200.

43 See inter alia Olivier De Schutter, ‘La procéduralisation dans le droit européen— Propositions institutionelles’, in De Schutter, Lebesis, and Paterson (eds), La gouvernance dans l’Union européenne, Luxembourg, Office des publications officielles des Communautés euro-péennes, 2001, pp 229– 31; Ellen Vos, ‘Reforming the European Commission: What Role to Play for EU Agencies?’, (2000) 37 CMLRev 5, p 1119. However, Versluis has noted that the geo-graphic dispersion of agencies can also make them less visible, see Esther Versluis, ‘Europese Agentschappen: de tentakels van Brussel?’, in Van der Vleuten (ed), De bestuurlijke kaart van de Europese Unie, Bussum, Coutinho, 2007, p 202.

44 Mark Thatcher, ‘The Creation of European Regulatory Agencies and Its Limits: A Comparative Analysis of European Delegation’, (2011) 18 JEPP 6, p 791.

45 Martin Shapiro, ‘The Institutionalization of European Administrative Space’, in Stone Sweet, Sandholtz, and Fligstein (eds), The Institutionalization of Europe, Oxford, OUP, 2001, p 281.


2.1.1 A neo- functionalist account of agencificationStill, Shapiro himself has not elaborated this observation into a neo- functionalist explanation of agencification. De Moor- van Vugt also alluded to a neo- functionalist motive when she described the institutionalization of networks of national regulators at EU level. In the more advanced stages of this process, these networks may be complemented by an EU agency and the Commission will strengthen its control over national authorities (cf. section III 3.1).46 At first sight, agencification indeed seems to exteriorize the results of spillover. Since the Single European Act (SEA), the competences of the EU have expanded with every Treaty revision, in part because of pressure caused by functional spillover. To illustrate: to achieve a genuine free movement of goods in the food, drugs, and chemicals markets, risks associated with these products should be assessed and managed at the European level. Similarly, to maximize the benefits brought by the free movement of persons, systematic controls at intra- EU borders should be abolished, but to do so necessitates a common regime for the control at the EU’s external border. In each of these areas, agencies have been established.

However, this does not mean that those agencies are the result of functional spillover. Functional spillover may indeed explain why integration proceeds in new policy areas, but it does not explain why an agency should be established. The notion of political spillover does not help much either, since it only explains centralization, while the distinctive element in agencification is that it is an atyp-ical form of centralization.

The direct route to further political integration, alluded to by Shapiro, would result in the conferral of implementing powers on the Commission and would cut off national administrations from the implementation of EU law. Logically, the Member States are averse to such a development.47 To nevertheless achieve further integration, the Commission has proposed the establishment of agencies through and in which national administrations can coordinate activities on a permanent basis. Because of the institutional design of these agencies, they are politically much more acceptable to the Member States than is the expansion of the Commission’s powers. The problem for neo- functionalist theory in explaining agencification then has to do with one of the general shortcomings of the theory, which has always been about ‘process’, not ‘product’.48 However, agencification is not just about more integration (process), but rather about a very specific form which this integration takes (product).

46 Adrienne de Moor- van Vugt, ‘Netwerken en de europeanisering van het toezicht’, (2011) 59 SEW 3, pp 100– 2.

47 Daniel Kelemen, ‘The Politics of “Eurocratic” Structure and the New European Agencies’, (2002) 25 WEP 4; Renaud Dehousse, ‘Delegation of Powers in the European Union: The Need for a Multi- principals Model’, (2008) 31 WEP 4, p 796; Merijn Chamon, ‘EU Agencies: Does the Meroni Doctrine Make Sense?’, (2010) 17 MJECL 3, pp 286– 8.

48 Philippe Schmitter, ‘Neo- neofunctionalism’, in Wiener and Dieze (eds), European Integration Theory, Oxford, OUP, 2004, p 69.


Although Shapiro’s assertion may have an intuitive appeal to it, neo- functionalist theory cannot fully explain agencification. Although it may explain the pressures leading to further integration, explaining the institutions’ choice for agencies falls outside the scope of the neo- functionalist model.49 In general, then, the traditional functional accounts of agency creation do not provide a forceful explanation for agencification. This further shows how the choice for an agency (among other alternatives) is not technical but rather political. A neo- functionalist account does not allow the identification of any political limits to agencification other than those identified for the legislative integration preceding administrative integration. The explanation of agencification should therefore be searched for elsewhere.

2.2 Non- functional accounts of agencification

Here it is useful to recall that agencification was already qualified as furthering administrative integration, whereby ‘integration’ may be defined as the process of increasing interaction obscuring the boundaries between the international and national environments.50 The reason why the agency instrument is relied on is the subject of disagreement between the authors proposing non- functional accounts of agency creation, incidentally showing how the different institutions may also have different reasons for supporting agencification.

Just as the functional accounts have not been grounded in a functional theory of regional integration, the authors invoking political reasons for agencification often do so in absence of a theoretical underpinning. Agencies are thus said to be created as a response to the continuing politicization of the Commission,51 or to put the principle of sincere cooperation into effect,52 since they build networks in which national bodies participate,53 fostering administrative integration.54 On the other hand, and seemingly completely opposed to the last view, some see agencification as a means to protect national law from Europeanization.55 Chiti proposes a dialectic synthesis to these positions, describing agencies as forming part of decentralized integration.56 Agencies are also said to increase

49 Similarly, see Kelemen and Tarrant, n 33, p 925.50 Ernst Haas, Beyond the Nation- state: Functionalism and International Organization, Stanford,

Stanford University Press, 1964, p 39.51 Marc Blanquet, ‘Agences de l’union et gouvernance européenne’, in Couzinet (ed), Les

Agences de l’Union européenne, Toulouse, PUSST, 2002, pp 48– 9.52 Ibid, pp 57– 61.53 Thomas Gross, ‘Die Kooperation zwischen europäischen Agenturen und nationalen

Behörden’, (2005) 39 Europarecht 1, p 58.54 Kreher, n 1.55 Stefan Kadelbach, ‘European Administrative Law and the Law of a Europeanized

Administration’, in Joerges and Dehousse (eds), Good Governance in Europe’s Integrated Market, Oxford, OUP, 2001, p 43.

56 Edoardo Chiti, ‘Decentralisation and Integration into the Community Administrations: A New Perspective on European Agencies’, (2004) 10 ELJ 4, pp 404– 5. See also Herwig Hofmann, ‘Agency Design in the European Union’, (2010) 28 Windsor Yearbook of Access to Justice, p 311.


the regulatory capacity of the Union,57 without choosing the centralizing supra-national option.58

The question whether agencification is a form of decentralization can be debated. Although the Commission is keen on presenting agencification as such,59 and the three main institutions have adopted the Common Approach on Decentralised Agencies, agencification is actually a form of centralization.60 However, for ob-vious political reasons, this process is presented as centrifugal.

A number of authors have suggested that EU agencies, as alternatives to the opaque comitology committees, are a response to the EU’s democratic deficit,61 but in reality EU agencies are often involved in procedures together with comitology committees.62 At least one commentator has even suggested in general terms that the increased reliance on agencies is a response to the EU’s democratic deficit,63 while agencies have at the same time been described as a danger to democracy.64

Lastly, on a different level, it cannot be excluded that agencification is attractive because it allows Member States to host these bodies in their own countries or because agencies may be used as solutions to policy crises or problems which are visible to the general public.65 The EFSA could then be seen as a response to the food scares of the 1990s, the European Maritime Safety Agency (EMSA) to the Erika and Prestige disasters,66 the ESAs to the financial crisis of 2008, etc.67 Here a logic of appropriateness seems to have developed whereby EU agencies are seen as legitimate solutions to any kind of policy problem.68

57 Groenleer, n 37, p 164.58 Dehousse, n 35, pp 521– 7.59 European Commission, COM (2002) 718 final, p 2. 60 See Chapter I, n 37.61 Daniel Kelemen, ‘The Politics of Eurocracy: Building a New European State?’, in Jabko and

Parsons (eds), The State of the European Union: With US or against US? European Trends in American Perspective, Oxford, OUP, 2004, Vol. 7, p 179; Maria Luisa Sanchez Barrueco, ‘L’Agence euro-péenne de défense, un organe intergouvernemental au service d’une institution communautaire?’, (2008) RDUE 3, pp 511– 12. See also Laurie Buonanno, ‘The Creation of the European Food Safety Authority’, in Ansell and Vogel (eds), What’s the Beef? The Contested Governance of European Food Safety, Cambridge, MIT Press, 2006, pp 261– 2.

62 See also Merijn Chamon, ‘Comitologie onder het Verdrag van Lissabon’, (2013) 61 SEW 2, p 73.

63 See Sophie Perez, ‘Les autorités administrative indépendantes: Rapport Union européenne’, (2009) 31 Annuaire européen d’administration publique 2008, p 298. This argument is odd at first sight but it makes more sense if agencies are seen as improving the output- legitimacy of the EU (cf. section II 2.1.3).

64 See Thomas Gross, ‘Unabhängige EU- Agenturen— eine Gefahr für die Demokratie?’, (2012) 67 JZ 22, pp 1087– 93.

65 Gross, n 53, p 67.66 See Wolfgang Kilb, ‘Europäische Agenturen und ihr Personal— die großen Unbekannten?’,

(2006) 17 EuZW 9, p 271.67 Similarly, see the revision of the EMSA Regulation in light of the Deepwater Horizon disaster.

See Regulation (EU) 100/ 2013, OJ 2013 L 39/ 30. In addition, following the Costa Concordia dis-aster in January 2012, suggestions for a stronger role for the EMSA were made: see Parliamentary Question (E- 508/ 2012) of MEP Niccolò Rinaldi on minimum safety distances of international shipping routes from coastlines, 25 January 2012.

68 See Merijn Chamon, ‘The European Railway Agency under the Fourth Railway Package: A Political and Legal Perspective’, in Rapoport (ed), L’espace ferroviaire unique euro-péen: Quelle(s) réalités, Bruxelles, Bruylant, 2015, p 173.


Just like the functional accounts of agencification, some of these political mo-tives may have an intuitive appeal, but fall short of explaining the political choices behind agencification. As Thatcher has noted, ‘creating an [EU agency] is an in-stitutional choice by policy- makers with their own interests and strategies from among competing alternatives’.69

By looking at EU agencies as conscious institutional choices, rather than mech-anical consequences of a functional logic, a Rational Choice perspective appears promising to explain how the agency instrument furthers the (individual) inter-ests of the competent actors. These different interests may also explain the dual nature of the phenomenon of agencification as an atypical form of integration. As Kelemen and Tarrant point out, the wider institutional context within which agencies are designed and embedded needs to be understood.70 This includes mapping the alternatives to agencies and the relevant actors.

2.2.1 General delegation theory in European integration studiesThatcher sees three main alternatives to agency creation: (i) networks, (ii) direct administration, or (iii) indirect administration.71 The latter two have indeed been the subject of most of the delegation research on the implementation of EU law.

Franchino attempted to explain the legislator’s choice between empowering the Commission or the national administrations.72 Although he is not oblivious to the existence of EU agencies,73 presenting the issue as a binary choice for the legislator has become too simplistic. Perhaps the very essence of the agency phe-nomenon from a political perspective is that the EU legislator has been able to escape the rigid dichotomy of direct and indirect administration. The studies by Thomson and Pollack suffer from the same shortcoming.74 As a result, while del-egations to the Commission have been studied to some extent they have not been linked to the question of delegation to agencies, even if both are ultimately linked.

2.2.1.1 Adding agencies to the equationThe fact that none of these three authors took account of agencification may relate to a more general critique on Rational Choice supranationalists, who allegedly do not leave room for ‘learning’ on the part of the Member States.75

69 Thatcher, n 44, p 792.70 Kelemen and Tarrant, n 33, p 926.71 Kelemen and Tarrant also list these alternatives but leave out the possibility of traditional in-

direct administration. See ibid. Blanquet provides two further alternatives: (i) (genuine) decentral-ization to national or transnational authorities and (ii) outsourcing to the private sector. An example of decentralization to national bodies may be seen in Decision (EC) 94/ 458 of the Commission, OJ 1994 L 189/ 84.

72 Fabio Franchino, The Powers of the Union: Delegation in the EU, Cambridge, CUP, 2007, p 17.73 Ibid, p 16.74 Robert Thomson, Resolving Controversy in the European Union: Legislative Decision- making

Before and After Enlargement, Cambridge, CUP, 2011, pp 252– 77; Mark Pollack, The Engines of European Integration: Delegation, Agency and Agenda Setting in the EU, Oxford, OUP, 2003, pp 152– 4.

75 Hussein Kassim and Anand Menon, ‘The Principal- agent Approach and the Study of the European Union’, (2003) 10 JEPP 1, pp 132– 3.


The establishment of agencies could then be seen as a way of working around the problem which an insufficiently controllable Commission poses. In another way, agencification could be seen as a way of sanctioning the Commission. As Pollack noted, it is hard to cut the budget to sanction the Commission.76 Similarly, it is hard for Member States to amend the Treaties. But what if the Commission’s constitutional position is undermined by the creation of other, rival, entities unforeseen in the Treaties? By establishing such entities the Member States substantively ‘amended’ the Treaties through secondary law and have sanctioned the Commission because the Union legislator is no longer obliged to charge the Commission with executive tasks, as is now also confirmed by the Court in Short- selling.

Does this mean that the process of agencification invalidates the research of Pollack and Franchino?77 Not necessarily, since the type of functions delegated to the Commission differ from the tasks generally attributed to the agencies. And if the delegation of tasks to EU agencies does not involve core Commission tasks, does it make sense to speak in terms of ‘sanctioning the Commission’? Still it should be noted that a small number of agencies— not coincidentally the more powerful ones— fulfil functions which Pollack and Franchino would also at-tribute to the Commission.

To explain why powerful agencies have been established in some sectors but not in others, Thatcher relies on previous patterns of delegation. Firstly, the Commission will try to safeguard its own position. As a result, new agencies will not be proposed where they might be a threat to the Commission’s powers or in any event the agency will have modest formal powers.78 Similarly, if the Member States have established powerful national agencies, the latter will plead for a form-alized EU network and against a strong EU agency.79 However, as Thatcher him-self notes, national agencies will not be able to veto agency creation at EU level. In addition, although the Commission does have this veto power,80 it could still agree to establish an agency which might damage its own institutional interests as part of the trade- off with greater integration within the EU context.81

2.2.1.2 Relevant actors for agency creationThatcher’s insights on previous patterns of delegation illustrate the importance of identifying the relevant actors in agency creation.

76 This is only so nominally. In real terms the budget has already shrunk. See European Commission, SEC (2007) 1188 final, p 4.

77 Franchino’s research is founded on the assumption that the EU legislator has to choose be-tween indirect and direct administration and that the preferences of national administrations are the same as their national ministers in the Council. See Franchino, n 72, p 26. The existence of agencies undermines the first and affects the second of these core assumptions. Agencies are new actors with possibly new preferences and add to the complexity of decisional reality.

78 Thatcher, n 44, pp 800– 2. 79 Ibid, pp 802– 5.80 See however Chapter V, n 100.81 See for instance the Commission’s difficult position when the legislator empowered the

ESAs: section I 1.2.1.3.1.f. See also Merijn Chamon, ‘EU Agencies between “Meroni” and “Romano” or the Devil and the Deep Blue Sea’, (2011) 48 CMLRev 4, p 1068.


Under the ordinary legislative procedure, the EU legislator consists of the Parliament and the Council. In addition, there is the Commission, who has the power of proposal.82 Because the Commission does not only have a specific bur-eaucratic interest but also supports deeper integration in general, it may often sup-port legislative initiatives which further integration, without furthering its own bureaucratic interests when securing both interests is difficult.83

According to Kelemen and Tarrant, for the Parliament the same two interests are complemented by a third, which is the interest Parliament has in maximizing its popularity among voters. According to Dehousse,84 the Parliament further shares the Commission’s lukewarm view to agencification because it could undermine its recently acquired legislative powers, while Kelemen and Tarrant note that the Parliament is attracted by the agency model because it ‘was granted with consider-able oversight powers with respect to agencies’.85 However, Parliament’s oversight powers can only be assessed in a relative sense, that is, compared to its powers over agency alternatives. The Parliament then actually loses if the comparison is made with its powers over the Commission.86 As a result, Dehousse’s observations seem to reflect the Parliament’s position better, being driven both by the concern for its legislative power and by its control function over the EU executive. In addition, the change in legal bases used for establishing agencies and the generalization of the ordinary legislative procedure have turned the Parliament from an outsider to an insider as regards agencification, resulting in a much less critical stance.87

The Member States then are caught between the need to gain benefits from col-lective action and the desire to control EU actors,88 the difference between an EU agency and an EU network being the control which Member States can exercise.89

2.3 Concluding remarks

The motives underlying EU agencification have puzzled authors since the 1990s. Now, a consensus is emerging on the inadequacy of the early functional accounts. Undoubtedly, agencies fulfil a functional need, that is, strengthening the EU’s regulatory and administrative capacity. However, establishing an agency is only one way in which this capacity may be strengthened. This is where functionalist

82 Kelemen and Tarrant, n 33, pp 927– 30.83 Ibid, pp 927– 8; Dehousse, n 47, p 796.84 Dehousse, n 47, pp 796– 7.85 Kelemen and Tarrant, n 33, p 929.86 See also Christopher Lord, ‘The European Parliament and the Legitimation of

Agencification’, in Rittberger and Wonka (eds), Agency Governance in the EU, London, Routledge, 2011, p 135.

87 In line with Franchino’s findings, increased conflict between the Parliament and Commission may also result in a Parliament more inclined to delegate tasks to an agency instead of to the Commission.

88 Kelemen and Tarrant, n 33, p 929. See also Edoardo Chiti, ‘Les agences, l’administration indirecte et la coadministration’, in Auby and Dutheil de la Rochère (eds), Droit administratif euro-péen, Bruxelles, Bruylant, 2007, pp 268– 9.

89 See Kelemen and Tarrant, n 33, pp 930– 1.


theory can no longer explain the process of agencification, since it is foremost con-cerned with process, that is, increasing integration, rather than the form which this integration takes.

In more recent years, a number of authors have concluded that the agency instrument is a conscious political choice as an alternative to strengthening the Commission. Although the latter alternative is prescribed by the Treaties, the Member States have rejected it, concerned about the autonomy of their own ad-ministrations and following a learning process of earlier delegations. Whereas the Commission and Parliament would preferably opt for a strengthened Commission, the objective of strengthening the EU’s regulatory capacity takes precedence over their institutional interests. As a result, both institutions acquiesce to a second- best option.

3 Agencification and Its Effects

What do the context and dynamics of agencification tell us about the ‘product’ of agencification? The dynamics of the process cannot a fortiori be equated with its effects. Indeed, it cannot be simply assumed that the result of agencification is in line with the institutions’ interests driving the process.

Here again the New Institutionalism school helps in understanding this rel-ationship, although the paradigm used will shift from a Rational Choice to a more historical perspective. This has to do with a critique on Rational Choice theory formulated early on by Simon,90 noting that human beings are not ‘hyper- rational’, since a perfectly rational actor would require both unlimited cognitive capacities and complete information.

Simon thus proposed the notion of bounded rationality:91 actors taking bounded rational decisions whereby they cannot foresee all present and future eventualities. In casu: following a decision to establish an agency, the agency becomes an actor in its own right and becomes part of the institutional con-text. Subsequent decisions are then taken in this new context. In historical New Institutionalism terms, the processes of ‘increasing returns’ and ‘lock- ins’ imply that initial choices become entrenched and will be difficult to alter, even if the policy environment changes and even if the institutions would no longer come to the same decision in the new environment.92 The increasing returns are processes of positive feedback which make a reversal of an initial decision increasingly un-attractive over time.93 In simple terms: many different solutions may be worked

90 Herbert Simon, ‘A Behavioral Model of Rational Choice’, (1955) 69 The Quarterly Journal of Economics 1, pp 99– 118.

91 Stefano Fiori, ‘Forms of Bounded Rationality: The Reception and Redefinition of Herbert A. Simon’s Perspective’, (2011) 23 Review of Political Economy 4, p 589.

92 See Mette Eilstrup- Sangiovanni, Debates on European Integration: A Reader, Basingstoke, Palgrave Macmillan, 2006, p 198.

93 This is because decisional rules form barriers to entry and because actors will adapt to the new environment, resulting in path dependency. See Eilstrup- Sangiovanni, n 92,

Agencification and Its Effects 115

out to respond to a policy problem but if such a problem arises in a policy field in which an EU agency has already been established, the solution which sees a reinforced role of the agency will have a head start against other, possibly more ‘rational’ solutions.94

This helps to explain that even if solutions do not serve the (bounded) rational actors’ interests anymore, they may still persist. In addition, the result of the pro-cess of agencification may be seen as more than the sum of the different decisions to establish individual agencies. The question what the actual result is of the on-going agencification may indeed be raised. The tension inherent in the phenom-enon is that between integration and national autonomy. It is this same tension which explains why both Kreher and Kadelbach may be right (cf. section III 2.2), even if the first emphasizes the aspect of administrative integration while the latter refers to agencification as a means to protect national law from Europeanization.95

3.1 EU agencies: Trojan horses, part of executive centre formation, or part of a multi- layered administration?

In comparison with Kadelbach, de Moor- van Vugt is less optimistic when she des-cribes European networks as ‘Trojan horses’.96 According to de Moor- van Vugt the EU strengthens its grip on the implementation of EU law through the Member States in successive stages. While the Member States first enjoy autonomy, the Commission will establish informal and later formal networks in order to coordinate national administrations. Then an EU agency with a modest mandate is established, the mandate later being enhanced. Ultimately all competences become centralized, whereby the executive competences will in turn be delegated back to the national level under very strict terms,97 with national administrations acting as the extended arm of the Commission.98

Caution is warranted when this embryonic model is taken as a general model for agencification, since it is not an iron law describing the process. Further, al-though de Moor- van Vugt sees the Commission as driving agencification, it is im-portant to note that each time a qualitative step is taken in the process, the three main EU institutions will need to agree on this.

pp 198– 9; Margaret Levi, ‘A Model, a Method, and a Map: Rational Choice in Comparative and Historical Analysis’, in Lichbach and Zuckerman (eds), Comparative politics: Rationality, Culture, and Structure, Cambridge, CUP, 1997, p 28; Siobhán Harty, ‘Theorizing Institutional Change’, in Lecours (ed), New Institutionalism: Theory and Analysis, Toronto, University of Toronto, 2005, p 57.

94 See Chamon, n 68, p 173. 95 See nn 54 and 55.96 de Moor- van Vugt, n 46, p 100.97 The notion of delegation is used here because de Moor- van Vugt also refers to it. However,

whether delegation is the apt term in casu is questionable (see section IV 5.7). For this, the constitu-tional foundations of the EU would need to be changed, clearly providing that the original power to implement EU law is vested at EU level. Only then could the EU institutions delegate these powers to the national administrations. See also however Dubos, Chapter I, n 266.

98 de Moor- van Vugt, n 46, pp 100– 1.


The idea that agencification strengthens the Commission’s hold over national administrations is also advanced by Egeberg, who sees two trends resulting in the gradual incorporation of the national administrations in the administrative ap-paratus of the EU: (i) a greater autonomy and reinforcement of the Commission, and (ii) a process of (national) agencification resulting in autonomous bodies at national level. After first having been detached from the traditional national min-istries, their autonomy allows them to be recoupled to the Commission and/ or EU agencies, within European networks.99

Of course, Egeberg’s first claim is debatable. He sees a strong Commission, compared to federal executive actors in other polities (eg US, Germany, UN, etc), because the Commission President’s role has been strengthened since Amsterdam, and because Commission officials are socialized in a European context.100 But one could also argue that the Commission as an institution is under continuous pressure,101 and that it no longer has the same prestige that it did under Delors. Furthermore, the way in which the Court sanctioned agencification in Short- selling has been a major blow for the Commission.102

Still, Egeberg is quite optimistic.103 According to him, the fact that more and more national agencies are created allows them to be placed under the joint super-vision of the respective Member State and the Commission. Egeberg then notes that ‘[l] a double casquette permet une exécution plus homogène des politiques européennes que la méthode indirecte. L’efficacité serait plus grande encore si la Commission disposait de ses propres agences ou si l’application de la loi euro-péenne était entre les mains de structures propres à l’Union européenne’.104 In further research, Egeberg and Trondal hypothesized that the EU agencies might contribute to a European ‘executive centre formation’.105

Linking the double trend of agencification in both the national and EU legal orders, it has also been suggested that in the existing multi- layered EU adminis-tration, a multi- layered agency may be created, composed of the national agen-cies and an EU agency linked in an EU network.106 The national agencification,

99 Morten Egeberg, ‘L’Administration de l’Union Européenne: Niveaux Multiples et Construction d’un Centre’, (2010) Revue française d’administration publique 133, p 18.

100 Ibid, pp 20– 21.101 As regards implementing powers Barents saw a weakening of the Commission by the Lisbon

Treaty, while Craig noted a reinforcement of the Commission. See René Barents, Het Verdrag van Lissabon, Achtergronden en commentaar, Deventer, Kluwer, 2008, pp 461– 2; Paul Craig, ‘The Hierarchy of Norms’, in Tridimas and Nebbia (eds), European Union Law for the Twenty- first Century: Rethinking the New Legal Order, Oxford, Hart, 2004, pp 80– 4.

102 Merijn Chamon, ‘Grenzen voor de EU- wetgever bij het machtigen van Europese agentschappen’, (2014) 29 RegelMaat 3, pp 139– 40.

103 Egeberg’s perspective seems to be that of other ‘executive’ bodies at the international level. See Egeberg, n 99, p 26. From this perspective, the Commission is indeed a powerful and autono-mous institution, but it would still be difficult to speak of a trend of a ‘reinforced’ Commission.

104 Ibid, pp 24– 5.105 See Morten Egeberg and Jarle Trondal, ‘EU- level Agencies: New Executive Centre

Formation or Vehicles for National Control?’, (2011) 18 JEPP 6, p 869.106 Merijn Chamon, ‘Verzelfstandiging in de nationale en Europese rechtsordes: nieuwe uit-

daging van een meergelaagde administratie’, (2013) TBP 2– 3, pp 117– 21.

Agencification and Its Effects 117

also identified by Egeberg, is often spurred by EU harmonization directives im-posing an obligation on the Member States to establish independent regulatory authorities.107 At the same time, the EU legislator adopts regulations establishing a European regulator for the same sector, where the heads of the national agen-cies will sit on the Board of the European agency. Agencies at both levels are then ordered to cooperate between themselves to regulate the economic sector concerned. This end result is different from the previous ones because it does not result in a strengthened Commission even if it contributes to an executive centre formation. Despite the fact that the Commission is often pictured as a techno-cratic body, the multi- layered agency would result in a much more technocratic EU administration.108


Egeberg and Trondal observe that ‘[t] he jury is still out’ with respect to the effects of agencification and whether it is a tool of national governments or instead results in executive centre formation,109 in a multi- layered agencified administration, or in still some other construct. In any event, the effects of agencification should be

107 See also Chapter V at n 73. In the telecoms sector for instance Article 3a was inserted in Directive 2002/ 21 by Directive (EC) 2009/ 140 of the European Parliament and of the Council OJ 2009 L 337/ 37. Recital 13 of the preamble to the latter Directive describes the purpose of the amendment as follows: ‘[E] xpress provision should be made in national law to ensure that, in the exercise of its tasks, a national regulatory authority […] is protected against external intervention or political pressure liable to jeopardise its independent assessment of matters coming before it.’ For the electricity sector a similar independence requirement is provided for in Article 35(4) of Directive (EC) 2009/ 72 of the European Parliament and of the Council, OJ 2009 L 211/ 55. Recital 33 of the preamble to that Directive explains: ‘[E]xperience shows that the effectiveness of regulation is frequently hampered through a lack of independence of regulators from government, and insuffi-cient powers and discretion.’ Identical provisions may be found for the gas sector in Article 39(4) and recital 29 of the preamble to Directive (EC) 2009/ 73 of the European Parliament and of the Council, OJ 2009 L 211/ 94. In the railway sector Directive 2012/ 34 further reinforced and elabo-rated the provisions on the regulatory bodies which the Member States need to establish. Article 55(1) now inter alia prescribes: ‘Each Member State shall establish a single national regulatory body for the railway sector […] this body shall be a stand- alone authority which is, in organisational, functional, hierarchical and decision- making terms, legally distinct and independent from any other public or private entity.’ See Directive (EU) 2012/ 34 of the European Parliament and of the Council, OJ 2012 L 343/ 32. Despite arguments to this end, no comparable provisions to the above have been included in EU harmonization directives on financial supervision. See Marc Quintyn and Michael Taylor, ‘Regulatory and Supervisory Independence and Financial Stability’, (2002) IMF Working Papers 46, pp 1– 53. The suggestion in the de Larosière report was not followed, in that report it was noted: ‘The [European System of Financial Supervision] must be independent from possible political and industry influences, at both EU and national level. […] In short, supervisory work must be independent from the political authorities, but fully accountable to them.’ See Jacques de Larosière, Report of the High- Level Group on Financial Supervision in the EU, 2009, http:// ec.europa.eu/ internal_ market/ finances/ docs/ de_ larosiere_ report_ en.pdf, p 47 (emphasis added). On the question how the independence of national and European supervisory authorities may be under-stood in the light of accountability, see Saskia Lavrijssen and Annetje Ottow, ‘Independent Supervisory Authorities: A Fragile Concept’, (2012) 39 LIEI 4, pp 419– 45.

108 See Claudio Radaelli, ‘The Public Policy of the European Union: Whither Politics of Expertise?’, (1999) 6 JEPP 5, pp 764 et seq.

109 Egeberg and Trondal, n 105, p 868.

http://ec.europa.eu/internal_market/finances/docs/de_larosiere_report_en.pdf





seen as distinct from the dynamics of agencification even if they form part of a circulatory process in which both are informed by the other.

Whether the process of agencification still supports the interests underlying the initial choice for agencification is then an open question. It seems inevitable that the process further restricts national administrative autonomy, but at the same time this is done gradually and the involvement of the national administrations themselves allows for this to be done in a context of cooperation. What seems clear nonetheless is that if agencification contributes to an executive centre for-mation, the process is not simply dictated by the Commission, and neither does it necessarily strengthen the latter. Instead, the Court’s ruling in Short- selling has made clear that the Commission may well be on the losing end in this process.

4 Identifying the Political Limits

Taking account of these dynamics and effects of agencification, the political limits to agencification may be identified. This in turn helps to understand the legal limits. After all, decision- making is a process whereby actors first bargain according to their interests (subject to political limits) and then translate this bargain into a decision following constitutional and legal rules.110 Sometimes the legal rules must thereby yield to the political compromise. The example of agencification in the EU is illustra-tive in this regard since the whole phenomenon is questionable from a legal perspec-tive (cf. Chapter IV).111 Precisely because the institutions do not seem too concerned about the legal limits to agencification, identifying the political limits becomes key.

As was noted previously, the process of agencification has analogies with comi-tology, which also developed through a continuous power struggle between the three political institutions, under the jurisprudence of the Court of Justice.112

Agencification, just like comitology, is a ‘mixed’ solution113 initially worked out without a clear legal basis in the Treaties, in the framework of a new policy field, the Common Agricultural Policy (CAP). Similarly, extensive reliance on the agency instrument coincided with the launch of the internal market, which is especially clear for the ‘internal market agencies’.114 Apart from such parallels,

110 Christopher Achen, ‘Institutional Realism and Bargaining Models’, in Thomson (ed), The European Union Decides, Cambridge, CUP, 2006, p 86.

111 Similarly, see the establishment of the EFSF and the European Stability Mechanism (ESM) in disregard of the no bailout clause of Article 125 TFEU. The Court sanctioned the ESM in the Pringle case: see Case C- 370/ 12, Pringle, ECLI:EU:C:2012:756.

112 See Carl Frederik Bergström, Comitology: Delegation of Powers in the European Union and the Committee System, Oxford, OUP, 2006; Carl Frederik Bergström and Adrienne Héritier, ‘Controlling the Implementation Powers of the Commission’, in Héritier (ed), Explaining Institutional Change in Europe, Oxford, OUP, 2007, pp 171– 227.

113 See Bergström, n 112, p 309.114 See also the EEA and the EU- OSHA. The former was established following the SEA whereby

an EEC environmental policy was identified as a necessary requirement for an internal market. According to the EU- OSHA Regulation, the agency is a measure under the action programme con-cerning the implementation of the Community Charter of Fundamental Social Rights for Workers.

Identifying the Political Limits 119

there equally exist important differences, and these may partly be explained by the institutions’ experiences following their conflict over the comitology system.

Recapitulating an important conclusion of the previous sections, the agency instrument corresponds to a functional need for more effective and efficient im-plementation of EU law as well as providing the EU with its own expertise. The agency instrument is then the preferred institutional choice for the Member States because it allows a modest elaboration of an EU administration, without directly strengthening a supranational Commission.115 For the Commission, the agency instrument allows further EU integration in areas which could otherwise be im-penetrable. The same goes for the Parliament, which acquires greater control over the implementation process compared to the typical status quo situation (national administrations being competent) even if it hypothetically loses control, since it is not the Commission which is empowered. The political limits to agencification are then a result of these interests on which the consensus between the three in-stitutions is built. Since each institution now has a veto power, agencification will not develop in ways which any of the three institutions assesses as fundamentally contrary to its own interest.

4.1 The Council

Just like comitology, agencification is a result of the Member States’ mistrust in a supranational institution and their desire to proceed with EU integration in a more intergovernmental manner. The fact that agencification is foremost tailored to the needs of the Member States has traditionally been reflected in (i) the legal basis which is used to establish an agency and (ii) the composition of the internal body directing the agency.

As was noted above (cf. section II.1) and as will be analysed from a legal per-spective in the following chapter, agencies used to be established on the single legal basis of Article 235 EEC requiring unanimity in the Council but no Parliamentary consent. Firstly, this meant that the establishment of an agency was solely an affair between the Commission and Council. Secondly, it meant that the process of agencification was controlled by the Member States rather than by the Council. The procedure therefore safeguarded both the intergovernmental character of the institutional solution and the possibility for every Member State to thwart integra-tion in those areas in which it preferred greater national autonomy. As a result, it would have been impossible to establish an EU agency with powers concurrent to those of national authorities, let alone in sensitive areas.

Every Member State having a veto power evidently puts a brake on agencifica-tion, turning the establishment of a new agency (or the revision of a mandate) into a rather cumbersome procedure. In a sense this goes against the very (functional)

The action programme aimed to develop the social dimension of the internal market following the SEA. See European Commission, COM (89) 568 final, p 3.

115 This similarly to the establishment of the comitology system. See Bergström, n 112, p 309.


rationale behind agencification, since agencies (merely) ensure a better applica-tion of EU legislation. This raises the pertinent question of why such substantive rules may be adopted using lighter procedures, such as codecision, while the institutional arrangement designed to ensure a proper implementation of these substantive rules should be established using unanimity in the Council.116 As a result, after a number of unsuccessful attempts in the 1990s the Commission suc-ceeded, from the third wave onwards, in its attempts to establish agencies without recourse to the current Article 352 TFEU.

This move was also inspired by the institutions’ experience with comitology. The latter, just like agencification originally, was an affair between the Council and the Commission, to the Parliament’s exclusion. At first this was not so much an anomaly in view of the Parliament’s generally marginal position among the EEC institutions. However, with the reinforcement of Parliament’s legislative pos-ition under the SEA and subsequent Treaties also came Parliament’s increased resistance against comitology. This resistance meant that the institutions had to spend a lot of time and energy in solving interinstitutional conflicts and also affected the efficiency of comitology itself since the Parliament went as far as freezing budgetary appropriations, preventing the committees from properly func-tioning.117 Starting from the mid- 1990s, the Parliament also used its budgetary powers vis- à- vis the agencies: before the (non- self- financed) agencies’ budgetary procedure was harmonized, the Parliament put part of the agencies’ subsidies in reserve for the years 1995 and 1996,118 to put pressure on the Council and the Commission to come to a harmonization.119

Because of the change in legal basis for establishing agencies, the institutions prevented agencification from becoming as contentious a subject as comitology. By involving the Parliament as an equal player in agencification, the individual agencies and the agency instrument itself acquired the necessary legitimacy (for the Parliament) to make it uncontested.

Evidently, the price to pay for this was the loss of veto power for individual Member States. Agencification was no longer subject to the individual Member States’ interests, but (only) to the common Member State interest. The potential

116 See the argument made by Mok, section IV 1.1.3.1.117 For a discussion of the Parliament’s difficult relationship with comitology, see Daniela

Kietz and Andreas Maurer, ‘The European Parliament in Treaty Reform: Predefining IGCs through Interinstitutional Agreements’, (2006) SWP Working Papers, pp 9– 17.

118 See lines B0- 4 0 at pp 1609– 11 and 1673– 5 respectively for the budgets of 1995 and 1996. See OJ 1994 L 369/ 1; OJ 1996 L 22/ 1. Further see point 23 of the European Parliament Resolution, OJ 1995 C 18/ 145; points 25– 7 of the European Parliament Resolution, OJ 1995 C 308/ 116.

119 This was only an indirect way of putting pressure, since it first of all hampered the agencies in their functioning. See inter alia Fernand Sauer, ‘European Medicines Evaluation Agency: Status Report’, in Kreher (ed), The New European Agencies, Florence, EUI, 1996, p 27; Peter de Rooij, ‘The European Training Foundation’, in Kreher (ed), The New European Agencies, Florence, EUI, 1996, p 46; Jean- Claude Combaldieu, ‘Histoire, experience et futur de l’Office de l’Harmonisation dans le Marché Intérieur’, in Kreher (ed), The New European Agencies, Florence, EUI, 1996, pp 49– 63. Parliament put this policy on hold when the Commission in 1997 proposed to amend the establishing regulations of these agencies: see European Commission, COM (97) 489 1- 9 final. In the end the issue went off the table when the 2002 Financial Regulation was adopted.


scope of agencification (both quantitative and qualitative) was therefore broad-ened and the move away from Article 352 TFEU as a legal basis was sanctioned by the Court in the ENISA case (cf. section IV 1.1.3.1). Whereas the move lowered the probability of interinstitutional conflict, it evidently did create a risk of con-testation by the Member States, since a minority of them could now be outvoted.

The veto power of the Council not only affects the question of which powers are granted to agencies, but evidently also has consequences for the internal func-tioning of an agency. The tasks and powers of the agencies’ Boards as well as the composition of these bodies have been set out previously (cf. section II 3.1). It was noted that as a rule, every Member State is represented on the Board. Even if this does not amount to a veto power for every Member State (given that unanimity voting is the rare exception), it is of fundamental importance for the Member States.120 This representation, in theory, allows them to keep track of the func-tioning of the agency and its planned activities.

Practice has shown that this over- representation of the Member States on the Boards of EU agencies is a political limit for the Council, even in Boards of ser-vice agencies such as the Translation Centre for the Bodies of the European Union (CdT) (discussed previously). When discussing the composition of the Boards (cf. section II 3.1.2) it was noted that the Commission has tried to introduce a parity between Commission and Council in several Boards,121 and that it at-tempted to establish this rule through the proposed interinstitutional agreement. All these attempts failed, however, and, following the Common Approach, the Commission has reconciled itself to agencies’ Boards being dominated by the Member States. The proposed parity would not only have the evident result of granting the Commission and Council equal votes on the Boards but would also have fundamentally altered the nature of the representation, as it would no longer be the Member States being represented but the Council. As the Commission noted in its draft interinstitutional agreement (IIA):

the agency’s involvement in exercising executive powers at Community level calls for the equal representation of the two branches of the Community executive within the adminis-trative board. The Commission and the Council should therefore designate an equal and limited number of members within the administrative board.122

Evidently, the Commission and the Council are only the two branches of the ex-ecutive under the system of direct administration, with the national administra-tions being the Community executive under indirect administration.

Further institutional practice has also confirmed this. In its original pro-posal on the CdT, the Commission had provided that the Board would only be

120 See Johannes Saurer, ‘Supranational Governance and Networked Accountability Structures: Member State Oversight of EU Agencies’, in Rose- Ackerman and Lindseth (eds), Comparative Administrative Law, Cheltenham, Edward Elgar, 2010, p 621.

121 For instance in the case of the EMSA (see COM (2000) 802 final, Article 11), ERA (see COM (2002) 23 final, Article 29), EIGE (see COM (2005) 81 final, Article 10), and ECDC (see COM (2003) 441 final, Article 14).

122 European Commission, COM (2005) 59 final, p 14 (emphasis added).


composed of its own representative and the representatives of the agencies using the services of the CdT.123 The Council changed the composition and provided that each Member State would also have a representative. Furthermore, the prob-lems resulting from the oversized Boards following successive enlargements have been solved by providing for Executive Boards which may take up most of the work of the Boards proper, while important matters such as the work programmes and budgets remain to be decided by the latter. The dual Boards in agencies such as the ACER and the ESAs, where Member States are only fully represented in the Boards of Supervisors or Regulators, also confirms their insistence on being in-volved in the policies of the agency. An interesting but rather exceptional further illustration of how the Member States secure their dominance in the Boards is the procedure for the adoption of the work programme in the EMSA and European Railways Agency (ERA) regulations (cf. section II.3.3.1). These prescribe that the Boards may still approve of a work programme rejected by the Commission if the Member States’ representatives do so unanimously.124 Lastly, during the second wave of agency creation the Commission proposed for three agencies that the Boards would be presided over by its own representative;125 however, as is also ap-parent from the overview provided (cf. section II 3.1.5), the Council consistently rejected these proposals.

4.1.1 ConclusionFor the Member States in Council, the link between them and the agencies is a clear red line. This link used to be safeguarded by the Member States’ represen-tation in the agencies and the unanimity requirement under Article 352 TFEU. The latter has been sacrificed to make agencification uncontested but the Member States have not allowed compromises to be made as regards their representation on the Boards. By rejecting the parity in the Board proposed by the Commission on numerous occasions, they have successfully prevented agencification from being recuperated as an instance of direct administration. A less clear red line which nonetheless emerges is the Council’s insistence on flexibility and its re-jection of the Commission and Parliament’s attempts at formalizing. Here again a parallel with the development of comitology may be noted, as for a long time the Council was not keen on delegating powers to the Commission, and when it did, before the first comitology decision of 1987, it established ad hoc comitology procedures resulting in thirty or so different procedures.126 The Member States

123 See Article 4 of the Proposal of the European Commission, OJ 1994 C 99/ 17.124 The Commission itself had proposed that the programme be drafted by the Director and

approved by the Commission.125 This was so for the ETF, EU- OSHA, and EMCDDA. See Article 5(4) of the ETF proposal

(OJ 1990 C 86/ 12); Article 8(4) of the EU- OSHA proposal (OJ 1991 C 271/ 3); Article 7(2) of the EMCDDA proposal (OJ 1992 C 43/ 2). On the other hand, this arrangement was accepted by the Council for the CdT.

126 Bergström, n 112, pp 179– 80.


through Council will therefore insist on as much flexibility as possible, since in-creased formalization decreases the attractiveness of the agency instrument. This would also explain why the Council did not want a binding IIA and why in the end the attempt to come to a horizontal framework only resulted in a non- binding Common Approach.

4.2 The Commission

The Commission’s starting position is fundamentally different from that of the Member States in Council. While the latter have secured an institutional ar-rangement outside the Treaties which is beneficial to them, the Commission is on the losing side. It will therefore draw its red lines to ensure that no more of its institutional position and powers are eroded and will attempt to gradually shift the limits away from its losing position, progressively codifying the advances it makes.

In this, however, the Commission is also involved in a delicate balancing exer-cise: since further agencification also means further integration, too much insist-ence on acquiring greater control over the process could result in no further inte-gration being achieved at all. Following Egeberg and Trondal’s argument on the central role of the Commission in the formation of an executive centre through agencification (cf. section III 3.1), another red line for the Commission would be whether its own institutional interests are directly served through agencification. However, this does not seem to be a conditio sine qua non for the Commission. Instead, the Commission’s red line simply seems to be that its direct institutional interest should not be negatively affected.

A number of limits can then be deduced. A first is monitoring the compliance of Member States with their obligations under EU law pursuant to Article 258 TFEU. While some agencies have a monitoring function inter alia by conducting inspection visits,127 the Commission will not allow (qualitative) agencification to develop in such a way that it should share its position as guardian of the Treaties with an agency.128

A second red line relates to another core task of the Commission, that is, its monopoly over the initiation of legislation. As has been noted in the overview of the powers conferred on EU agencies (cf. section I.1.2.1.3), a number of EU agencies may ‘assist’ the Commission in the preparation of new legislative pro-posals. While this ‘assistance’ might in some cases restrict the Commission in its power of legislative proposal, this competence would seem too fundamental to the

127 See eg Antje David, Inspektionen im Europäischen Verwaltungsrecht, Berlin, Duncker & Humblot, 2003, pp 185– 97; Esther Versluis, ‘Catalysts of Compliance?’, in Busuioc, Groenleer, and Trondal (eds), The Agency Phenomenon in the European Union, Manchester, Manchester University Press, 2012, pp 172– 90.

128 See also Martijn Groenleer, Michael Kaeding, and Esther Versluis, ‘Regulatory Governance through Agencies of the European Union? The Role of the European Agencies for Maritime and Aviation Safety in the Implementation of European Transport Legislation’, (2010) 17 JEPP 8, p 1227.


Commission’s institutional interest to allow any encroachment by EU agencies, even if such a scenario has not been tested yet.

This is different for the delegated and implementing acts which are reserved to the Commission under the Treaties.129 Here again one would expect the Commission to take a firm stance and not allow an encroachment by EU agen-cies on its function under Articles 290 and 291 TFEU. To a certain extent this hypothesis was confirmed when the ESAs were established. As noted previously (section I 1.2.1.3.1.f), the Commission, in a statement annexed to the ESAs’ Regulations, resisted the restrictions which those regulations imposed on its com-petences under Articles 290 and 291 TFEU.130

Three main reasons may explain why the Commission did not use its veto power here. Firstly there is the continuous balancing exercise which the Commission is engaged in. Secondly, under the pre- existing arrangements the Commission was already restricted in its freedom when adopting measures regulating the fin-ancial services sector.131 Lastly, the regulations still leave some margin to the Commission to change the ESAs’ proposals.

The situation is different however for Regulation 236/ 2012 granting a power to the ESMA to ban, or impose conditions on, shorting on the financial markets (cf. section I 1.2.1.3.2.b).132 The instruments which the ESMA would adopt pur-suant to this power would seem to meet all the requirements of an implementing act under Article 291 TFEU even if the ESMA would only exceptionally use these powers of intervention, and even if the Commission through delegated acts may further elaborate the framework in which the ESMA would exercise these powers.133 Since the Commission proposed to grant this power to the ESMA it has not challenged it itself; yet this was done by the UK. The Short- selling case will therefore be dealt with in the following chapter. The outcome of this case was almost bound to affect how the Commission will draw a red line for agencifica-tion in relation to implementing and delegated acts in the future.

As it is, the Commission suffered a serious blow in Short- selling. As Orator notes, the Commission intervened in support of the defendants, but it did not always make clear supporting arguments against the pleas advanced by the UK.134 As regards the plea on the possibility to elaborate the system as laid down by Articles 290 and 291 TFEU by granting powers to EU agencies rather than to the Commission, the latter argued that Article 290 TFEU (unlike Article 291) powers

129 With the exception of the Council adopting implementing acts in exceptional circumstances (and in the CFSP).

130 See Council Documents 15649/ 10, ADD 1, 15647/ 10 ADD 1, and 15648/ 10 ADD 1, all dated 10 November 2010.

131 See Declaration 39 annexed to the Final Act of the IGC which adopted the Treaty of Lisbon on Article 290 TFEU.

132 See Regulation (EU) 236/ 2012, OJ 2012 L 86/ 1.133 Evidently, because of ESMA’s general powers under the ESMA regulation, such delegated

acts would also be drafted by the ESMA itself.134 Andreas Orator, ‘Die unionsrechtliche Zulässigkeit von Eingriffsbefugnissen der ESMA

im Bereich von Leerverkäufen’, (2013) 24 EuZW 22, p 855 at note 44.


could not be granted to agencies and should remain an exclusive Commission com-petence.135 As will be discussed later, AG Jääskinen followed the Commission’s lead in his Opinion but the Court found that Articles 290 and 291 TFEU were both part of an open system, meaning Short- selling weakened the Commission in its relationship with the legislator.

As regards the Commission’s powers in the EU’s competition policy, it may be noted that the Commission successfully defended these against attempts to push agencification also in this field.136 Based on further core tasks of the Commission, one may imagine another red line even if no problems have arisen yet on the issue. This red line would concern the EU’s external relations and the representation of the EU by the Commission as foreseen in Article 17 TEU. As noted previously (section II 3.4.3), possible bureaucratic competition from the agencies in this area was also addressed in the Common Approach.

Even if the Common Approach is non- binding, this is a good example of the Commission trying to secure certain advances by codifying and formalizing them. Another example is the early warning system provided for by the Common Approach. In itself it is nothing new, since the Commission’s representatives on the Boards could anyway ask a Board to refrain from taking a decision and, with or without the Common Approach, that Board is under no obligation to do so.137 The provision in the Common Approach does however introduce the idea that it is the Commission which provides the link between the agency and the Council and Parliament.

On other issues the Commission has been less successful. As regards the ap-pointment of the Director, for instance, the Common Approach recognizes that the Commission will propose candidates to the Board, but it equally provides that ‘[e] xceptions to this approach can be foreseen if justified in specific cases’. As regards the work programme, the Common Approach only provides that the Commission should issue a formal opinion, not that the Commission should agree with the draft programme or that a higher majority threshold is needed to adopt a work programme on which the Commission has issued a negative opinion. At the same time, the existing institutional practice and its codification through the Common Approach now form a rather solid basis for the Commission to present its involvement in the adoption of the work programme and nomination of the Director as an absolute minimum requirement, which can still be elaborated upon.

4.2.1 ConclusionUnlike the Council, the Commission finds itself in a defensive position. Since agen-cification puts its executive position as defined by the Treaties under pressure, the

135 Case C- 270/ 12, UK v. Parliament and Council, ECLI:EU:C:2014:18, para 76.136 See Chapter I, n 171.137 Unless this is explicitly provided in the agency’s establishing regulation or in the Rules of

Procedure of the Board.


Commission will only agree to further agencification if its core tasks are kept intact, performing a difficult balancing exercise between furthering EU integration and safe-guarding its own institutional interests. To recuperate lost terrain, it will attempt to advance new practices which are more congenial to its interests and will subsequently try to codify them. Following Short- selling, however, the Commission suffered a major setback (cf. section IV 7.2.4.1). As will be elaborated, Short- selling requires the institutions and ultimately the primary legislator to address the question of the agencies’ position in the EU’s institutional setup. Specifically for the Commission, it should provoke a more considerate reflection on a shared understanding among the different DGs to come to a clearer strategic position in the process of agencification.

4.3 The European Parliament

Based on its experience with the comitology system, one would expect that the Parliament has also taken a critical stance vis- à- vis agencification. Indeed, in 1961, when the CAP was being set up, the Parliament, alarmed by the Council’s in-tention to diminish the Commission’s executive role by setting up a committee, adopted a resolution in which it expressed its resistance against any (indirect) weakening of the Commission, the Commission being the only one account-able before the Parliament.138 Later, in 1967, the Parliament remarked that the Commission and Council should not establish new committees and asked its own legal committee to produce a report.139

The latter report came to be known as the Jozeau- Marigné Report of 1968,140 which gave an elaborate overview of the then existing procedures to implement sec-ondary legislation. In its subsequent resolution, the Parliament noted that Article 155 EEC provided for competences for the Council and the Commission and that as a result ‘these two institutions should exercise these competences, without a body not provided for by the Treaty violating these competences’.141

In a number of resolutions from the 1970s, Parliament was already showing its more pragmatic side,142 recognizing that the functions to be performed could justify the establishment of a decentralized institution but remaining at the same time very sceptical because of the effects such ‘decentralization’ would have on its powers of control vis- à- vis the Commission and through the budget.143 As a result, during the second wave of agency creation the Parliament went as far as confirming

138 See Résolution de l’Assemblée parlementaire, JO 1962 72/ 62.139 Résolution du Parlement européen, JO 1967 268/ 20.140 Rapporteur Jozeau- Marigné, Verslag nopens de communautaire procedures ter uitvoering van het

afgeleide gemeenschapsrecht, Zittingsdocumenten van het Europees Parlement, Document 115/ 68.141 See Résolution du Parlement européen, JO 1968 C 108/ 37.142 See Resolution of the European Parliament, OJ 1977 C 183/ 48; Resolution of the European

Parliament, OJ 1977 C 299/ 63.143 Even if the Parliament became a half- hearted supporter of agencification during the second

wave of agency creation, it kept underlining the undemocratic character of agencies, stressing the analogy with comitology committees: ‘democratic control over bodies (such as […] decentralized EU agencies) […], is often only indirect or even non- existent.’ See Resolution of the European Parliament, OJ 1997 C 200/ 153.


its continued support for the specialized agencies of the European Union; [it] recalls, how-ever, that all the agencies, and in particular the newly established ones, have to resp ect the general constraints to the budget and the principles that govern the European institu-tions; [it] reaffirms that the agencies may expand their activities and budgets only after demand has been demonstrated.144

This position was also already apparent in 1993 when the Parliament adopted two separate resolutions on agencification. The first was devoted to the budgetary (control) implications of agencification, in which the Parliament urged a harmonization.145 In the second resolution,146 the Parliament com-mented on the general process of agencification, raising fundamental issues which are still relevant today. From a political perspective, the Parliament again expressed its principled support for agencification in the light of the new policy areas unlocked by the SEA and foreseeing an acceleration in the process with the advent of the Maastricht Treaty. It then deplored the lack of a frame-work and the ad hoc way in which the agencies had so far been established, demanding that the Commission draft ‘a framework- statute to regulate the specialist bodies and agencies’.147 Strategically, it referred to the Meroni ruling (cf. section IV 5) to establish the Commission’s responsibility (and therefore indirectly its own control) for the agencies and demanded a representative of Parliament on each agency Board. It made the establishment of new agencies subject to a number of other requirements148 and undertook not to approve of new agencies if these requirements were not met.

The reference to Meroni would have meant that the Parliament erroneously held the Commission to be the delegating authority (cf. section IV 5.7.2.2) but, interestingly, the Parliament further noted that it supported ‘the constitutional principle whereby the decision to delegate tasks falls upon the executive power; nevertheless [it] takes account of the fact that at the present stage of European construction this decision is taken by the Council and consequently asks to be fully associated with the process’.149 The Parliament’s strategic behaviour in the following years was very much inspired by this resolution, as is apparent from the amendments which it filed on several proposals for agency creation.

During the second wave of agencification, the Parliament combined its mod-erate support for agencification with attempts to acquire a say in the agencies’ functioning. As regards the EEA, for instance, the Parliament proposed that the Director would be appointed following a proposal by ‘the Commission with the collaboration of the European Parliament’; that the Parliament could decide to replace the Director for important reasons; and that the Parliament would send

144 Resolution of the European Parliament, OJ 1995 C 308/ 116.145 Resolution of the European Parliament, OJ 1993 C 42/ 61.146 Resolution, OJ 1993 C 42/ 63. The report on which the resolution was based deals with these

issues in more depth. See Rapporteur Boissière, Report on the Setting Up and Operation of Specialized Bodies and Agencies, 8 December 1992, European Parliament Session Documents, PE 201.993/ fin.

147 See point 15 of the Resolution. 148 See point 11 of the Resolution.149 See point 9 of the Resolution.


four representatives to the EEA Board.150 In addition, the Parliament also pro-posed that the EEA should be granted further powers and insisted on a study on the feasibility of inter alia granting inspection powers to the EEA.151 In the end, however, the Parliament only succeeded in securing the right to send two scientific personalities as representatives to the Board. What is more, because the (European) Council could not decide on a seat for the EEA, it took another three years before the agency started working, despite numerous Parliamentary Questions and crit-ical Parliamentary Resolutions.152

During the negotiations on the EMA, the Parliament again proposed that it should approve the Director and that it would have an equal number of repre-sentatives on the Board to that of the Commission.153 Here the Parliament again only succeeded in getting its representatives on the Board, and its position was further undermined by the Council, which changed the legal basis from Article 100a EEC to Article 235 EEC. With this, the cooperation procedure changed to the advisory procedure.154

The same happened with the proposal for the Community Plant Variety Office (CPVO), for which the Council changed the legal basis from Article 43 EEC to Article 235 EEC, even if this did not affect the Parliament’s position. The latter however had proposed to change the legal basis to Article 100a EEC, and again proposed that it would both be involved in the appointment of the CPVO’s President and have a representative on the Board.155 Ultimately the Parliament failed to push through any of its amendments.

For the EU- OSHA, the Parliament proposed inscribing a review clause with the prospect of granting further powers to the agency. It also suggested that the relevant Parliamentary committee would have to agree with the appointment of the Director and that the Commission would submit an annual report to the Parliament on the work of the agency. Finally it suggested that the draft work programme would have to be approved by the Commission and the relevant Parliamentary committee.156 However, only the amendment on the review clause made it to the final regulation.

In the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) proposal the Commission itself foresaw a representation of the Parliament on the Board,157 and the Parliament further suggested that it should give its opinion on the multiannual work programme, that the annual report of the agency would be submitted to it for approval, and that the Director would report twice a year to

150 See Amendments 48, 52, and 53 of the European Parliament, OJ 1990 C 68/ 50.151 See Compromise Amendment 3 of the European Parliament, OJ 1990 C 96/ 112.152 See Resolution of the European Parliament, OJ 1990 C 175/ 231; Resolution of the European

Parliament, OJ 1992 C 241/ 226; Resolution of the European Parliament, OJ 1993 C 150/ 291.153 See Amendments 85, 88, 168 of the European Parliament, OJ 1991 C 183/ 145.154 See Legislative Resolution of the European Parliament, OJ 1993 C 176/ 127.155 See Amendments 1, 13, and 14 of the European Parliament, OJ 1992 C 67/ 148. See also

Legislative Resolution of the European Parliament, OJ 1993 C 305/ 56.156 See Amendments 24, 27, 29, 30, 35 of the European Parliament, OJ 1994 C 128/ 503.157 European Commission, COM (91) 463 final.


its relevant committee.158 Again, none of these amendments made it to the final regulation.

The Parliament’s amendments to the CdT proposal are different from those to the other agency proposals, since it suggested that the CdT should be established not as an agency but as a distinct body within the Commission,159 similar to Eurostat. Evidently this meant the Parliament did not put forward its otherwise typical amendments.

The EFSA then was the first agency established through the codecision pro-cedure, which immediately translated Parliament’s reinforced formal position into the text of the final regulation. In its amendments the Parliament had suggested that the candidates for the post of Director would be invited to a Parliamentary hearing; that the work programme would be drafted in consultation with the Parliament; that the Director maintain a regular dialogue with the relevant com-mittees; that he would present the annual report and (multi- )annual work pro-gramme in the Parliament; and that it would be able to take part in the work of the EFSA Advisory Forum.160 In the final text, the candidate selected by the Board is obliged to make a statement before and answer questions from the Parliament;161 the Director may invite representatives from the Parliament to the Advisory Forum; and the Director develops and maintains contact with the Parliament, further en-suring a regular dialogue with its relevant committees. In addition, the Parliament’s role in the appointment of the Board members was also strengthened.162

In light of the discussion on the agencies’ statutes (cf. section II 3) and the Common Approach, the EFSA case shows that the Parliament has succeeded, from the third wave onwards, in including a number of typical provisions in the establishing regulations which it failed to have included for the agencies of the second wave. Apart from promoting these provisions, the Parliament also insisted on, and secured, the agencies’ incorporation in the budgetary procedure.163 In addition, as was to be expected, the Parliament’s position also shifted as a result of its changed role in the legislative procedure. Whereas in the 1999 Herman report, just before the third wave, it still took on a rather principled approach— similar to the one in its 1993 resolutions— it substituted its principles for pragmatism once it became co- legislator, even if it kept its critical discourse.

The Herman report suggested different ways to improve the institutional func-tioning of the EU without recourse to a Treaty revision. Part of the report therefore also focused on the EU agencies. The Parliament urged a more thorough debate in the future when new proposals are discussed and again called for a ‘coherent overall institutional framework’ for the agencies.164 During the Convention on the Future

158 See Amendments 20, 21, and 23 of the European Parliament, OJ 1992 C 150/ 48.159 See Amendment 2 of the European Parliament, OJ 1994 C 341/ 44.160 See Amendments 107, 108, 109, 111, 114, and 220/ rev of the European Parliament, OJ 2002

C 53E/ 47.161 See also Amendment 17 of the European Parliament, OJ 2002 C177E/ 65.162 See ibid, Amendment 37. 163 See n 119.164 See Chapter II, n 57.


of Europe, the Parliament adopted a resolution on the hierarchy of norms in the EU legal order. In its resolution it proposed that EU agencies would be able to adopt what are now called implementing acts, with a scrutiny mechanism which would have allowed the two arms of the legislature and the Commission to repeal such acts.165 This again shows the analogies between the Parliament’s positions vis- à- vis comitology and agencification (cf. section III 4.1). The Parliament in general kept the pressure on the Commission (and the Council), at least in its resolutions, urging a fundamental rationalization of the agencification process. The Parliament continued to insist on the Commission’s responsibility for the agencies, which would allow it indirect control of agencies, and also proposed to renounce its rep-resentation on the Boards if the EFSA model was generalized.166

During the third and fourth waves, however, Parliament agreed to establish new agencies as co- legislator, despite the fact that a framework was still lacking. Because of this, the Parliament’s criticism, in its resolution following the failed IIA project, on the lack of a framework while quantitative agencification contin-ued167 seems out of place.

Further, despite the fact that the Parliament had referred to its role as budgetary authority, it has only recently re- started using its budgetary powers to coerce some of the agencies to adapt their functioning even if this seems to be incited foremost by the positions taken by individual rapporteurs in the budgetary con-trol committee rather than by a clear policy formulated by the Parliament as a body.168

4.3.1 ConclusionOf the three political institutions, the Parliament probably has the least clear position vis- à- vis agencies. Early on in the process it was sceptical, since (i) it was not really involved in the legislative process, (ii) agencies undermined its scrutiny powers, and (iii) for a long time agencies frustrated its budgetary powers. Today agencies are generally established following codecision and most EU agencies, except the wholly self- financed, come under a harmonized budgetary procedure. As a result, Parliament has become remarkably pragmatic in agencification. The only problematic issue for the Parliament is that its control powers over EU agen-cies are not horizontally defined. If the Parliament could remedy this, it would have even more reasons to support agencification, since agencies are controllable EU actors and Parliament would be confronted with mere subsidiary organs in its confrontation with the executive rather than with an institution proper. However, the Parliament’s internal organization has so far seemed to prevent it taking up a clear and consistent position in agencification.

165 See point 17 of the Resolution of the European Parliament, OJ 2004 C 31E/ 126.166 See point 17 of the Resolution of the European Parliament, OJ 2004 C 92E/ 119.167 See point 5 of the Resolution of the European Parliament, OJ 2010 C 15E/ 27.168 The assertive stance of MEP Macovei in the discharge procedure for the budgetary year 2010

is exemplary in this regard. See Chapter V, n 107.


4.4 The role of the Court of Justice

Unlike the political institutions, the Court has not had so much involvement in agencification. This is an important difference between agencification and comi-tology, since the latter has been influenced much more by the Court’s case law. On this jurisprudence Bergström notes that the Court ‘has been far from active: not only has it proved itself unwilling to reduce the room for political negotiations but it has also demonstrated a surprising ability to adapt itself to their result’.169 This observation also seems to capture the attitude of the Court in agencification.

In its agency- related jurisprudence before the Short- selling case, the Court could have taken a more activist approach, but instead it left a number of under-lying contentious issues untouched.170 And when the Court sanctioned the use of Article 114 TFEU in ENISA this elicited some criticism, since several authors questioned whether the decision to set up an agency could really be qualified as ‘measures for approximation’.171 Should the Court be reproached for its passive stance pre- Short- selling? Although a more active approach would have been inter-esting for academic purposes, this does seem doubtful given the broad consensus among the political actors involved.

Unsurprising in light of the Court’s jurisprudence on comitology, but remark-able nonetheless, was the fact that the Court continued its passive stance on agencification in Short- selling. Although the Court might have at least taken a firm stance on the plea related to Articles 290 and 291 TFEU,172 it totally def-erred to the legislator. The Court’s ruling in Short- selling has resulted in two major victors. The first was the ESMA, which saw its powers sanctioned; but from a more general perspective the ruling also bolstered the position of the legislator (Parliament and Council) at the expense of the Commission in its ex-ecutive capacity.173


The political limits to agencification are determined by the greatest common div-isor of the different interests of the institutional veto players. However, from the legislative practice it appears that the Council and the Commission have identi-fied their interest in a clearer manner than the Parliament has.

For the Council, the link between an individual agency and the individual Member States is of primary importance. For the process of agencification this translates into the Council’s rejection of the process being an instance of direct administration. The link between agencies and Member States is then exemplified

169 Bergström, n 112, p 319.170 This was certainly the case in ENISA, but also in minor cases such as Thomae v. Commission,

FNAB and others v. Council, etc (cf. Chapter IV, n 761 and Chapter V, n 159).171 See Vetter and Ohler specifically, Chapter IV, n 63.172 Merijn Chamon, ‘Le recours à la soft law comme moyen d’éluder les obstacles constitution-

nels au développement des agences de l’UE’, (2014) RUE 576, p 159.173 Chamon, n 102, pp 139– 40.


in the Council’s insistence on a full and equal involvement of Member States’ representatives in the policy body of an agency.

The Commission is involved in a more delicate balancing exercise, blur-ring its red lines. The delicate balance is a direct result of the essence of agen-cification itself: continued integration without strengthening the supranational Commission. The Commission will therefore continuously compromise. Of course, some issues are regarded as so fundamental to its institutional interest that the Commission will not compromise on them. While the adoption of delegated and implementing acts would seem to be such a core issue, the Commission has also made significant concessions in this area.

The Parliament then seems the most pragmatic institution. Initially it took on a principled approach. However, even before it became a co- legislator in agenci-fication it took on a more nuanced position, which became even more pragmatic from the third wave onwards. Even if the Parliament remains critical vis- à- vis agencification and individual agencies from time to time in its (discharge) resolu-tions, this depends more on the individual rapporteurs than on a clear position of the Parliament as an institution. Since the agencies were included in the general budgetary procedure, no further red lines seem to have been formulated by the Parliament, even if it has every reason to emphasize its non- budgetary control powers to a greater extent.

The Court has taken on a passive approach, much like it has in the interinstitutional power struggle over comitology. The Court’s ruling in Short- selling confirmed this and further shifted the balance to the disadvantage of the Commission as the EU’s executive actor.

5 Conclusion on the Political Limits

While EU agencies may fulfil functional needs, their establishment and setup cannot simply be explained through a functional account. Essentially, the instru-ment of the EU agency is the result of a conscious political choice for a specific institutional form allowing the EU to build and strengthen its administrative capacity, without relying on traditional direct administration and while retaining national autonomy as much as possible.

Agencification being a political choice, the question on the political limits to the process reveals itself as important. While initially these political limits are informed by rational considerations based on the relevant actors’ interests, they are subsequently determined not merely by a logic of consequentialism, but also by a logic of appropriateness and path dependency. As a result, the effects of agen-cification are not necessarily in line with the original motives driving this pro-cess. Still, caution is warranted when identifying trends describing the effects of agencification: building administrative capacity at EU level cannot automatically be equated with executive centre formation or the wholesale transfer of adminis-trative competence to the EU.

Conclusion on the Political Limits 133

After all, the relevant actors themselves, that is, the institutions, do not share a common understanding on the finality of agencification. Instead, each time an agency is established or an agency mandate is reviewed, agencification is shaped by recurring political negotiations characterized by pragmatism and a set of red lines on which an institution will not compromise.

IVThe Legal Limits to Agencification

The present chapter aims to answer the following basic questions:

(i) Does the EU have a competence to establish subsidiary bodies (principle of conferral)?a. If so, what is the proper legal basis to do so?b. If so, which legal instrument can be used to do so?

(ii) If the EU has such a competence, when and how can it exercise this com-petence (principles of subsidiarity and proportionality)?

(iii) How does the empowerment of EU agencies fit into the existing typology of delegations in the EU legal order?

(iv) Which rules apply to the empowerment of EU agencies?a. Can the Meroni and/ or Romano jurisprudence be applied?b. Is the institutional balance a limit to such empowerments?

The search for the legal limits to agencification starts by verifying the existence of any relevant rules in the Treaties. As was already noted previously, the latter do not contain a general enabling clause explicitly conferring power on the EU to establish or empower an agency. This is for instance in contrast to the EFTA Convention1 or the UN Charter, which, in its Article 7(2), provides: ‘Such subsidiary organs as may be found necessary may be established in accordance with the present Charter.’ Meier has described this clause as an allgemeine Organisationskompetenz restricted by the provisions in Articles 22 and 29 of the Charter which provide that it is up to the General Assembly or the Security Council to establish these bodies.2 Since EU primary law lacks such provisions, the possibility to supple-ment the EU’s institutional system has been analysed extensively in the academic debate, even if the focus has been more on the powers which may be conferred on subsidiary bodies than on the possibility of establishing these bodies in the first place.

The Treaties as they stand do contain references to agencies. The Treaty of Lisbon amended a number of provisions on the institutions, broadening their

1 See Article 43(3) of the EFTA Convention.2 Gert Meier, ‘Das Recht internationaler Organizationen zur Schaffung und Bevollmächtigung

eigener Organe’, (1964) 12 AVR 1, pp 22– 3.

The Legal Limits to Agencification 135

scope to both the institutions and the ‘bodies, offices and/ or agencies’3 of the Union, even if not every language version refers to ‘agencies’.4 In any event, these amendments brought by the Treaty of Lisbon have not fundamentally affected the issue, since even beforehand, references to agency- like bodies could be found in the Treaties.

A first could be found, as early as 1957, in the Article on the preliminary ruling procedure, which inter alia referred to ‘bodies established by an act of the Council’.5 In 1975, the Treaty on financial provisions establishing the Court of Auditors6 provided that Court would also ‘examine the accounts […] of all bodies set up by the Community […]’. Lastly, the Amsterdam Treaty introduced Article 213b EC (later Article 286 EC, now Article 16 TFEU), which provided that rules ‘on the protection of individuals with regard to the processing of personal data and the free movement of such data shall apply to the institutions and bodies set up by, or on the basis of, this Treaty’.

These provisions suggest that the Treaty authors foresaw that subsidiary bodies could be established. On the other hand, up until the entry into force of the Treaty of Lisbon, the original Article 4 EEC contained the list of the EU’s institutions, preceded by the following statement: ‘The tasks entrusted to the Community shall be carried out by the following institutions […]’. A strict reading of this provision would then suggest that the EU institutions have to make do with the EU’s institutional machinery as provided for by the Treaties, precluding the possibility to entrust the execution of the EU’s tasks to bodies other than the institutions. However, such a (potentially) prohibitive reading is no longer possible following the Treaty of Lisbon’s rewording of Article 13 TEU.

This raises the question of whether the Treaties now contain a recognition of the competence for the EU to establish and empower EU agencies. Evidently the starting point to answer this question is the principle of conferred powers which governs the competences of the EU.

3 See Article 9 TEU and Articles 15, 16, 71, 123, 124, 127, 130, 228, 263, 265, 267, 282, 287, 298, 325 TFEU as well as Articles 24, 40, 42 and Article 1 of Annex 1 of the Statute of the Court of Justice, Articles 4, 5, 7, 21 of Protocol No 4, Protocol No 6, Article 9 of Protocol No 36 and Articles 41, 42, 51, 52 of the Charter of Fundamental Rights. Michel finds that these references anchor the agencies in primary law: see Katja Michel, Institutionelles Gleichgewicht und EU- Agenturen— Eine Analyse unter besonderer Berücksichtigung der European Banking Authority, Berlin, Duncker & Humblot, 2015, p 112. However, it is argued here that being mentioned in the Treaties is something different from being provided for or anchored in the Treaties.

4 Apart from the English- language version, the Bulgarian, Danish, Estonian, Greek, Gaelic, Slovenian, Slovak, Romanian, and Swedish- language versions refer to ‘agencies’. The other language versions, inter alia the German, Polish, French, Spanish, and Italian, only refer to the institutions and ‘einrichtungen und sonstigen Stellen’, ‘organy i jednostki organizacyjne’, ‘organes et organismes’, ‘órganos y organismos’, ‘organi e organismi’.

5 Article 177 first sentence (c) EEC, later Article 234 first sentence (c) EC. This provision was deleted by the Lisbon Treaty.

6 See OJ 1977 L 359/ 1.

The Legal Limits to Agencification136

1 Limits Flowing from the Principle of Conferred Powers

The principle of conferred or attributed powers is laid down in Article 5 TEU and makes clear that the EU’s competences are limited and that it only has those compe-tences explicitly or implicitly conferred on it. Fischer- Appelt notes that the question on the EU competence to establish an agency breaks down into three legally distinct questions. Thus, (i) the competence to establish an agency should be proven and (ii) the agency’s powers should in the first place have been conferred on the EU.7 These two elements then need to be linked by a third which allows (iii) the grant of competences to the agency.8 Since the last question relates to the horizontal rather than the vertical distribution of powers, it will be dealt with separately later.

Of the first two elements, the first is the most problematic. The second seems much less contentious since the powers granted to EU agencies are typically dep-endent upon an earlier exercise of competence by the EU, that is, the EU legislator adopting new legislation which is then implemented. Normally, the question as regards the material competence will therefore have been determined before the agency is established. On a second level there is the question of whether the agen-cies’ executive tasks and powers fall under competences which have been conferred on the EU. In her study on this topic, Schreiber concluded that the Treaty provi-sions on the EU’s competences are worded broadly enough to encompass not only the power to enact legislation, but also the power to implement such legislation, even if the latter is more confined since it is in principle up to the Member States to take care of the implementation of EU law. An EU power in this respect is then only permissible when the realization of the Treaty objectives so requires, and this should be proven every time the EU takes on an executive task.9 This further relates to the principles of subsidiarity and proportionality, which will also be discussed in a following section.

1.1 The competence to establish a new body within the EU’s institutional architecture

Do the institutions have the power under the Treaties to supplement the EU’s institutional architecture with bodies meeting the definition of EU agency? The first question here is whether the Treaties lay down an exhaustive list of EU insti-tutions and bodies or whether a ‘tertiary organizational structure’ for the EU may be developed.10 Most EU agencies should indeed be situated in this third layer,

7 On this second element, see also Rainer Vetter, ‘Die Kompetenzen der Gemeinschaft zur Gründung von unabhängigen europäischen Agenturen’, (2005) 58 DÖV 17, p 723.

8 Dorothee Fischer- Appelt, Agenturen der Europäischen Gemeinschaft, Berlin, Duncker & Humblot, 1999, p 83.

9 Stefanie Schreiber, Verwaltungskompetenzen der Europäischen Gemeinschaft, Baden- Baden, Nomos, 1997, Vol 196, p 185.

10 See Meinhard Hilf, Die Organisationsstruktur der Europäischen Gemeinschaften, Berlin, Springer Verlag, 1982, p 132.

Limits under the Principle of Conferral 137

as only a handful, for example the future European Public Prosecutor Office, are explicitly foreseen in the Treaties (cf. section I 1.1.3). The latter and the EU’s insti-tutions’ competence to establish them are of course uncontested.11

This is different for the other agencies. The internal power of organization which comes to each of the institutions results in a great number of bodies making up the EU’s tertiary organizational structure,12 but this internal power may not create ‘external’ legal effects.13 This was also concluded by Everling, who further noted that the Council cannot establish such bodies either, since it is not empowered to intervene in the institutional system.14 In other early writings only the powers exercised by subsidiary bodies were found to be problematic, rather than their establishment as such.15 In any case, opinions in legal doctrine seem to have evolved in line with institutional practice, leading Nettesheim to conclude that it is ‘beyond doubt that the EU has the competence, in the framework of the limits set by the Treaties, to alter the institutional structure of the union of integration’.16

Yet, even if this competence is accepted in doctrine, the fundamental question of where the EU’s competence in this matter resides is still open. Because agencies have their own separate legal personality, recourse to formal legislation becomes necessary.17 Given the lack of an explicit provision, three types of legal bases have been used to supplement the EU’s institutional architecture: the flexibility clauses of Articles 352 and 114 TFEU and the specific legal bases.

1.1.1 Article 352 TFEU as a legal base for agency creationEarly on in the integration process recourse to Article 352 TFEU to establish agencies was ruled out, inter alia because ‘appropriate measures’ was found to refer only to material rules.18 However, these conclusions were superseded in

11 However, their functioning and specific powers may well be contested. On the Public Prosecutor, see for instance Merijn Chamon, ‘Eerste toepassingen van de subsidiariteitstoets’, (2014) 62 SEW 6, pp 278– 81.

12 See for instance Case 46/ 72, De Greef v. Commission, [1973] ECR 543, para 14.13 Michael Berger, Vertraglich nicht vorgesehene Einrichtungen des Gemeinschaftsrechts mit ei-

gener Rechtspersönlichkeit, Baden- Baden, Nomos, 1999, Vol. 219, p 61.14 Ulrich Everling, ‘Zur Errichtung nachgeordneter Behörden der Kommission der

Europäischen Wirtschaftsgemeinschaft’, in Hallstein and Schlochauer (eds), Zur Integration Europas, Karlsruhe, Müller, 1965, p 42.

15 Hubert Ehring, ‘Art. 235’, in Von der Groeben, von Boeckh, Thiesing, and von Arnim (eds), Kommentar zum EWG- Vertrag, Baden- Baden, Nomos, 1974, Vol. 2, p 790; Charles- Albert Morand, La législation dans les communautés européennes, Paris, Librairie Générale de Droit et de Jurisprudence, 1968, p 211; Richard Lauwaars, ‘Auxiliary Organs and Agencies in the E.E.C.’, (1979) 16 CMLRev 3, p 371.

16 Martin Nettesheim, ‘Kompetenzen’, in von Bogdandy (ed), Europäisches Verfassungsrecht, Berlin, Springer, 2004, p 463. Translated from German by the author.

17 See Christoph Ohler, ‘Rechtmäßige Errichtung der Gemeinschaftsagentur ENISA’, (2006) 17 EuZW 12, p 373.

18 See Everling, n 14; Peter Schindler, Delegation von Zuständigkeiten in der Europäischen Gemeinschaft, Baden- Baden, Nomos Verlagsgesellschaft, 1972, p 76. On the other hand, Reuter noted that ‘Il est difficilement contestable que [l’article 235 CEE] peut servir de fondement à l’ institution de tout “organe subsidiaire” dont le développement de la Communauté Economique ferait apparaître


the next decade, since the early agencies were all established based on Article 352 TFEU. The difficulty here is that it seems impossible to make a distinc-tion between the formal arrangements (ie the establishing of an agency) and the material rules (ie the tasks of the agency) when verifying whether the creation of an agency is an ‘appropriate measure’ under Article 352 TFEU. Just before the first wave of agency creation, Pescatore indeed remarked that the Council had a complete liberty in its choice for the measures to be adopted, adding that ‘[r] ien n’empêcherait le Conseil de prendre même des dispositions de caractère organique, en vue de créer des organismes nouveaux dans le cadre de la structure institutionnelle’.19

Under its old wording (Article 235 EEC and 308 EC), reference was made to action necessary to attain one of the objectives of the Community in the course of the operation of the Common Market. Dashwood criticized the generous inter-pretation of this enabling clause as being at odds with the principle of conferred powers and suggested that, following Maastricht, the institutions should exercise greater restraint in relying on this legal basis. According to Dashwood the key to such restraint was to take the reference to ‘in the course of the operation of the Common Market’ seriously; he further suggested that this requirement could be interpreted as having the ‘objective [of] directly affecting the establishment or functioning of the common market’.20

The analysis of an agency’s legal base would then become two- stepped, since the legislator would have to show first that the agency’s tasks are a necessary and appropriate means and directly affect the establishment or functioning of the Common Market. Secondly, it should be shown that to properly exercise these tasks it is necessary and appropriate to establish a new EU body. Under such a strict reading doubt may be cast on whether some agencies with ‘modest’ tasks meet the requirements of the flexibility clause. However, in any event, despite Dashwood’s appeals, the flexibility clause continues to confer a wide and ill- defined power with limited significant limits.21

While it is difficult to distinguish the issues of agency establishment and agency empowerment under Article 352 TFEU, there is also the difference between the issues of agency and agencification. As regards Article 352 TFEU, Vetter recalls the Court’s observations in Opinion 2/94 to the effect that ‘Article [352] cannot be used as a basis for the adoption of provisions whose effect would, in substance, be to amend the Treaty without following the procedure which it provides for that purpose’.22 According to Vetter, should the EU legislator establish an increasing

l’opportunité.’ See Paul Reuter, ‘Les organes subsidiaires dans la Communauté Economique Européenne’, (1962) 8 AFDI 8, p 780.

19 Pierre Pescatore, L’ordre juridique des Communautés européennes: étude des sources du droit communautaire, Liège, Presses Universitaires de Liège, 1975, p 137.

20 Alan Dashwood, ‘The Limits of European Community Powers’, (1996) 21 ELRev 2, p 123.21 Trevor Hartley, The Foundations of European Community Law, Oxford, OUP, 2007, pp

109– 10.22 Opinion 2/ 94, re Accession by the Community to the ECHR, [1996] ECR I- 1759, para 30.


number of agencies based on Article 352 TFEU, this could fundamentally alter the division of competences between the Member States and the EU in the ad-ministration of EU law.23 Whereas the establishment of a single agency would not be problematic per se, a process of agencification could be. Today, one can qualify the EU agencies no longer as a curiosité institutionnelle,24 but indeed as a new method of administration which does not find a basis in the Treaties.

1.1.2 Specific sectoral legal bases for agency creationFollowing the use of Article 352 TFEU as a legal basis for agencification, a number of authors excluded the possibility of recourse to any other legal basis to estab-lish agencies. Lauwaars for instance held Article 352 TFEU to be indispensable if legal personality is granted to a newly created body.25 Following the entry into force of the Lisbon Treaty, such an argument could be revived. After all, it may appear awkward that the Lisbon Treaty brought a simplification by merging the EC and EU into a single legal personality, while the legislator could at the same time partially undo this simplification by establishing a plethora of new bodies with separate legal personalities resulting in the increasing organizational plur-alism noted by Ruffert (cf. section I 1.1.4). Hilf left open the question on legal personality but opined that Article 352 TFEU is in any case necessary for the con-ferral of powers to these bodies.26 According to Priebe, the legal bases for the EU’s sectoral policies are insufficient because they lack ‘clues that the Treaty authors wished to grant the power to adopt organisational arrangements by conferring material competences’.27 Sacchettini on the other hand noted that recourse to Article 352 TFEU was merely required because the substantive provisions of the establishing regulations were not covered by other Treaty provisions.28 From a legal perspective, this was confirmed by the opinions of the Council Legal Service (CLS) during the second wave of agency creation, when the Council almost con-sistently changed the legal bases proposed by the Commission to the legal basis of Article 352 TFEU (cf. section IV 1.1.4). Yet, it needs no explaining that, from a political perspective, Article 352 TFEU was also of interest to the Member States to remain in full control of agencification (cf. section III 4.1).

Most of these observations on Article 352 TFEU were then again largely superseded by institutional praxis, since the legislator started to rely on the specific legal bases to establish agencies from the third wave onwards. The

23 Vetter, n 7, pp 730– 1.24 Merijn Chamon, ‘Le recours à la soft law comme moyen d’éluder les obstacles constitution-

nels au développement des agences de l’UE’, (2014) RUE 576, p 153.25 Lauwaars, n 15, p 376.26 Meinhard Hilf, ‘Die abhängige Juristische Person des Europäischen Gemeinschaftsrechts’,

(1976) 36 ZaöRV 2, p 559.27 Reinhard Priebe, Entscheidungsbefugnisse vertragsfremder Einrichtungen im Europäischen

Gemeinschaftsrecht, Baden- Baden, Nomos, 1979, p 81. Translated from German by the author.28 Antonio Sacchettini, ‘Les Agences Européennes: Perspective Juridique du Conseil’, in

Kreher (ed), The New European Agencies, Florence, EUI, 1996, p 70.


European Aviation Safety Agency (EASA) for instance has been established based on Article 80(2) EC (current Article 100(2) TFEU), which provides that the legislator ‘may lay down appropriate provisions for sea and air transport’. For the European Environment Agency (EEA) the legal basis was Article 130s EEC (current Article 192 TFEU), which provides that the legislator ‘shall decide what action is to be taken by the Union in order to achieve [its environmental] objectives’.

The European Centre for Disease Prevention and Control (ECDC) has been established pursuant to Article 152(4) EC (current Article 168(4) TFEU), pre-sumably on point c of that paragraph which allows for the adoption of ‘incentive measures designed to protect and improve human health’. When the European Food Safety Authority (EFSA) was established as the first agency of the third wave, this was done based on the current Articles 43, 114, 207, and 168(4) (b) TFEU, as originally proposed by the Commission. This marked the start of a change in the choice for legal bases. In its opinion the CLS noted that establishing an agency in itself does not require recourse to Article 352 TFEU.29 It found the four proposed legal bases appropriate and differentiated the EFSA from the European Medicines Agency (EMA) case (cf. section IV 1.1.3) since the EFSA regulation would not set up a centralized system replacing national authorities. Instead it would take over the tasks of existing EU committees and its advice would contribute to harmon-izing national measures.30

While Berger notes that the question of whether all these Treaty provisions in-clude the competence to make organizational arrangements or whether the latter flow from the implied powers doctrine is irrelevant,31 taking such a view would ignore part of the debate on the implied powers doctrine in the EU legal order, which is in any case different from the implied powers of international organiza-tions under international law.32

As Hartley explains, there is a narrow and a broad formulation of the doc-trine. Under the narrow formulation, ‘the existence of a given power implies also the existence of any other power which is reasonably necessary for the exercise of the former’.33 Under the broad formulation, ‘the existence of a given objective or function implies the existence of any power reasonably necessary to attain it’.34 Whether new structures within the EU organization are ‘reasonably necessary’ under a narrow formulation of the doctrine may be doubted for some agencies and even under the broad formulation this may appear problematic, depending on the interpretation ‘reasonably necessary’. On this point, Hartley refers to the

29 Legal Service of the Council of the European Union, 18 May 2001, Doc 8891/ 01, p 2 at para 3.30 Ibid, p 4 at paras 11– 12.31 Berger, n 13, pp 62– 3.32 As von Bogdandy and Bast explain, implied powers under international law are seen as a

counter- concept to the doctrine of attributed powers, whereas in the EU legal order the former is seen as an integral part of the latter. See Armin von Bogdandy and Jürgen Bast, ‘The Federal Order of Competences’, in von Bogdandy and Bast (eds), Principles of European Constitutional Law, München, Hart Publishing, 2010, p 282.

33 Hartley, n 21, p 105 (emphasis added). 34 Ibid, p 105 (emphasis added).


importance of Germany v. Commission, which dealt with the Commission’s power of consultation under Article 118 EEC35 (currently this power may be found in Article 156 TFEU). In that case, the Court noted:

It must be emphasized that where an article of the EEC Treaty confers a specific task on the Commission it must be accepted, if that provision is not to be rendered wholly inef-fective, that it confers on the Commission necessarily and per se the powers which are indispensable in order to carry out that task. Accordingly, the second paragraph of Article 118 must be interpreted as conferring on the Commission all the powers which are neces-sary in order to arrange the consultations.36

The Court in its reasoning thus replaced ‘indispensable’ with ‘necessary’, allowing for a more generous doctrine of implied powers.

However, extensive reliance on a generous doctrine of implied powers would expose the EU legislator (and EU integration) to the obvious critique that it side-lines the principle of conferred powers.37 An agency will seldom be really ‘indis-pensable’ for the exercise of certain competences explicitly attributed to the EU, barring recourse to the implied powers doctrine. In addition, if agencies may be established without recourse to the doctrine of implied powers, this would bolster their legitimacy.

1.1.3 Article 114 TFEU as a legal basis for agency creationThe broad formulation of the implied powers doctrine, referring to objectives, is akin to the provisions of Article 352 and 114 TFEU.38 These two functional provisions differ from the provisions on the other powers of the EU because the latter are defined in terms of a particular sector. The functional provisions on the other hand ‘are defined in terms of a cross- sectoral policy objective to be achieved, [ie] the establishment and functioning of the internal market’.39 In addition, both functional provisions are also residual provisions, to be used when the more spe-cific Treaty provisions are insufficient.40 De Búrca and de Witte note that the rel-iance on Article 114 TFEU was rather uncontested during the intensive period of internal market construction, but following the completion of the internal market core, the legislator relied upon it to legislate in more contentious areas.41 Indeed,

35 Ibid, p 105.36 Joined Cases 281, 283– 5, 287/ 85, Germany e.a. v. Commission, [1987] ECR 3203, para 28

(emphasis added).37 See the discussion in Luigi Corrias, The Passivity of Law: Competence and Constitution in

the European Court of Justice, Dordrecht, Springer, 2011, pp 19– 20. See also the discussion in ‘The Court of Justice in the Limelight— Again’, (2008) 45 CMLRev 6, pp 1571– 9. For a critique by Dashwood, see Dashwood, n 20, p 124.

38 Corrias, n 37, p 11.39 Gráinne de Búrca and Bruno de Witte, ‘The Delimitation of Powers Between the EU

and its Member States’, in Arnull and Wincott (eds), Accountability and Legitimacy in the European Union, Oxford, OUP, 2002, p 214.

40 Paul Craig and Gráinne de Búrca, EU Law: Text, Cases, and Materials, Oxford, OUP, 2011, pp 90 and 590.

41 de Búrca and de Witte, n 39, p 215.


Article 114 TFEU has been relied upon also for agency creation.42 According to de Búrca and de Witte one important limit to the use of Article 114 TFEU is that its object must be the harmonization of national laws,43 an objective arguably nar-rower than that of Article 352 TFEU.

The problem here again is that the text of the Article only hints at material rules which may be adopted at EU level, the issue of the possible formal arrange-ments not being explicitly addressed. According to Vos, establishing an agency as such is not possible under Article 114 TFEU, but adopting harmonization measures, together with provisions which establish an agency, is possible under a centre- of- gravity test.

The fate of the agency then depends on the answers to the questions of whether the harmonization measures ‘seek to eliminate existing or future differences be-tween national legislative provisions, and [whether] the agency [is] dependent upon such harmonization measures’.44 However, such an argument focuses on the tasks exercised by an agency, not on the actual establishment. This further be-comes clear when Vos remarks that the agency ‘would no longer possess the sup-plementary character if proper decision- making powers were delegated to it and these powers were exercised without Commission supervision’.45 The argument that only non- decision- making agencies could be established based on Article 114 TFEU seems detached from the content of that Article, which only revolves around the notion of ‘measures for approximation’.46 In any event, the EU legis-lator subsequently established decision- making agencies such as the European Chemicals Agency (ECHA), Agency for the Cooperation of Energy Regulators (ACER), and European Supervisory Authorities (ESAs) on the sole legal basis of Article 114 TFEU.

During the second wave of agency creation, the Commission had indeed pro-posed to establish the EMA based on Article 114 TFEU (then 100a EEC). The Council however changed the legal basis to Article 352 TFEU. The CLS invoked two reasons for this. Firstly and most importantly, it noted that certain measures at EU level could never be measures for approximation, ‘on the grounds that no

42 ACER, BEREC, and the BEREC Office, the three ESAs, ECHA, ENISA, and the SRB were established based on Article 114 TFEU. For the EFSA and EMA the Article functions as a joint legal base together with other Treaty provisions.

43 de Búrca and de Witte, n 39, p 215. Despite this limit, clear ‘red lines’ seem difficult to draw in practice as regards the use of Article 114 TFEU. See Walter Frenz and Christian Ehlenz, ‘Rechtsangleichung über Art. 114 AEUV und Grenzen gem. Art. 5 EUV nach Lissabon’, (2011) 22 EuZW 16, p 625.

44 Ellen Vos, ‘Reforming the European Commission: What Role to Play for EU Agencies?’, (2000) 37 CMLRev. 5, p 1122.

45 Ellen Vos, Institutional Frameworks of Community Health and Safety Legislation: Committees, Agencies and Private Bodies, Oxford, Hart Publishing, 1999, p 199.

46 Similarly, following ENISA (cf. section IV 1.1.3.1), Neergaard noted: ‘It follows from this caselaw that the legislature may confer non binding tasks […] On the other hand, this judgment does not allow the conclusion that such a body could be empowered to adopt acts that would bind Community institutions, national authorities or private undertakings.’ See Anders Neergaard, ‘European Supervisory Authorities— A New Model for the Exercise of Powers in the European Union?’, (2009) EUREDIA 4, p 618 (emphasis in original).


Member State would be in a position to adopt them by virtue of their purpose’.47 It then continued: ‘the creation of a Community procedure involving the granting of authorizations by the Commission with horizontal effect and the establishment of a European agency with legal personality recognized in all Member States are unfeas-ible actions and therefore inconceivable in national law’.48 Secondly, the CLS noted that the measures proposed were not approximation measures either, ‘as national action is completely superseeded [sic]’.49 Following the first of these reasons, one would assume that Article 114 TFEU can never function as the (sole) legal basis for establishing an agency, since the Member States separately could never establish an EU body. In its opinion on the EFSA legal basis, however, the CLS noted that the old practice of establishing EU committees through decisions separate from the directives rested on the assumption that EU bodies could not be established by direc-tives.50 That practice was later superseded by subsequent institutional practice where directives were used to establish EU bodies and the CLS noted that Article 114 TFEU could be used for measures which did not themselves amount to harmoniza-tion, but which do contribute to such harmonization.51 These issues figuring in the CLS’s opinions on the EMA and EFSA were later brought to the Court in ENISA.

1.1.3.1 The ENISA caseIn ENISA the UK questioned Article 114 TFEU as the legal basis of the ENISA Regulation, arguing that the regulation’s provisions took effect at the level of the EU’s institutional law whereas the purpose of Article 114 TFEU is that of approximation, similarly to the CLS’ reasoning in its opinion on the EMA (cf. section IV 1.1.3).

The Court thus had to deal with two issues raised by the UK: whether the ENISA’s tasks (when exercised) could be qualified as ‘approximation measures’ and whether formal arrangements, such as setting up a body, are permitted under Article 114 TFEU. To put this question into its context, AG Kokott referred to the conclusions of AG Stix- Hackl in the European Cooperative Society case,52 which were followed by the Court when it concluded that the new form of cooperative society would leave the national laws on cooperative societies unchanged, meaning no approximation was at issue.53

A clear parallel may be drawn between the creation of a new form of cooperative society and the creation of a new body at EU level. Still, whereas the European cooperative society acquires its practical legal effects within the nat-ional legal order, the UK had argued that the legal effects of creating an agency should be situated at EU level. According to AG Kokott, then, the ‘approximation provisions’ in the ENISA regulation were to be found in the tasks entrusted to

47 Legal Service of the Council of the European Communities, 19 November 1991, Doc 9525/ 91, p 12 at para 14.

48 Ibid, pp 12– 13 at para 15. 49 Ibid, p 14 at para 17.50 Legal Service of the Council of the European Union, Doc 8891/ 01, p 5 at para 13.51 Ibid, p 5 at para 14.52 Opinion of AG Stix- Hackl in Case C- 436/ 03, Parliament v. Council, [2006] ECR I- 3733,

paras 96– 8.53 Case C- 436/ 03, Parliament v. Council, [2006] ECR I- 3733, para 44.


ENISA, but these were overshadowed by the provisions on the establishment and organization of ENISA which ‘do not contribute directly to the approximation of provisions of the Member States’.54

The conclusion would already have been that the regulation could not be est-ablished based on Article 114 TFEU, were it not for the following observation by the AG: ‘[However,] the establishment of ENISA cannot be separated from its tasks, but is a means to the end. The ENISA Regulation thus pursues only a single aim, which is to be derived above all from the provisions on the tasks of ENISA’55— thus suggesting that the organizational arrangement was a power implied by the power contained in Article 114 TFEU, without entering into the question of whether the agency instrument was indeed necessary, or indispens-able, for the material arrangements.56 In the end, however, the AG found that ENISA’s tasks did not qualify as measures for approximation in the first place.57

The Court itself merely remarked that ‘[t] he legislature may deem it neces-sary to provide for the establishment of a Community body responsible for con-tributing to the implementation of a process of harmonisation’.58 It therefore spoke in very general terms of ‘Community bodies’, without making a distinc-tion between permanent or temporary bodies and bodies with or without legal personality. Having dealt with the issue of establishment, it did not further make a clear distinction between the material and organizational rules in the Regulation, unlike the AG, but seemingly took the establishment and the tasks of the agency as a single measure. It further remarked that such a Community body could contribute to the implementation of a harmonization process where ‘the adoption of non- binding supporting and framework measures seems appro-priate’.59 After having concluded that the ENISA’s tasks could be situated within the broader regulatory framework of network security, it hinted at making a distinction by stating that ‘it needs to be determined […] whether the establish-ment of the Agency and the objectives and tasks which are assigned to it by the regulation may be regarded as “measures for approximation”’.60 However, it then continued by observing that the ‘legislature considered that the establishment of a Community body such as the Agency was an appropriate means of preventing the emergence of disparities likely to create obstacles to the smooth functioning of the internal market’.61 This seems questionable, since the establishment of a new EU body as such can never be a means to prevent disparities from emerging and it stands in contrast with the opinion of the AG, who, rightly, qualified the establishment of the ENISA as the instrument allowing it to carry out its tasks.

54 Opinion of AG Kokott in Case C- 217/ 04, United Kingdom v. Parliament and Council, [2006] ECR I- 3771, para 26.

55 Ibid, para 27.56 According to Vetter it should be shown that the establishment of an agency, based on Article

114 TFEU and the implied powers doctrine, is compellingly necessary (zwingend erforderlich). See Vetter, n 7, p 729.

57 Opinion of AG Kokott in Case C- 217/ 04, n 54, para 46.58 Case C- 217/ 04, United Kingdom v. Parliament and Council, [2006] ECR I- 3771, para 44.59 Ibid, para 44. 60 Ibid, para 59 (emphasis added). 61 Ibid, para 62.


Because it remained uncertain whether the ENISA would actually contribute or facilitate harmonization, the Court concluded by emphasizing that the ENISA was established only for five years and that its performance would be reviewed by the Commission. According to Bouveresse this element was decisive in the Court coming to its conclusion.62

Following ENISA, Article 114 TFEU may be presumed to provide a legal base to supplement the EU’s institutional structure. Still, the ruling received a mixed reception. Ohler observed: ‘If the reference point of Article 114 TFEU is the difference in Member State provisions which should be approxi-mated, it is hard to see how an organisational act establishing a Community agency may be qualified as the approximation of legal rules and administra-tive measures.’63 Even if Mok welcomed the Court’s ruling more favourably, he admits that the Court ‘extended’ the scope of Article 114 TFEU,64 giving a rather pragmatic justification for doing so: since the agency’s purpose is to help implement directives adopted pursuant to Article 114 TFEU, it should be established pursuant to that same Article.65 Of course, this reasoning also depends on whether the ENISA (when it exercises its limited tasks) indeed contributes to the effective implementation of that legislation. In his analysis of the ruling, Randazzo then identified a two- tier test whereby the Court verified firstly whether the ENISA’s tasks could readily be situated within the broader legislative framework and, secondly, ‘whether those […] tasks may be regarded as supporting and providing a framework for the implementation of that legislation’.66 The two- tier approach identified by Randazzo shows how the Court, once the test is passed, seemingly leaves it to the institutions to decide which organizational form is best suited as the formal instrument to deliver the policy envisaged.

Still, Randazzo, but also Gutman, deplored the fact that the Court did not enter into the general question of agency creation without recourse to Article 352 TFEU.67 Sander aptly summarizes the ruling by observing: ‘Through this CJEU- ruling it has been clarified that the creation of bodies and the delegation thereto of (executive) powers is permissible on the general legal basis of Article 114 TFEU, when the tasks of this body, simply put, serve the continued integration through

62 Aude Bouveresse, ‘Bases juridiques autorisant la création d’organismes dotés d’une person-nalité juridique propre’, (2006) Europe 203, p 10.

63 Ohler, n 17, pp 373– 4. Translated from German by the author. See in the same vein Vetter, n 7, p 729. In addition, Vetter is of the opinion that the powers granted to agencies cannot find their basis in Article 114 TFEU either.

64 Mok actually admits that AG Kokott’s arguments were legally more convincing, but he found the Court’s solution more appealing for pragmatic reasons.

65 Robert Mok, ‘Nos 498 & 499: Annotatie bij de zaken C- 436/ 03 en C- 217/ 04’, (2006) Nederlandse Jurisprudentie 40, p 4103. This was also remarked by the European Parliament in ENISA, see Case C- 217/ 04, n 58, para 23.

66 Case C- 217/ 04, n 58, para 47; Vincenzo Randazzo, ‘Case C- 217/ 04, United Kingdom v. European Parliament and Council of the European Union’, (2007) 44 CMLRev 1, pp 161– 5.

67 Katleen Gutman, ‘Case C- 66/ 04, Smoke Flavorings: Case C- 436/ 03, SCE; & Case C- 217/ 04, ENISA’, (2006) 13 CJEUL, pp 176– 7; Randazzo, n 66, pp 168– 9.


the harmonization of legal rules.’68 At the same time he draws attention to an interesting effect of the Court’s emphasis on the legislative and regulatory con-text in which an agency would exercise its tasks. Thus not only is the question on the institutional form left to the institutions; the first precondition (identified by Randazzo) also lies in the hands of the legislator, who may ‘unlock’ the agency instrument by firstly (or simultaneously) adopting a ‘sufficient’ body of material rules, providing the necessary context in which the agency would operate.69

1.1.3.2 The Short- selling caseSince the UK, in the Short- selling case, mainly questioned the empowerment of the European Securities and Markets Authority (ESMA) in the light of the Court’s Meroni and Romano jurisprudence, the circumstances and details of the case will be presented in the sections on those rulings.70 Here it suffices to note that the regulation, adopted pursuant to Article 114 TFEU, empowered ESMA to regulate shorting on the financial markets. The UK raised four pleas, only the last of which is relevant for the present section. According to the UK, ‘if and to the extent that Article 28 [of Regulation 236/ 2012] were interpreted as empowering ESMA to take individual measures directed at natural or legal persons, it would be ultra vires Article 114 TFEU’.71

The question in this case differed from that in ENISA. Whereas ENISA re-volved around the question of establishing an agency, Short- selling dealt with the empowerment of an agency. The UK inter alia took issue with the possibility that the ESMA would directly instruct a natural or legal person to disclose his shorting position(s) through individual measures. This was different from the situation in General Product Safety where the Court had sanctioned the adoption of individual measures based on Article 114 TFEU,72 since the Commission and Council in that case had emphasized that the individual measures at issue were addressed to the Member States and did not have direct effect with respect to individuals.73

AG Jääskinen proposed to the Court to solve the case on the plea related to the legal basis. Similar to the logic structuring the present chapter, the AG rightly noted that from ‘a constitutional law point of view, the assessment of the legal basis [of the contested power granted to the ESMA] precedes subordinate legal issues relating to its content’.74 In his analysis, the AG first set out the existing

68 Peter Sander, ‘Europäische Agenturen: Rechtsgrundlagen in und sonstige Berührungspunkte mit dem primären Unionsrecht’, in Raschauer (ed), Europäische Agenturen, Wien, Jan Sramek Verlag, 2011, p 26. Translated from German by the author.

69 See ibid, p 27.70 The present discussion is partially based on Merijn Chamon, ‘The Empowerment of Agencies

under the Meroni Doctrine and Article 114 TFEU: Comment on United Kingdom v Parliament and Council (Short- selling) and the Proposed Single Resolution Mechanism’, (2014) 39 ELRev 3, pp 383– 9.

71 See Case C- 270/ 12: Action brought on 1 June 2012, OJ 2012 C 273/ 3.72 Case C- 359/ 92, Germany v. Council, [1994] ECR I- 3681, para 37.73 See ibid, para 16.74 Opinion of AG Jääskinen in Case C- 270/ 12, UK v. Council and Parliament,

ECLI:EU:C:2013:562, para 26.


jurisprudence of the Court, emphasizing the Smoke Flavourings and ENISA cases (cf. section IV 1.1.3.1). AG Jääskinen was convinced that in the light of these rul-ings there was no problem in relying on Article 114 TFEU to establish ESMA, since the agency undoubtedly contributes to the implementation of a process of harmonization in the financial sector.75 However, whether the power under Article 28 of the regulation could be based on Article 114 TFEU was a different matter.76

AG Jääskinen rightly noted that this decisive issue depended on whether the ESMA’s decisions either contributed or amounted to internal market harmoniza-tion.77 The contested provision then failed the AG’s test because ‘the conferral of decision making powers […] in substitution for the assessments of the competent national authorities, cannot be considered to be a measure [of approximation] within the meaning of Article 114 TFEU’.78 The AG came to this conclusion following a reasoning which shared similarities to the one followed by the CLS in its opinion on the EMA (cf. section IV 1.1.3). The AG similarly found that Article 28 simply ‘elevated’ intervention powers to the EU level.79 The AG further rejected the Commission’s reliance on General Product Safety because the ESMA would not further develop specific and detailed rules but would simply intervene in the financial market similarly to national financial supervisory authorities.80 In short: for the AG the outcome of ESMA’s powers under Article 28 was ‘not harmonisation […] but the replacement of national decision making […] with EU level decision making’.81 Without referring to that author, the AG thereby followed Michael’s critique on EIOPA.82

After having rejected Article 114 TFEU as an appropriate legal basis, the AG suggested the use of Article 352 TFEU to grant the ESMA the contested powers, since the contested powers were still necessary to prevent the internal market in financial services from being distorted. The AG added to this that Article 352 TFEU would have enhanced democratic input, given the unanimity requirement in the Council.83 Orator questioned this reasoning,84 and in any event it should

75 Ibid, paras 32– 4.76 Article 9(5) of the ESMA Regulation itself only contained a general clause stating that fur-

ther legislative acts could confer the power on the ESMA to temporarily prohibit or restrict certain financial activities. Article 28(2) provides that the power under paragraph 1 may only be exercised if no national authority has taken adequate measures to address a threat to the orderly functioning and integrity of the whole or part of the financial system in the Union and if there are cross- border implications.

77 Opinion of AG Jääskinen in Case C- 270/ 12, n 74, para 36. 78 Ibid, para 37.79 Ibid, para 39. 80 Ibid, para 45. 81 Ibid, para 52.82 See Dirk Looschelders and Lothar Michael, ‘Europäisches Versicherungsrecht’, in

Ruffert (ed), Europäisches Sektorales Wirtschaftsrecht, Baden- Baden, Nomos, 2013, p 694. Even before the establishment of the ESMA, Neergaard noted that ‘[i] t is difficult to see that a Community act approximates national provisions if it directly provides the applicable provisions in a matter and also that the application in the individual cases of this act is to take place on Community level.’ See Neergaard, n 46, p 619.

83 Opinion of AG Jääskinen in Case C- 270/ 12, n 74, para 58.84 Andreas Orator, ‘Die unionsrechtliche Zulässigkeit von Eingriffsbefugnissen der ESMA

im Bereich von Leerverkäufen’, (2013) 24 EuZW 22, p 854. See also Rob van Gestel, ‘European Regulatory Agencies Adrift? Case C- 270/ 12’, (2014) 21 MJECL 1, p 192.


be noted that Article 352 reduces the input from the European Parliament com-pared to Article 114 TFEU,85 rendering the AG’s argument very debatable indeed.

Even if the effects of this case for the process of agencification could have been limited by solving it solely on the plea related to Article 114 TFEU, a number of other EU agencies had also been empowered under Article 114 TFEU and may adopt individual decisions addressed to natural and legal persons.86 Should the Court indeed have followed the suggestion of AG Jääskinen, this could have had repercussions for those agencies as well. The AG sought to argue differently by noting that those other agencies ‘cannot make legally binding decisions directed at individual legal entities in substitution for either a decision, or the inaction, of a competent national authority which may well disagree with a decision taken by ESMA’.87 Here the AG seems to have overlooked the ACER, which in a number of cases may also take decisions in substitution of the national regulatory auth-orities, inter alia when those authorities cannot agree on a decision between themselves.88

Alternatively to the AG’s analysis, a reading of Article 114 TFEU taking into account the Court’s prior jurisprudence could also have led to a different conclu-sion. In Smoke Flavourings the Court had validated a centralized procedure at EU level for the authorization of certain food products,89 and in General Product Safety the Court sanctioned the adoption of individual measures addressed to the Member States. Only in European Cooperative Society did the Court reject the use of Article 114 TFEU to establish a new legal form existing alongside the existing national legal forms for cooperative societies.90

In the light of these cases Article 114 TFEU might therefore have allowed for the centralized procedure of Article 28 of Regulation 236/ 2012, covering both the adoption of general and individual measures as ‘harmonization measures’ that do not co- exist alongside national measures. The questions remaining would be whether the essential elements of the harmonization measure were already con-tained in the basic act, as prescribed by Smoke Flavourings, and whether the in-dividual measures may also be addressed to private parties, unlike in General Product Safety.

Probably because it had already ruled in relation to the first three pleas that the ESMA had not been accorded discretionary powers (cf. section IV 5.8.1.3), the Court did not dwell on the first of these questions. As to the second question, the Court recalled its observations in General Product Safety and ENISA, finding that nothing in the wording of Article 114 TFEU ‘implies that the addressees of the measures adopted by the Community legislature on the basis of that provision can only be the individual Member States’.91

85 The AG noted this himself in footnote 78 of the Opinion.86 These are for instance the other ESAs, the ECHA, and the ACER. See also Orator, n 84,

p 853.87 Opinion of AG Jääskinen in Case C- 270/ 12, n 74, para 24 (emphasis in original).88 See Article 8(1)a ACER.89 Case C- 66/ 04, UK v. Parliament and Council, [2005] ECR I- 10553.90 Case C- 436/ 03, n 53, para 44. 91 Case C- 217/ 04, n 58, para 44.


The Court then dismissed the UK’s plea by testing the contested power in light of the express Treaty provision, noting that ‘a legislative act adopted on that legal basis must, first, comprise measures for [approximation] and, second, have as its object the establishment and functioning of the internal market’.92 On the first of these requirements, the Court stressed the large margin of discretion left to the legislator to choose the most appropriate method of approximation and found that the legislator ‘may delegate to a Union body, office or agency powers for the implementation of the harmonisation sought’.93 With this first issue out of the way, the result of the test was already determined, since there was no doubt that the measure in casu had as its object the proper functioning of the internal market.

The Court’s finding was in a sense unsurprising as it followed the logic of prior jurisprudence on Article 114 TFEU,94 thereby also exposing itself to the exist-ing critique on the Court’s Article 114 TFEU jurisprudence.95 The result of the Court’s ruling is foremost a sanctioning of agencification as developed during more than two decades (cf. section IV 1.1.4). Nevertheless the question could be raised as to whether the Court should not (have) placed greater emphasis on the EU legislator’s obligation to properly reason its decision to extraordinarily em-power an agency, rather than emphasizing the large margin of discretion left to the legislator to choose the most appropriate method of approximation.96

1.1.4 Institutional practice of agency creationFrom the perspective of the principle of conferred powers it appears that the Treaty authors did not want to withhold the power to establish new bodies from the EU institutions, in light of the references to other bodies in the Treaty pro-visions on the tasks of the Court of Auditors and the preliminary ruling pro-cedure. The Court’s rather lax stance on the issue of the legal basis was criticized in the preceding section, since it sits uneasily with the established canon that the choice of a legal basis is of constitutional significance.97 From this perspective, the

92 Case C- 270/ 12, UK v. Parliament and Council, ECLI:EU:C:2014:18, para 100.93 Ibid, para 105.94 For a different view see Dariusz Adamski, ‘The ESMA Doctrine: A Constitutional

Revolution and the Economics of Delegation’, (2014) 39 ELRev 6, p 828. Maletić has discussed this jurisprudence more elaborately, drawing attention to the Court’s broad interpretation of ‘measures for approximation’ and the legislator’s broad discretion in choosing the most appropriate method of harmonization. See Isidora Maletić, The Law and Policy of Harmonisation in Europe’s Internal Market, Cheltenham, Edward Elgar, 2013, pp 28– 67, pages 33– 43 specifically. Clément- Wilz also notes: ‘[Á] l’appui d’un raisonnement fragile, la Cour confirme son interprétation large de la norme d’habilitation donnée dans l’article 114 TFUE, donnant un quasi- blanc- seing aux institutions de l’Union pour créer des agences dotées de pouvoirs normatifs.’ See Laure Clément- Wilz, ‘Les agences de l’Union européenne dans l’entre- deux constitutionnel’, (2015) 51 RTDE 2, p 339.

95 Specifically when it comes to the establishment and empowerment of EU agencies, see n 63.96 Maletić also concludes that further agencification based on Article 114 TFEU will likely

remain to be contested. See Isidora Maletić, ‘Delegating Harmonization of the Internal Market: the Ruling in Case C- 270/ 12’, (2014) 33 YEL 1, p 514.

97 See eg Opinion 2/ 00, re the Cartagena Protocol, [2001] ECR I- 9713, para 5.


Court would do well to take a less passive approach on this issue in its scrutiny of agencification.

In line with the dominant opinion in legal doctrine, the early agencies had all been established based on Article 235 EEC. However, in the 1970s the Commission also made a number of proposals for agencies which were never adopted by the Council. In three out of four of these proposals the Commission had suggested Article 235 EEC as a (joint) legal basis.98 But in its proposal for a regulation establishing a Community guarantee system, the Commission put forward Article 113 EEC (current Article 207 TFEU) as the single legal basis to establish a European Private Investment Guarantee Office, since the material rules in the proposal were predominantly trade- related.99 The difference in legal basis was quite significant since Article 113 EEC provided that the Council would decide on Commission proposals by qualified majority, not by unanimity as under Article 235 EEC. The proposals on a European export bank and a European agency for trade cooperation also related to the Common Commercial Policy, but here the Commission suggested the joint legal bases of Articles 113 and 235 EEC. Because these proposals were never adopted by the Council,100 they merely shed some light on the Commission’s position and on that of the Parliament, which also adopted opinions.

For the Parliament, there would be no difference at first sight between the two alternative legal bases, with it only holding an advisory power under both proce-dures. This makes it all the more interesting that it called upon the Commission to change the legal basis for the guarantee scheme from Article 113 EEC to Article 235 EEC.101 The original report drafted by MEP Armengaud did not contain such a provision,102 and unfortunately no reasons were cited for the amendment changing the legal basis.103 One can only guess therefore whether the Parliament was of the opinion that (i) establishing an agency necessitated recourse to Article 235 EEC, (ii) Article 235 EEC was necessary because the proposal also contained aspects related to development policy,104 or (iii) Article 235 EEC was preferable

98 Proposal of the Commission for a Council Regulation, OJ 1975 C 258/ 2; Proposal of the Commission for a Council Regulation, OJ 1976 C 76/ 2; Proposal of the Commission for a Council Regulation, OJ 1976 C 256/ 2.

99 European Commission, COM (72) 1461 final.100 According to the Commission, the Member States in Council were not enthusiastic about

its proposal on a Community guarantee scheme because of the economic climate at that time. See Answer of the European Commission to Parliamentary Question 217/ 75, OJ 1975 C 233/ 13. Given that these three proposals on new agencies should be situated in the same economic sphere, it is likely that the reason for the non- adoption of the other proposals should be situated along the same lines.

101 See Resolution of the European Parliament, OJ 1974 C 23/ 31.102 Rapporteur A. Armengaud, Report of 27 November 1973 on the Proposal from the Commission

to the Council for a Regulation Establishing a Community Guarantee System for Private Investments in Third Countries, European Parliament Working Documents, Document 208/ 73.

103 See Amendment proposed by MEPs Van der Hek, Ariosto, Harmegnies, and Spenale, PE 35.636/ rev.

104 This is what the Commission’s Legal Service itself suggested when it invoked the (almost) precedent of the Office to argue that the sole recourse to Article 130s EEC (without Article 235 EEC) would not be problematic. See n 106.


because it gave a formal role to the Parliament instead of the informal advisory role the Parliament had under Article 133 EEC.105

The first time an EU agency was established without recourse to Article 352 TFEU was the creation of the EEA.106 On this occasion, the Council followed the Commission’s suggestion to establish the EEA using Article 130s EEC. The establishment of the EEA was not the EU’s first step in the field of environmental data collection. The aims of the Corine programme, in the 1980s, were the same as that of the EEA.107 The Corine programme was originally based on Article 235 EEC, but under the Single European Act (SEA) the EEC had been conferred a proper competence and the Commission took the opportunity to propose the EEA based on this new environmental competence. In the explanatory memo-randum to its proposal,108 the Commission stressed not only this new competence but also the Member States’ previous commitment to use the new competences under the SEA.109

As early as the beginning of the second wave of agency creation, a consensus could thus be identified among the three main institutions in that recourse to Article 352 TFEU was not indispensable to establish EU agencies. Still, it took until the new millennium before recourse to the sector- specific legal bases became the rule for agency creation. During the 1990s the Commission tried to build on the EEA precedent in its proposals for the Community Plant Variety Office (CPVO) and the EMA,110 but the Council insisted on using Article 235 EEC.111

From a legal perspective, it may be noted that Article 352 TFEU indeed pro-vides a more secure legal basis covering all elements of the legislative act, whereas this would be less clear if a specific legal basis was to be used. From a political perspective, however, it is clear that recourse to Article 352 TFEU allowed every Member State a firm grip on the process of agencification. At the same time, with the Parliament’s position being strengthened from the SEA and the Maastricht Treaty onwards, this practice came under increasing pressure. In addition,

105 Still, even under these three options it is not entirely clear why Parliament would propose a single instead of a dual legal basis. Perhaps under the third option this could be explained by the fact that discussions on the proposal in casu predated the Court’s jurisprudence emphasizing the democratic importance of the Parliament’s involvement in the advisory procedure.

106 The Commission’s legal service cited the ‘precedent’ of the Guarantee Office for establishing an agency using a sector- specific legal basis. See Service Juridique de la Commission, 2 juin 1989, JUR (89) D/ 3370, pp 1– 2.

107 See Decision (EEC) 85/ 338 of the Council, OJ 1985 L 176/ 14. The decision to extend the period of the Corine programme to allow the EEA to take over those activities once it became fully operational was based on Article 130s EEC. See Decision (EEC) 90/ 150 of the Council, OJ 1990 L 81/ 38.

108 See European Commission, COM (89) 303 final.109 See the Declaration on the Environment, annexed to the Presidency Conclusions of the 1988

Rhodes European Council.110 See Proposal of the European Commission, OJ 1990 C 244/ 1; Proposal of the European

Commission, OJ 1990 C 330/ 1.111 See also Alexander Kreher, ‘Agencies in the European Community— A Step towards

Administrative Integration in Europe’, (1997) 4 JEPP 2, p 232; Ellen Vos, ‘Agencies and the European Union’, in Zwart and Verhey (eds), Agencies in European and Comparative Perspective, Antwerpen, Intersentia, 2003, p 119.


Geradin and Petit note that the Commission continued to support a move away from Article 308 EC because agencies established on specific legal bases may be controlled more easily112 and, obviously, because a move away from Article 308 EC allowed for majority voting in the Council.113 Finally, Mok’s parallelism argu-ment put forward in the context of the ENISA case (cf. section IV 1.1.3.1) seems insightful here, since it does not make much sense to resort to qualified majority voting (QMV) to adopt material rules while insisting on unanimity to adopt the arrangements which merely help the Member States to meet their obligations under the Treaties.114

From the third wave of agency creation onwards, Article 352 TFEU has only been relied on to establish the Fundamental Rights Agency (FRA). In addition, when the existing agencies’ regulations are recast, their legal basis may be changed as well, such as the EMA’s legal basis changing to Articles 114 and 168(4)(b) TFEU.115 This decision raises the question of how the CLS’s original opinion on the EMA should be assessed in the current legal framework. Assuming Article 114 TFEU still does not allow a superseding of national authorities’ competences, this aspect would be covered by Article 168(4)(b) TFEU which simply provides that the ordinary EU legislator may adopt ‘measures in the veterinary and phyto-sanitary fields which have as their direct objective the protection of public health’ (emphasis added).

So far, the Parliament’s reinforced role under Article 352 TFEU by the Lisbon Treaty has not changed the existing practice, since the unanimity requirement for the Council was kept intact and the Parliament has ‘merely’ received the power of consent, rather than a power of codecision.

1.1.5 ConclusionGiven that the Treaty provisions are drafted in general terms, the question of whether the agencies’ executive powers fall within the powers conferred on the EU in the first place is not that contentious. In addition, the question as to the ex-istence of a material competence will normally already be resolved at the moment the EU adopts the legislation which agencies help implement.

The question of whether the EU institutions are competent to establish new EU bodies in the absence of a specific provision in the EU Treaties is a different matter. On this issue an interesting evolution, influenced by the process of agencification,

112 According to Geradin and Petit, this is so because the specific legal basis immediately limits the future scope of action of the agency, whereas Article 352 TFEU does not have such a ‘pre- emptive’ function.

113 Damien Geradin and Nicolas Petit, ‘The Development of Agencies at EU and National Levels: Conceptual Analysis and Proposals for Reform’, (2005) 23 YEL, p 177.

114 On the other hand, one could argue that Member States might be persuaded to agree to (vague) material rules which still allow them a significant margin of manoeuvre in the implementa-tion phase. The establishment of an agency could then threaten this margin.

115 However, the 2009 OHIM Regulation is still based on the single legal basis of Article 308 EC.


may be noted in legal literature. Early on it was argued that the EU institutions had to make do with the organizational armamentarium as defined in the Treaties. Later, Article 352 TFEU was presented as the only legal basis available to the insti-tutions to establish new EU bodies. Following the second wave of agencification, some authors argued that other legal bases, such as Article 114 TFEU, could also be used unless the agency was empowered to take binding decisions.

In the ENISA case, the Court sanctioned the use of Article 114 TFEU to est-ablish an agency and, following the establishment of agencies such as the ECHA and ESAs, decision- making agencies have been founded on the single legal basis of Article 114 TFEU. If legal opinion continues to follow institutional practice, the conclusion must be that the power to establish new EU bodies is possible under Articles 352 and 114 TFEU and under virtually all the legal bases for the EU’s sectoral policies. Under the reasoning of the Court in ENISA, the legisla-tor’s decision to establish such a new body would be largely discretionary and the establishment of a new institutional entity as such incidental to tasks granted to that entity.

Because of this, and assuming the Court would not give the legislator carte blanche as to when an agency may be established, a number of issues remain unresolved. If the competence to exercise certain powers at EU level implies the competence to establish a new body to which these powers are then granted, should there not be a more rigorous test as to the indispensability or necessity of establishing that body?

This is something the Court refrained from doing in ENISA, when it merely checked whether the tasks conferred on ENISA were linked to the relevant har-monization directives. Ideally the Court would also have checked whether these tasks necessarily have to be exercised by a body such as an EU agency.

A second issue is that which Vetter hinted at in the context of Article 352 TFEU, but which may be generalized: does the analysis change when the process of agencification is looked at rather than the agencies individually? Obviously this is something which the Court could not enter into in ENISA but, as Vetter noted, agencification may alter the rule of indirect administration of EU law contained in the Treaties.116

A Treaty amendment whereby a provision is written in the Treaties allowing for the establishment and empowerment of agencies could address this issue, but only if this is embedded in an updated approach towards the implementation of EU law, away from the traditional dichotomy between direct and indirect admin-istration (cf. Chapter VI).

1.2 The choice of instrument

The question of which legal instrument may be used to establish EU agencies is discussed under the section devoted to the principle of conferred powers because

116 See Vetter, n 7, p 731.


of the link between this principle and the EU’s legal instruments. Pre- Lisbon, this link was already implicit in Article 249 EC which provided that the institu-tions could adopt the instruments listed, ‘in order to carry out their task and in accordance with the provisions of the Treaty’.117 Article 288 TFEU made the link more explicit, providing that ‘to exercise the Union’s competences, the institu-tions shall adopt regulations, directives, decisions, recommendations and opin-ions’. Furthermore, the link was also confirmed by the Court in Faccini Dori, in which it refused to establish a horizontal direct effect for directives, noting that to do so ‘would be to recognize a power in the Community to enact obligations for individuals with immediate effect, whereas it has competence to do so only where it is empowered to adopt regulations’.118

The question of the strength of this link is a different matter. Von Bogdandy et al119 rightly note that there is a stark contrast between this alleged function and the readiness with which the activism of the institutions in creating new instru-ments has been accepted by legal scholars120 and further tolerated by the Court. To say that there has been a prolific activism on the part of the institutions in cre-ating new instruments is an understatement.121 Consequently it is a worthwhile exercise to identify those EU instruments which could be used to establish EU agencies.

At first sight, institutional practice seems to have resolved this question, as EU agencies have all been created through regulations. The regulation would indeed seem to be the instrument par excellence with which an agency should be estab-lished. The main reason for this lies in the agencies’ legal personality (cf. section I 1.1.4): the most straightforward way to effectuate the agencies’ ‘most extensive legal capacity’ is by introducing these provisions into the most pervasive legal in-strument at the EU’s disposal— the regulation.

Obviously, an EU directive could also instruct the Member States to ensure the agency enjoys the most extensive legal capacity in their legal orders, but this would be more cumbersome. Still it should not be excluded, since directives may also contain ‘organic’ provisions, setting up new EU bodies.122 This option could indeed be envisaged if the agency’s legal personality may be conceived as a bundle

117 According to de Witte, Article 249 EC did not provide the institutions with a general auth-orization to adopt the acts listed in that Article; instead ‘[w] hether they may adopt an act, and in what form they can do so, will depend on the description of the powers granted to them in spe-cific Treaty provisions’. See Bruno de Witte, Adrie Geelhoed, and Jan Inghelram, ‘Legal Instruments, Decision- Making and EU Finances’, in Kapteyn (ed), The Law of the European Union and the European Communities: With Reference to Changes to be Made by the Lisbon Treaty, Alphen Aan Den Rijn, Kluwer, 2008, p 275.

118 Case C- 91/ 92, Faccini Dori, [1994] ECR I- 3325, para 24.119 Armin von Bogdandy, Felix Arndt, and Jürgen Bast, ‘Legal Instruments in European

Union Law and their Reform: A Systematic Approach on an Empirical Basis’, (2004) 23 YEL, p 96.120 See for instance Hans Peter Ipsen, Europäisches gemeinschaftsrecht, Tübingen, Mohr,

1972, p 425.121 See Silvère Lefèvre, Les actes communautaires atypiques, Bruxelles, Bruylant, 2006, p 552.122 As was noted previously, the EMA finds its origins in two EEC comitology committees

which had been set up by Directives: see Chapter III, n 2.


of legal personalities recognized in twenty- eight different national legal orders.123 The use of a directive would however also be more cumbersome on other points, since the acts establishing agencies are never confined to mere organic provisions on the agency’s structure, tasks, and functioning but also contain material pro-visions. Generally, the directive would be unsuitable as an instrument for these provisions since the ‘choice of form and methods’, which a directive leaves to the Member States, should not be left to the Member States for these matters. In addition, an agency may only be properly empowered to take binding decisions vis- à- vis Member States or private parties through the ‘directly applicable’ regu-lation.124 Since these material rules are separable from the organic provisions, the former could still be adopted in a separate legal instrument, but such a complex solution would again show why a regulation would be the more preferable instru-ment to establish agencies.

Apart from the regulation and the directive, Article 288 TFEU also mentions the decision as a binding act. The Article 288 TFEU decision must be differen-tiated from the Article 249 EC decision. The latter was ‘binding in its entirety upon those to whom it is addressed’ whereas the former may be binding in its entirety. The Article 288 TFEU decision is the result of the merger of the Article 249 EC decision (Entscheidung or Beschikking in German and Dutch) and the ‘addresseeless’ or sui generis decision (Beschluß or Besluit in German and Dutch), an atypical act.

At first sight, the Entscheidung decision does not seem to lend itself as an in-strument to establish agencies, since it is not a measure of general scope and nei-ther may it result in measures of general scope, unlike the directive. It is sur-prising therefore that the Commission’s original proposal for the first EU agency provided that the European Monetary Cooperation Fund (EMCF) would be es-tablished pursuant to a Council Entscheidung decision,125 even if the Monetary Fund was eventually established by regulation. However, as Mager noted in her study on the Entscheidung decision, that instrument has fulfilled different functions when addressed to the Member States. Apart from perhaps the most well- known decisions in state aid cases and in the Common Agricultural Policy (CAP), Mager notes that they have also been used to make the implementation of EU law, through indirect administration, more uniform.126 Similarly to the directive, an Entscheidung decision could be used to instruct Member States to recognize the legal personality of the agency and to instruct them to set up certain bodies or units which are then organized in a European network managed by an

123 This issue intersects with that of the legal basis. Following the CLS’ legal opinion on the EMA, one could argue that Article 114 TFEU is suitable if the agency’s legal personality is a bundle of national legal personalities. If the agency’s legal personality is an EU legal personality, however, it is clear that individual Member States (acting collectively) could not themselves achieve this result.

124 See to this effect also Case C- 270/ 12, n 92, para 110.125 See European Commission, COM (73) 69 final.126 Ute Mager, ‘Die staatengerichtete Entscheidung als supranationale Handlungsform’,

(2001) 36 Europarecht 5, p 668.


agency. However, the Entscheidung decision would be ill suited both to establish the agency as such and to set out material provisions.

Whether the Beschluß decision lent itself as an instrument to establish an agency, and whether the general decision of Article 288 TFEU does so now, is a different matter. As was noted previously, this act may be of general application, just like the regulation. Unlike the regulation, however, the Beschluß decision has a more limited use, since ‘it cannot impose, either directly or indirectly, obliga-tions on individuals or Member States’.127 Desomer notes that this type of act is used for (i) decisions of an organic nature, (ii) decisions approving international agreements, and (iii) decisions setting up EU programmes.128

The Beschluß decision has indeed been used in the past to establish new EU bodies, such as the European Statistical Governance Advisory Board129 or the European Administrative School.130 What is more, some of the forerunners of EU agencies were also established by the Commission using Beschluß decisions. In the introductory chapter, the examples of the European Regulators' Group for Electricity and Gas (ERGEG) and the European Regulators Group for Electronic Communications Networks and Services (ERGECNS) as forerunners to the ACER and Body of European Regulators for Electronic Communications (BEREC) (Office) were given.131

So far, the major difference between these bodies and the EU agencies is that the former are not endowed with their own legal personality. One would conclude that this is the limit to the use of the Beschluß decision, were it not for the EU executive agencies which also have legal personality and which are established by Commission decisions. This elicited the following observation by Desomer: ‘It is not clear why the unaddressed decision is not being used more systematically for the adoption of such organic acts [setting up EU agencies], like the Commission does for setting up executive agencies.’132

A number of elements should be highlighted in trying to answer Desomer’s question. As was noted, the acts setting up EU agencies, unlike the acts setting up executive agencies, are rarely purely ‘organic acts’ as referred to by Desomer, since they also contain material rules. Further, in post- Lisbon terms, the decisions setting up executive agencies are not legislative but implementing decisions which cannot be dissociated from the Council’s (legislative) regulation. Even if the ex-ecutive agencies are established by decision this cannot be invoked as positive proof that a regulation is not necessary for their establishment, since the decisions’ legal basis is a regulation.

127 Jürgen Bast, ‘Legal Instruments and Judicial Protection’, in Von Bogdandy and Bast (eds), Principles of European Constitutional Law, München, Hart Publishing, 2010, p 365. Bast notes, how-ever, that the Beschluß decision can create actionable individual rights vis- à- vis the EU institutions.

128 Marlies Desomer, The Reform of Legal Instruments of the European Union, Leuven, KUL 2009, PhD Thesis, pp 120– 1. See also de Witte, who also mentions the budget as a typical Beschluß decision: de Witte, Geelhoed, and Inghelram, n 117, pp 288– 9.

129 Decision (EC) 235/ 2008, OJ 2008 L 73/ 17.130 Decision 2005/ 118, OJ 2005 L 37/ 14. 131 See section III 1.132 Desomer, n 128, p 192.

Limits under the Subsidiarity Principle 157

In addition, it is the regulation itself which prescribes that the agencies have legal personality and that the Commission should rely on decisions,133 the deci-sions themselves not even mentioning the agencies’ legal personality or containing material rules.

Were it not for the separate legal personality of executive agencies, there would be no discussion as to whether the decisions come under the notion of ‘internal organization’ of the Commission. As the Commission itself explained in its ori-ginal proposal for the Council regulation: ‘the sole raison d’être of such agencies is to participate in managing Community programmes’.134

The Beschluß decisions thus seem ill suited to establish EU agencies, since they cannot cover the legal personality of agencies. The same applies to the executive agencies, since their legal personality is provided by the Council’s framework regula-tion. Lastly, it may be that the idea of the regulation resembling the law at national level deters EU decision- makers from establishing EU agencies without recourse to a regulation, even if such an idea has no immediate basis in EU constitutional law.135

1.3 Conclusion

From legal practice it could easily be deduced that the regulation is the sole instru-ment to establish EU agencies, but this did not answer the question of whether agency establishment requires, as a conditio sine qua non, the adoption of a regula-tion. It has been found that, in theory, this is not the case, since recourse to a dir-ective or a directive and a Beschluß decision could also be envisaged. The fact that this solution would be much more cumbersome, without any added value, means it remains only a theoretical option. In addition, the regulation is the preferred instrument because the establishing acts of EU agencies often deal with material rules as well, which cannot all be laid down in a directive or a decision. For these practical reasons, the EU institutions have found the regulation to be the best in-strument for agencifying the EU administration.

2 Limits Flowing from the Principle of Subsidiarity

A second important constitutional principle of the EU legal order is the principle of subsidiarity. The discussion of this principle follows that of the principle of con-ferred powers because it should first be determined whether the EU is competent to act, following which the subsidiarity principle answers the question of whether the EU should exercise that competence.136

133 See Articles 3(3) and 4(2) of Regulation (EC) 58/ 2003, OJ 2003 L 11/ 1.134 European Commission, COM (2000) 788 final, p 18 (emphasis added).135 Jürgen Bast, ‘On the Grammar of EU Law: Legal Instruments’, (2003) Jean Monnet

Working Papers 9, p 28.136 Renaud Dehousse, ‘Réflexions sur la naissance et l’évolution du principe de subsidiarité’, in

Delpérée (ed), Le principe de subsidiarité, Bruxelles, Bruylant, 2002, p 362.


Article 5(3) TEU explains that the principle is only relevant in so far as the EU’s competence is not exclusive. The subsidiarity test is then composed of two stages. The first is a national insufficiency test, the objectives of the proposed action not being sufficiently achievable by the Member States. The second is a comparative efficiency test, the objectives being better achievable at EU level.137 These two tests are a novelty introduced by the Treaty of Lisbon, since pre- Lisbon national insufficiency automatically resulted in (greater) comparative efficiency, whereas today a proposed action should cumulatively pass the two tests before the Union may act.138

According to Schütze, subsidiarity should not be confined to the question of whether the EU should act but should also address that of how the EU should act, since reducing subsidiarity to the ‘whether’ question would result in a dual fed-eralism, whereby either the EU or the Member States may exercise their compe-tences, instead of in a cooperative federalism.139 Under a cooperative federalism, subsidiarity would come down to federal proportionality whereby a check is kept on the EU, preventing it from unnecessarily restricting national autonomy.140 Subsidiarity as federal proportionality would indeed give (back) substantive meaning to the principle, reducing the discretion left to the EU legislator on how it should act. While interesting, this perspective should not obfuscate that sub-sidiarity, in EU law, is about more than safeguarding national autonomy. After all, the EU could in theory exercise a competence without unnecessarily restricting national autonomy while not meeting the national insufficiency or comparative efficiency test, in which case the EU action would still violate the subsidiarity principle. For the purposes of this chapter then, the ‘how’ question is brought under the proportionality principle.

Leaving further musings on the principle of subsidiarity to the specialized literature on this topic,141 it is useful to verify how the decision of establishing an EU agency and granting it powers should be assessed from a subsidiarity perspective.

137 Robert Schütze, From Dual to Cooperative Federalism: The Changing Structure of European Law, Oxford, OUP, 2009, p 250.

138 This raises the question of what should happen if only the first of these two tests is passed, leading Barents to conclude that the two tests in fact are the one and the same, the Lisbon Treaty not having changed the old test. See René Barents, Het Verdrag van Lissabon, Achtergronden en commentaar, Deventer, Kluwer, 2008, p 388.

139 In addition, Schütze notes that the ‘whether’ question is already answered by the principle of conferral which would nullify subsidiarity’s usefulness. That the Court, in its jurisprudence, has dealt with the ‘how’ question under the principle of proportionality then explains why challenges before the Court based on subsidiarity have, so far, been fruitless.

140 Schütze, n 137, pp 262– 3.141 See inter alia David Edward, ‘Subsidiarity as a Legal Concept’, in Cardonnel, Rosas, and

Wahl (eds), Constitutionalising the EU Judicial System— Essays in Honour of Pernilla Lindh, Oxford, Hart Publishing, 2012, pp 93– 103; Alexia Herwig, ‘Federalism, the EU and International Law: On the Possible (and necessary) Role of Subsidiarity in Legitimate Multilevel Trade Governance’, in Cloots, De Baere, and Sottiaux (eds), Federalism in the European Union, Oxford, Hart Publishing, 2012, pp 65– 82; Paul Craig, ‘Subsidiarity: A Political and Legal Analysis’, (2012) 50 JCMS S1, pp 72– 87.


In the section on the principle of conferred power, two issues in relation to agency creation were distinguished: (i) the competence to establish a new body and (ii) the tasks and powers exercised by such a body. The principle of subsidi-arity is only relevant as regards the second issue, then, since the competence to establish an EU agency necessarily is an exclusive EU competence— this even if the Commission is of the opinion that for example the competence to establish a European Public Prosecutor (from Eurojust) is not an exclusive competence by nature.142

2.1 The agency instrument under the subsidiarity principle

Before commenting on the agencies’ powers and tasks, the general agency instru-ment will first be scrutinized under a subsidiarity perspective. In the introductory chapter it was noted that agencies are a form of administrative integration,143 bringing together national administrations in a formalized structure, stream-lining national implementation of EU legislation. De Moor- van Vugt even spoke of Trojan horses whereby the process of further legislation and agencification ulti-mately negates national administrative autonomy. Such a representation of agen-cification would indeed have red subsidiarity lights flashing.

However, the question is best appreciated when national administrative au-tonomy is put into its broader context of the implementation of EU law. The rule of indirect administration may itself be regarded as inspired by subsidiarity— not so much because of a conscious choice, but in the first place because subsidi-arity is a rule of common sense.144 At the beginning of European integration in the EEC framework, it was a matter of common sense to rely on national ad-ministrations. The first great achievements related to the abolishment of customs duties, for which no new significant administrative structures were needed, or the coordination of existing (national) legislation in the field of social security, for which one evidently best relies on the national administrations already involved in social security. Unsurprisingly, the comitology system which acts as a check on the Commission’s power in direct administration originated in one of the first major ‘new’ EU policies, that is, the CAP.145

If agencies fall between direct and indirect administration, and if the latter is inspired by the principle of subsidiarity itself (cf. section I 2.1), how does this affect an appreciation of agencification from a subsidiarity perspective? At first sight agencification seems a move away from subsidiarity, since, as noted previously, it

142 See European Commission, COM (2013) 851 final, p 4. As discussed elsewhere, this quali-fication by the Commission is not subscribed to and seems foremost inspired by the fact that en-hanced cooperation between a limited number of Member States is only possible in areas of non- exclusive competences. See Chamon, n 11, p 279.

143 Because of the specific tasks which they exercise, this excludes those EU agencies identified as ‘service agencies’.

144 See Dehousse, n 136, p 361.145 Merijn Chamon, ‘Verzelfstandiging in de nationale en Europese rechtsordes: nieuwe uit-

daging van een meergelaagde administratie’, (2013) TBP 2– 3, p 115.


is a form of centralization. The typical form of EU centralization in the adminis-trative sphere, however, is direct administration. Through agencification, the EU legislator has therefore succeeded in escaping the Treaty’s dichotomic choice be-tween direct and indirect administration: hybrid EU bodies are established which are involved in complex composite procedures, linking national authorities with EU authorities.

This evolution, whereby the formal options provided for in the constitu-tional charter were found to be insufficient and not adapted to the needs on the ground, is surprisingly parallel to an evolution which may also be noted in the federal polity of the United States. As was noted previously, the US fed-eral system is not built on the idea of Vollzugsföderalismus and the Supreme Court has elaborated an anti- commandeering doctrine from the tenth amend-ment,146 whereas ‘commandeering’ is a basic feature of EU federalism.147 Subsidiarity has thus inspired the rule of indirect administration in the EU, whereas in the US similar considerations have inspired a rule prohibiting in-direct administration.148

In both polities, however, constitutional theory is lagging behind develop-ments on the ground. As Bendor and Farmer note, in the 1960s and 1970s the US federal government passed an unprecedented number of regulatory statutes without a similar increase in federal administration.149 As was noted previ-ously, following the SEA, the EU also saw a significant increase in legislative

146 The tenth amendment provides: ‘The powers not delegated to the United States by the Constitution, nor prohibited by it to the States, are reserved to the States respectively, or to the people.’ As such it is not merely concerned with the exercise of competences (as is the subsidi-arity principle) but by the existence of competences. However, just like the subsidiarity prin-ciple, it does aim to protect the lower entity’s autonomy against encroachment by the higher entity.

147 Daniel Halberstam, ‘Comparative Federalism and the Issue of Commandeering’, in Kalypso and Howse (eds), The Federal Vision: Legitimacy and Levels of Governance in the United States and the European Union, Oxford, OUP, 2001, pp 214– 15. According to Schütze, how-ever, commandeering is prohibited in the EU as well, since the EU may only give instructions to the Member States and not to individual national officers. See Robert Schütze, European Constitutional Law, Cambridge, CUP, 2012, p 253. Yet, such commandeering is different from the commandeering doctrine in the US, which prohibits the federal legislator from requiring the states to govern according to its instructions.

148 As was noted in the previous footnote these are not identical considerations. Commentators have indeed warned that the anti- commandeering doctrine could have perverse effects whereby Congress would be inclined to rely more on its power of preemption, completely pushing out state authority. Obviously such a solution would not be in line with the spirit of the subsidiarity principle. Bulman- Pozen and Gerken note that the dissenting Justices in Printz explicitly noted this issue and preferred commandeering to preemption because the former still leaves some power to the states. See Jessica Bulman- Pozen and Heather Gerken, ‘Uncooperative Federalism’, (2009) 118 The Yale Law Journal 7, p 1296. The risk that jurisprudence such as New York v. United States and Printz would lead to greater reliance on preemption also left Bermann concluding that the tenth amend-ment as interpreted in the former ruling did not fully reflect the subsidiarity principle: see George Bermann, ‘Subsidiarity as a Principle of U.S. Constitutional Law’, (1994) 42 American Journal of Comparative Law Supplement, p 560.

149 Josh Bendor and Miles Farmer, ‘Curing the Blind Spot in Administrative Law: A Federal Common Law Framework for State Agencies Implementing Cooperative Federalism Statutes’, (2013) 122 The Yale Law Journal 5, pp 1286– 7.


activity and a deepening of integration without a similar increase in resources being made available to the EU executive, resulting in an institutional def-icit.150 In the US, the Congress solved this problem by developing a form of cooperative federalism, which involves state agencies implementing federal policies. The advantages such a cooperative federalism brings151 would also fit well as justifications for the rule of indirect administration in the EU. The difference between this cooperative federalism and the (prohibited) comman-deering lies in the formally voluntary character of the states’ cooperation. The choice left to the states is that between (helping) regulate within the frame-work set by Congress— possibly in return for federal funding, called condi-tional funding— or facing preemption by federal legislation, also called con-ditional preemption.152

In a way US Congress’ increased reliance on cooperative federalism finds a parallel in the agencification process in the EU.153 The agencies’ involvement is necessary because relying on the Member States to implement EU law under in-direct administration has become increasingly inadequate to effectively deliver policy. From a subsidiarity perspective, the EU could then resort to direct ad-ministration, but this is only so because the Treaties only provide for two op-tions. Should the Treaties provide for more options for implementing EU law, the ‘national insufficiency’ and ‘comparative efficiency’ parts of the subsidiarity test would no longer necessarily point to direct administration. Through agencifica-tion, the legislator has created such an option, which is more efficient compared to purely indirect administration but also less intrusive on Member States’ autonomy than direct administration.

As a result, agencification serves the ethos of the subsidiarity principle since it restricts intervention by the higher entity,154 compared to such intervention under direct administration. As such, despite the fact that agencification amounts to centralization and despite concerns expressed about the restriction of national autonomy, the agency instrument should be seen as an instrument of subsidiarity. Whether every single power or task which has been granted to EU agencies would pass the subsidiarity test is a different matter, however, which will be addressed in general terms in the next section.

150 See Everson and Majone, n 476.151 Philip Weiser, ‘Cooperative Federalism and its Challenges’, (2003) Law Review of Michigan

State University Detroit College of Law 3, p 729.152 Seth Davis, ‘Conditional Preemption, Commandeering, and the Values of Cooperative

Federalism: An Analysis of Section 216 of Epact’, (2008) 108 CLRev 2, pp 412– 15. Of course, the question then is whether the choice offered to the states is a mere formal option or a genuine option left to the state (not) to assist in federal regulation. Davis refers to South Dakota v. Dole to argue that the Supreme Court indeed recognises the difference between a purely formal and a genuine choice.

153 Still, EU agencification as a subsidiarity- inspired alternative to direct administration is much more in line with the EU’s constitutional charter’s spirit compared to cooperative federalism under the US constitution, which is essentially based on a dual federalism.

154 Antonio Estella, The EU Principle of Subsidiarity and Its Critique, Oxford, OUP, 2002, p 104.


2.2 Individual tasks and powers of EU agencies under the subsidiarity principle

Previously (cf. section I 1.2.1.3) it was found to be more interesting to work out an instrumental classification of agency powers given the diverse tasks of any single agency. For the same reason it is difficult to speak of individual agencies meet-ing the subsidiarity test, and instead one should rather verify whether individual agencies’ powers meet this test.

Here the service agencies need to be distinguished from the other types of agencies identified in the first chapter. From the nature of these services, it follows that they should be provided centrally.155 Fischer- Appelt noted something similar for those (decision- making) agencies which grant authorizations in centralized procedures, since in those cases a single (EU) administration follows necessarily from a single harmonized procedure.156 The same applies to the inspection powers granted to EU agencies, in so far as the national authorities, rather than private operators, are being inspected. As such, these powers would automatically pass the subsidiarity test. This flows from the fundamental elements of the test itself, as was remarked in general by Van Nuffel, who noted that the Union legislator can legislate proactively, safeguarding its legislation from challenges based on the principle of subsidiarity, by setting the objectives of its proposed action at such a high standard that it becomes impossible for individual Member States to achieve them without action at Union level.157

However, some of the other powers granted to EU agencies have led to criti-cism based on subsidiarity grounds.158 Geradin and Petit have expressed doubts on those agencies which have been accorded the most modest powers, such as the precursor of the FRA and the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA). Had the European Institute for Gender Equality (EIGE) been established at the time of writing, they would undoubtedly have expressed serious doubts on that agency as well.

Geradin and Petit emphasize the economic logic behind subsidiarity and cen-tralization, referring to externalities and economies of scale,159 in light of the com-parative efficiency component in the subsidiarity test. Combining this insight with that of Van Nuffel, subsidiarity concerns will be the greatest for those agen-cies which have been granted modest powers in a field where economies of scale and externalities are difficult to identify. For Geradin and Petit, the racism issues

155 A question here may be whether the EU could not rely on the private sector to provide these services, but evidently this falls outside the scope of a subsidiarity perspective.

156 Fischer- Appelt, n 8, p 164.157 Piet Van Nuffel, ‘Gebruiksaanwijzing voor subsidiariteit— Een bijsluiter bij de eerste toe-

passing door het Hof van Justitie’, (1997) Rechtskundig Weekblad 9, p 290.158 Geradin and Petit, n 113, p 173.159 Referring to Roger Van den Bergh, ‘The Subsidiarity Principle in European Community

Law: Some Insights from Law and Economics’, (1994) 1 MJECL 3, p 341. See Geradin and Petit, n 113, pp 173– 4.


with which the FRA precursor dealt do not have a real EU dimension, and the same was true for the EMCDDA.

As regards the criticism that some issues should be tackled at a higher level than that of the EU, it should be noted that the principle of subsidiarity as laid down in the TEU only refers to one supranational level, that is, that of the Union. As a result, whereas the FRA may be problematic from a general subsidiarity perspec-tive, inter alia given the work of the Council of Europe, under the EU principle of subsidiarity this criticism seems more difficult to uphold.

Looking at the sub- EU level, the transformation of the European Monitoring Centre on Racism and Xenophobia (EUMC) into the FRA has alleviated some subsidiarity concerns, since the FRA generally devotes special attention to spill-over effects from EU policies. However, for some other issues on which the FRA works, the criticism voiced by Geradin and Petit still stands.160 For these issues, such as children’s rights, discrimination on the grounds of ethnicity or religious beliefs, access to efficient and independent justice, etc, it seems difficult to uphold the idea that the modest power of drawing up reports and gathering information cannot be achieved at national level but should instead be situated at EU level.

What is more, the EU legislator has imposed obligations on the Member States to establish national agencies in some of the same fields in which FRA is active, granting them more extensive powers than those granted to FRA.161 Article 28 of the Directive on data protection provides that each Member State shall establish a data protection authority which should be granted investigative powers and the power to start legal proceedings in cases of violation of data protection rules.162

How then may the gathering of information on these issues not be achieved by the Member States, and why should this instead be undertaken at EU level? The Commission, in the explanatory memorandum to its proposal, stressed that the FRA would help the EU itself ‘to fully respect fundamental rights in its action’.163 Given that the FRA does not only provide information to the EU institutions, the Commission added: ‘the Agency is designed to provide information, which enable the effectiveness of policies within and between the Member States to be assessed and thus adds value in terms of devising and targeting policies.’164 In its impact assessment, the Commission stressed the importance of comparable and reliable data between the Member States, but in its proposed policy options the Commission does not appear to have explored the possibility of harmon-izing data gathering at national level and establishing an EU agency which only monitors EU institutions, rather than EU institutions and Member States.165 The Commission’s discourse in its proposal and impact assessment shows not only the pro- active approach mentioned by Van Nuffel but also how difficult it is to object-ively assess the subsidiarity merits of a ‘proposed action’. Although the subsidiarity

160 See the issues discussed in the FRA Annual Report of 2011.161 See Article 13 of Directive (EC) 2000/ 43 of the Council, OJ 2000 L 180/ 22.162 See Directive (EC) 95/ 46 of the European Parliament and of the Council, OJ 1995 L 281/ 31.163 European Commission, COM (2005) 280 final, p 7. 164 Ibid, p 7.165 See European Commission, SEC (2005) 849.


procedure involving national parliaments established by the Treaty of Lisbon was obviously not in place when the proposal for the FRA was discussed, at the time at least one national chamber objected to the proposal on subsidiarity grounds.166

As was noted previously, whether the EIGE passes a subsidiarity test could be doubted as well, but this agency will be scrutinized in greater detail in the section on proportionality.

2.3 Conclusion

From the perspective of subsidiarity, the EU legislator’s reliance on the instrument of the EU agency is laudable. The brief comparison of how federal policy is imple-mented in the EU and the US has shown how both federal polities struggle with the challenge of effective implementation of federal legislation. In both polities, subsidiarity- inspired solutions were conceived which were not immediately fore-seen in the respective constitutional charters. Generally, agencification was then found to be in line with the ethos of the subsidiarity principle.

Turning to individual agencies, the establishment of an agency itself falls out-side the scope of the subsidiarity principle since the competence to do so is ex-clusive to the EU. In any case, the subsidiarity principle only makes a distinction between action at EU level and action at national or sub- national level. Once it is determined under a subsidiarity test that action at EU level is warranted, the question of who (at EU level) should act seems irrelevant.

On the other hand, the tasks exercised by EU agencies may be scrutinized. Here a general rule was identified: drawing on the general rule identified by Van Nuffel, the intensity of the powers granted to agencies stands in relation to the probability of passing the subsidiarity test. For agencies which have been granted only modest powers, it will be more difficult to show that the national insuffi-ciency and comparative efficiency tests are passed.

3 Limits Flowing from the Principle of Proportionality

Turning to proportionality, the function of this principle is often seen as pro-tecting the individual interests or rights of citizens or economic operators.167 However, the proportionality principle may also be extended to the entire EU legal order, becoming a genuine constitutional principle, prescribing that the means and the ends of government action need to be weighed against each other ‘in the sense that the means are only lawful if they bear an appropriate relation to the aim pursued’.168 According to Tridimas, the Maastricht Treaty, by virtue of

166 This was the Dutch Senate: see the motion proposed by MP C.D. Dees on 7 March 2006 and approved in plenary session on 14 March 2006, available at http:// www.eerstekamer.nl/ motie/ motie_ dees_ vvd_ c_ s_ inzake_ het_ niet/ document2/ f=/ w22112o.pdf.

167 See eg Hartley, n 21, pp 151– 2.168 Manfred Zuleeg, ‘The Advantages of the European Constitution’, in von Bogdandy and

Bast (eds), Principles of European Constitutional Law, München, Hart Publishing, 2010, p 774.

http://www.eerstekamer.nl/motie/motie_dees_vvd_c_s_inzake_het_niet/document2/f=/w22112o.pdf




Limits under the Proportionality Principle 165

current Article 5(4) TEU, indeed elevated proportionality to such a fundamental principle.169 Of the principle’s three primary functions identified by Tridimas, the focus in this section will be on the function by which it ‘governs the exercise by the [EU] of its legislative competence’.170

The discussion of the principle of proportionality logically follows those of the principles of conferral and subsidiarity. Together they form a trinity whereby in the first place, a competence to act is identified; secondly, the question is ans-wered whether the EU should act; and lastly the manner in which the EU acts is addressed.

The principles of proportionality and subsidiarity therefore govern the EU’s exercise of competences, albeit at a different level, whereby the principle of pro-portionality is also relevant when exclusive competences are exercised.171

Tridimas further notes that the principle regulates the extent of EU action, but does not limit the objectives or does not require a restrictive interpretation of that action.172 As a result, the principle leaves it to the political discretion of the legislator to pursue the objectives of the Union, only prescribing that the content and form of the measures adopted should be ‘appropriate’. As to the rationale of the principle in this function, Tridimas notes that it was included in the Treaty ‘primarily with a view to protecting the interests of the Member States’,173 as it seeks to limit the burdens imposed on them by the EU.174 In view of the previous observations on the EU’s legitimacy, this rationale should further be broadened. The principle should then be understood as part of the fabric of the EU legal order, which (partially) derives its legitimacy from the quality of its legislation, which includes the proportionality of such legislation.

Such is the background against which the provision in Article 5(4) TEU may be understood. The first sentence of that provision reads: ‘Under the principle of proportionality, the content and form of Union action shall not exceed what is necessary to achieve the objectives of the Treaties’, thereby clarifying that the principle of proportionality applies both to the content of the Union measure and its form.

Under the distinction between the competence of establishing an agency and empowering an agency, it may be noted that the principle of proportionality ap-plies to both the choice for an agency itself and the tasks and power granted to that agency. In theory the proportionality test is three- staged, whereby a measure is tested on its suitability, on its necessity, and then on its proportionality stricto sensu. As Tridimas notes, the proportionality principle allows the European judge

169 Takis Tridimas, The General Principles of EU Law, Oxford, OUP, 2006, p 138.170 The other two being proportionality as a ground of review for EU and national measures.

See ibid, p 139.171 Koen Lenaerts and Patrick van Ypersele, ‘Le principe de subsidiarité et son con-

texte: étude de l’article 3B du Traité CE’, (1994) 30 CDE 1– 2, p 62.172 Tridimas, n 169, p 176. Beck also qualifies proportionality as an indeterminate principle: see

Gunnar Beck, The Legal Reasoning of the Court of Justice of the EU, Oxford, Hart Publishing, 2013, pp 201– 2.

173 Tridimas, n 169, p 138. 174 Ibid, p 176.


to review not only the legality of a Union act but also, to a certain extent, its merits.175 Of course, whether the Union judge will use the principle as such also depends on the standard of review applied. This standard varies depending on the context and, if confronted with legislative measures involving economic policy choices, only a marginal review will be applied under established jurisprudence. A measure will then be found disproportionate if it is manifestly inappropriate with regard to the objectives which it pursues. The proportionality test is therefore both one of the most powerful as well as flexible tools available to the Union judge.

3.1 EU agencies’ tasks and powers under the principle of proportionality

Just like the discussion of the agencies’ tasks and powers under the subsidiarity principle, it would be unworkable in this section to comment on all the different tasks and powers exercised by agencies, even more so since proportionality is very context- specific.

What is important to point out, however, is that agencies are always established through formal secondary legislation, and generally in rather complex policy con-texts. As result, scrutinizing the proportionality of empowering agencies will typ-ically involve a mere ‘manifest inappropriateness’ test and thus only a marginal review. Indeed, in the ENISA case the Court clearly deferred to the legislator’s appraisal whereby the ENISA’s advisory powers were found suitable and necessary to bring about harmonization. Agencies usually also operate in technical areas, which would further deter the Court from zealously undertaking its proportion-ality review. In some cases the empowerment of agencies could also be qualified as urgent. This was most clearly the case for the ESAs (and the Single Resolution Board, or SRB) which have been empowered to tackle the negative effects of the financial crisis.

As regards the agencies’ powers, there is also an important difference between proportionality and subsidiarity. Under the latter, concerns lessen as agencies’ powers become greater, incentivizing the legislator to legislate proactively. Under proportionality, however, greater powers may be more problematic. As a result, whereas it is doubtful that modest powers meet the requirements of subsidiarity it will be less onerous to show their proportionality, a mere power of opinion being one of the least restrictive means of achieving any policy objective.

The conclusion of this section would be that there is no immediately apparent reason why EU agencies’ empowerment should receive special scrutiny under the principle of proportionality, the legislator’s decision to empower an agency not being fundamentally different from other measures which could be adopted. Just like other legislative measures in complex areas, the decision to empower an agency would further only be marginally reviewed by the Court. Yet this could be different if the larger context in which agencies are empowered is taken into

175 Ibid, p 140.


account. As was noted previously, agencies are increasingly involved in complex composite procedures which bring together different actors from different levels. Depending on how complex these procedures are, they may become dispropor-tionate if the added value which the agency’s involvement brings is minimal.

Here again a possible tension may be noted between the principles of subsidi-arity and proportionality. For instance, should a complex composite procedure involving a national administration, an EU agency, and the Commission be sim-plified by eliminating the role of the national administration, this would probably reduce complexity and costs, thus easing proportionality concerns. At the same time, however, such a move could raise subsidiarity concerns.

The difficult relationship between subsidiarity and proportionality may also be seen in Merlin’s discussion of the ESAs, when he contemplated the possibility of giving direct supervisory powers to the ESAs:

Faire passer l’essentiel de la supervision directe du secteur du niveau national au niveau européen aurait certes constitué une réponse aux difficultés de coopération observées entre superviseurs nationaux. Mais cela eût sans doute été une réponse disproportionnée, qui n’aurait pas pu se justifier au regard du critère de subsidiarité […] Pour être efficace, la supervision doit s’effectuer par des autorités qui sont proches de ces acteurs.176

Merlin seems to connect proportionality and subsidiarity, but both issues merit a separate analysis and should not be confused. As it is, it could well be argued that the exercise of direct supervisory powers at EU level may be proportionate and, if limited to those financial institutions which are of systemic importance, that it would also be in line with subsidiarity.177

3.2 EU agencies’ establishment under the principle of proportionality

Despite the critical remark at the end of the last section on the context in which agencies operate, the conclusion for that section may still be that the agencies’ em-powerment in general should not receive any special scrutiny under the principle of proportionality. This is different for the decision on establishing an agency. As noted previously, this decision fell outside the scope of the subsidiarity principle because the competence to establish a new EU agency is exclusive to the EU.

An interesting preliminary step in such an assessment is answering the question on how the decision to establish an agency should be understood and conceptual-ized. In the introductory chapter it was noted that a reference to the tasks of an EU

176 Martin Merlin, ‘Le nouveau système européen de supervision financière’, (2011) RDUE 1, p 25.

177 In its proposal for a Single Supervisory Mechanism, however, the Commission proposed to grant the ECB the power to supervise all credit institutions in the Eurozone. See European Commission, COM (2012) 511 final. In the end, the ECB’s supervision powers now depend on the size of the supervised entity, pursuant to Article 6 (4) of Regulation (EU) 1024/ 2013 of the Council, OJ 2013 L 287/ 63.


agency would not be included as one of the elements of a definition, since it would be pointless to establish an agency without at the same time granting it tasks.178 This idea could also be identified in the opinion of AG Kokott in the ENISA case (cf. section IV 1.1.3.1). An EU agency then is nothing more than an extraordinary organizational form, created to allow the functions (of that agency) to be exercised properly. If the agency were to be seen as an end in itself, the repercussion would be that the principle of proportionality is useless, since there would only be one way to attain the objective of establishing an agency, that is, by establishing an agency.

Seen from this perspective, a proportionality analysis would enquire whether that organizational form is suitable, necessary, and proportionate stricto sensu to fulfil the functions which the agency would exercise. The question on the neces-sity of establishing an agency was already briefly touched upon previously (see section IV 1.1) when the repercussions of a more strict approach to using the flexibility clause of Article 352 TFEU and the doctrine of implied powers were discussed.

The decision to establish an agency may thus only be dissociated in theory from the decision to empower an agency, since the former decision is intrinsically linked to the latter. As a result, an analysis of the former decision’s proportion-ality should be done in the light of the latter, meaning the EU legislator should be able to justify (i) why a given set of tasks or functions necessarily require the establishment of an agency, (ii) that an agency is a suitable instrument to exercise those tasks and functions, and (iii) that there are no other ‘lighter’ organizational arrangements which would allow the same tasks and functions to be executed to the same standard.

In practice, the second requirement will be the least problematic. However, the first and third requirements are a different matter. Typically, the preambles of the acts setting up agencies and even the Commission’s proposals do not really ex-plain why an agency would be necessary compared to other alternatives. On this issue, the Common Approach seems to bring a relevant new element, since one of its very first provisions reads: ‘The decision to create a new agency should be based on objective impact assessments of all relevant options.’

In reality this was again a codification of existing practice, since detailed impact assessments have accompanied the Commission’s major legislative proposals since the turn of the century, with the Commission’s White Paper on Governance also devoting special attention to better rule- or law- making.179 A key element in such ‘Better Regulation’ was a more prominent role for the impact assessment (IA) prepared by the Commission for its legislative proposals.180

The IA accompanying the proposal for the ACER, for instance, was quite exten-sive on the issue of why an agency was necessary to ensure a sufficient level of coordination.181 The IA for the eu- LISA was even more elaborate on this issue, since

178 See section I 1.1.179 See European Commission, COM (2002) 275 final.180 Craig Robertson, ‘Impact Assessment in the European Union’, (2008) EIPA Scope 2, p 17.181 See section 5.3 of European Commission, SEC (2007) 1179 final.


the problem was also more focused.182 The three databases (SIS II, Eurodac, VIS) needed to be managed in any event and the sole major issue to be decided upon was which organizational arrangement should be worked out to do so: management by the Commission, by a single Member State, by Frontex, by Europol, by a new agency, etc. Yet, however elaborate the assessment, it remains difficult to quantify such a com-parison on all aspects, and in the IA for the eu- LISA proposal the Commission indeed admitted itself that the choice for a new agency as the preferred option was based on a qualitative assessment.183 In the preparation of its proposals for new legislation on medical devices, the Commission rejected the policy option of a new agency in its IA since such a (small) agency would not reach the required critical mass, which the Commission set at an agency staffed by a minimum of 100 functionaries.184

From these IAs one would conclude that the Commission clearly takes into account the proportionality of establishing a new agency whenever it submits a proposal to this end. But this does not always seem to be the case.

As regards the FRA, for instance, the Commission commences its IA by refer-ring to the impetus of the European Council, which concluded in 1999 that ‘the question of the advisability of setting up a Union agency for human rights and democracy should be considered’.185 Later, it noted that the Member States had ‘agreed to build upon the existing European Monitoring Centre on Racism and Xenophobia and to extend its mandate to make it a Human Rights Agency to that effect’.186 As a result, apart from the status quo option, all four other options considered by the Commission envisaged the establishment of an agency.187 Since it was not an institution until the Lisbon Treaty, the European Council might not have felt itself bound by the principle,188 but the Commission could not simply refer to the European Council’s ‘decision’ to change the EUMC into the FRA without itself entering into such fundamental questions as to the EU’s compe-tence in this field and the proportionality and subsidiarity of such a measure.

A similar problem plagues the IA report for the EIGE proposal, where the Commission referred to a consensus among stakeholders on the need for an au-tonomous EU body responsible for gender equality. As a result, the Commission only worked out three policy options, one being the status quo option and the other two foreseeing empowering an agency. The only time the Commission (im-plicitly) referred to the proportionality principle was in its justification for opting for a separate agency instead of entrusting gender issues to an existing one. The Commission thus remarked:

The establishment of a separate Institute for Gender Equality will have some financial implications […] However, any financial savings that might result from the inclusion of

182 See European Commission, SEC (2009) 837. 183 Ibid, p 43.184 European Commission, SWD (2012) 273 final, p 63.185 Council of the European Union, Presidency Conclusions of the Köln European Council (3 and

4 June 1999), 150/ 99 REV 1, p 18 para 46.186 Council of the European Union, Doc 5381/ 04, p 27.187 European Commission, SEC (2005) 849, pp 12– 13.188 The protocol on proportionality and subsidiarity only applying to ‘the institutions’.


gender equality within the scope of the future Fundamental Rights Agency or an existing agency would be outweighed by the disadvantages highlighted above.189

The Commission thus simply posited, without any quantification, that establish-ing the EIGE would be proportionate because it would deliver better results to the effect that its higher cost would be offset.

Comparing the IA reports for the FRA and the EIGE with those of the eu- LISA and the ACER shows how the approach in the assessments may vary. Taking a broader perspective, this could undermine the credibility of the IA process itself as it might give the impression that an IA is drafted in function of a pre- established policy option, rather than being a tool to identify the best policy option.

This brings us to a second issue. Even if the Commission devotes special at-tention to the question of proportionality in its IA, does this constitute proof of the proportionality of the establishment of an agency? Popelier has, in gen-eral, argued for more rigorous scrutiny by the Court of Justice of legislation in the light of the principles of governance and Better Regulation.190 While an IA itself is not a reviewable act, challenging the legislative act for which an AI was undertaken is a different matter. As Alemanno explains, the self- binding effect of the soft law norms which underpin an AI could be invoked,191 or the Court could use the IA when it assesses challenges based on proportionality.192 In this regard, Alemanno draws attention to Spain v. Council, in which the Court struck down a Regulation in the CAP related to cotton production because the Commission had not conducted an IA, unlike in its proposal for reforming the tobacco market.193

In the follow- up case where a number of cotton producers filed for damages, the General Court remarked that ‘it was not the contested provisions themselves, but the failure to take account of all the relevant factors and circumstances, in particular by carrying out a study of the reform’s impact, before their adop-tion which was criticised from the point of view of an infringement of the prin-ciple of proportionality’.194 This would suggest that once the Commission does undertake such a study or IA, the Court would be inclined to grant deference to the Commission’s assessment. According to Alemanno exactly this happened in the Vodafone case, where the Court was asked to assess the validity of the roaming Regulation195 in the light of the proportionality principle.196 The Court

189 European Commission, SEC (2005) 328, p 13.190 Patricia Popelier, ‘Governance and Better Regulation: Dealing with the Legitimacy

Paradox’, (2011) 17 EPL 3, p 565 et seq.191 See for instance the interinstitutional Common Approach on Impact Assessment of

November 2005; European Commission, SEC (2009) 92.192 Alberto Alemanno, ‘The Better Regulation Initiative at the Judicial Gate: A Trojan Horse

within the Commission’s Walls or the Way Forward?’, (2009) 15 ELJ 3, p 396.193 Case C- 310/ 04, Spain v. Council, [2006] ECR I- 7285, paras 103– 4.194 Joined Cases T- 252/ 07, 271/ 07 & 272/ 07, Sungro, SA e.a. v. Council and Commission, [2010]

ECR II- 55, para 60.195 Regulation (EC) 717/ 2007 of the European Parliament and of the Council, OJ 2007 L 171/ 32.196 See Alberto Alemanno, ‘A Meeting of Minds on Impact Assessment’, (2011) 17 EPL 3,

pp 485– 505.


started by observing that ‘before it drafted the proposal for the regulation, the Commission carried out an exhaustive study, the result of which is summarised in the impact assessment […] It follows that the Commission examined various options […]’.197

The decision on the organizational arrangement, while at first sight perhaps rather straightforward, in reality may turn out to be rather complex, as the IAs prepared for agencies such as the ACER and eu- LISA show. As a result, this de-cision would merit the same degree of judicial deference as the decision on the material rules. From the perspective of proportionality, the decision to establish an agency would then remain safely in the hands of the political actors and the Court’s role would be limited to ensuring that the decision was preceded by a proper assessment of all relevant elements.

This would in turn mean that the role of the political institutions in control-ling each other becomes even more important. At times, there may be calls for the establishment of a new agency, possibly in response to a mediatized event or policy crisis, but this should not deter the Commission from rigorously scrutin-izing the policy option of establishing an agency on its merits. Since it is the in-stitution holding the power of legislative proposal, a special responsibility falls to the Commission, but of course this responsibility also extends to the two branches of the legislator.

3.3 The proportionality of agencification in legislative practice

Legislative practice from the past suggests that a close scrutiny of proposals in the light of the proportionality principle has not always been a priority of the institutions.198 Still, Paradis observes that the Commission’s IAs have brought transparency on the justifications for the latest agencies’ establishment.199 At the same time, Paradis draws attention to a number of problems which have persisted, such as the unexplored possibility of merging or closing down agencies and the dominant model of the agencies’ Boards which he describes as costly, ineffective, and unnecessary.200

Establishing an agency is a significant decision to which a number of important fixed costs are attached. From a proportionality perspective it is important that the agency makes up the extra total costs, by providing significant benefits in terms of better or more efficient (implementation of) EU law. Logically, this chal-lenge will be more daunting for the agencies of limited scope and with limited powers.201

197 Case C- 58/ 08, Vodafone Ltd e.a., [2010] ECR I- 4999, para 55.198 The FRA and EIGE cases discussed previously are in themselves good examples of this.199 Eric Paradis, ‘Les agences et organismes divers de l’union européenne: une apprécia-

tion d’ensemble en 2010’, in Molinier (ed), Les agences de l’union européenne, Bruxelles, Bruylant, 2011, p 209.

200 Ibid.201 Logically assuming there is an inverse relation between having limited powers and the cap-

acity of unlocking significant benefits.


The fixed costs will then also take up an increasing share of the total costs the smaller an agency is, or, as the Parliament noted in a 2009 study: ‘small agen-cies have to devote more resources to administrative tasks than large agencies, whereas larger agencies benefit from scale effects.’202 The nature of the fixed costs is quite diverse. Among these one can think of the separate seat of the agency, 203 and further additional costs may result from the need to establish li-aison offices in Brussels.204 Agencies further require Boards composed of at least one representative for each Member State, a Director and staff. Further, a great deal of administrative requirements imposed on the institutions are also im-posed on EU agencies. Given their separate legal personality, EU agencies will also have to conclude their own service- level agreements with service provid-ers, including with the Commission itself.205 An agency’s existence also creates extra costs for the institutions themselves. For instance, given their budgetary autonomy, new agencies result in new separate procedures for the budgetary authority and new reporting requirements for the Court of Auditors.206 As a result, ‘[g] iven the limited share of the EU Budget represented by the agencies [sic] combined expenditure, the audits place a disproportionate burden on the Court’s resources’.207

The problems which a small agency will encounter in this respect were also noted by the Parliament in 2011 in relation to the European Police College (CEPOL), when it observed an ‘impossibility to apply the complex EU finan-cial regulations and Staff Regulations, given the limited mission and correspond-ing small size of the College’.208 As regards the Staff Regulations, Pradines and Feugier have noted that they are not adapted on all points to the particularity of the agencies.209

202 Directorate- General for Internal Policies, 7 April 2009, ‘Study on the Opportunity and Feasibility of Establishing Common Support Services for EU Agencies’, PE 411.264, p 20.

203 See for instance the study commissioned by the Parliamentary Committee on Budgets on the agencies’ buildings: Policy Department on Budgetary Affairs, 29 April 2009, ‘Agencies’ Buildings’, PE 392.972.

204 In 2010 some seven out of twenty- nine agencies had such a liaison office in Brussels. See Answer of the European Commission to Parliamentary Question (E- 0492/ 10) by MEP Ingeborg Gräßle.

205 Kilb mentions this cost for the executive agencies, but they also apply to the decentral-ized agencies. See Wolfgang Kilb, ‘Europäische Agenturen und ihr Personal— die großen Unbekannten?’, (2006) 17 EuZW 9, p 273. See also interinstitutional Working Group on regula-tory agencies, Common administrative support for agencies, 2010, Analytical Fiche Nr° 16, p 3.

206 In 1996, the Parliament noted that the Court suffered from a shortage of auditors because of ‘the growing volume of work subsequent to the audits and annual reports to be produced for each of the agencies of the European Union’. See Resolution of the European Parliament, OJ 1996 C 347/ 135.

207 Friedemann Zippel, ‘The Mushrooming of European Agencies and Other Decentralised Bodies’, (2013) Journal Cour des comptes européenne 11, p 22.

208 See para 39 of the Resolution of the European Parliament, OJ 2011 L 250/ 261.209 See Marie- Hélène Pradines and Jean- Luc Feugier, ‘Le cas des agences européennes’, in

Le Theule and Leprêtre (eds), La fonction public européenne, Strasbourg, ENA, 2012, pp 128– 33. For an earlier discussion of this problem, see Wolfgang Kilb, ‘Europäische Agenturen und ihr Personal— die großen Unbekannten?’, (2006) 17 EuZW 9, pp 272– 3.


Among the fixed costs, those related to providing information to the public should also be taken into account. Article 41 of the Charter of Fundamental Rights for instance deals with the right to good administration and provides that ‘[e] very person may write to the institutions of the Union in one of the languages of the Treaties and must have an answer in the same language’. The reference to the institutions is interesting, since in other provisions of the Charter reference is made to ‘the institutions, bodies, offices and agencies of the Union’. In the same way, Article 24 TFEU provides that:

[e] very citizen of the Union may write to any of the institutions or bodies referred to in this Article [ie Parliament, Council and Ombudsman] or in Article 13 of the Treaty on European Union in one of the languages mentioned in Article 55(1) of the Treaty on European Union and have an answer in the same language.

Under primary law, an EU body, office, or agency therefore need not reply in the same language as the one in which it was addressed.210 In fact, this issue was brought before the Ombudsman when a Polish citizen lodged a complaint be-cause the Office for Harmonisation in the Internal Market (OHIM) replied in English to his e- mail in Polish. The OHIM tried to defend its policy by referring to the costs of providing information in all the official EU languages, despite it being one of the biggest EU agencies. The Ombudsman did not agree with the OHIM and recalled that ‘good administration requires that, as far as possible, the institutions, bodies, offices and agencies of the EU should provide information to citizens in their own languages’.211 By arguing that the OHIM should do so ‘as far as possible’, the Ombudsman acknowledged that practical and financial reasons could justify a limit on the equality of languages. In the light of the principle of proportionality this is an interesting issue, as one could argue that it is dispropor-tionate to require EU agencies to treat all official languages equally. However, this in turn raises the question of whether the legislator acts in conformity with the principle of proportionality by establishing small agencies.

In light of all these significant fixed costs, establishing an agency only makes sense from a proportionality perspective if those costs can be offset through sig-nificant added value delivered by the agency. However, that added value would seem to depend on whether significant powers were also accorded to the agency, whereas it was found earlier that proportionality seems to require the legislation to grant only modest powers to agencies.

3.4 Conclusion

Because of the complex and technical areas in which EU agencies are usually em-powered and because the choice for an EU agency as the preferred organizational

210 This would also mean that the institutions can resort to EU agencies to bypass the more stringent language requirements.

211 See Decision of the European Ombudsman closing his inquiry into complaint 2413/ 2010/ MHZ against the OHIM.


‘vehicle’ also necessitates a rather complex assessment, scrutinizing the propor-tionality of these decisions seems an affair for the political institutions rather than for the Court. While the latter has jurisdiction, it would in most cases only exer-cise a marginal review.

That the political institutions have so far not devoted enough attention to this issue was already apparent from the ad hoc way in which agencification has developed. The lack of a clear horizontal framework has almost inevitably led to the disproportionality of the process of agencification. Some of the decisions to establish agencies may further also be questioned, especially in the case of smaller agencies and where different agencies operate in the same policy fields. On the other hand, no special appraisal seems to be in order for decisions em-powering agencies, since it does not appear relevant who is made responsible at EU level to exercise a certain task, only how this is done. In this respect it was noted that increasingly complex procedures, through the involvement of inter alia agencies, could go against the need for administrative simplification. And where this added complexity does not come with added value, such procedures could be disproportionate.

Throughout the preceding sections, it was also noted that the analyses of the same issue from the perspectives of subsidiarity and proportionality will not nec-essarily point in the same direction. Obviously this complicates the work of the political institutions when they aim to balance subsidiarity and proportionality requirements but it does not relieve them from the obligation to look into these issues, which, up until now, they have rather failed to do.

4 Concluding Remarks: Limits Flowing from the General Principles of EU Law

The principles of conferred powers, subsidiarity, and proportionality determine whether, when, and how the EU may act. In the case of the agencies, it does not seem that these principles have truly governed the EU legislator’s actions. As re-gards the principle of conferred powers, today almost any legal basis would seem adequate to establish an agency, and legal doctrine has progressively followed in-stitutional practice on this point. From the ENISA case it further seems that the Court leaves it to the legislator to determine whether an agency is the proper org-anizational form to promote whatever material rules it lays down.

As regards proportionality and subsidiarity there are some conflicting tensions, since the hybrid nature of agencies means they are a subsidiarity- friendly alterna-tive to simple direct administration. However, if only modest powers are granted to agencies the efficiency advantage of the EU level may be questioned under a subsidiarity perspective. This is different for the principle of proportionality, where modest, less intrusive powers will be less problematic. However, the pro-portionality principle, unlike the subsidiarity principle, is also relevant to the dec-ision on the establishment (rather than the empowerment) of an agency. Here it should be noted that an agency is a ‘heavy’ organizational form and that it appears

Meroni and the Notion of Delegation 175

disproportionate to establish small agencies (with modest powers) that cannot make up the fixed costs associated with the establishment of the body as a separate agency, revealing also a tension within proportionality.

Subsidiarity is a political principle rather than a judicially enforceable prin-ciple, and although the principle of proportionality is a potent instrument in the hands of the EU judge, it could be argued that both the decisions to establish and to empower an agency depend on complex economic assessments, restricting the Court’s scrutiny.

The result is that for all three of these fundamental principles, the Court’s scru-tiny has been or would be rather marginal. This evidently puts an even greater responsibility on the shoulders of the political institutions to exercise both self- restraint and scrutiny vis- à- vis each other to guarantee that these principles remain to be respected. Here, however, the political institutions seem to have failed.

5 The Meroni Jurisprudence and the Notion of Delegation

One does not have to be very familiar with the topic of European agencies to have noticed that the European institutions themselves and virtually every academic writing on agencies mention the Meroni ruling, delivered by the Court in 1958 under the ECSC (European Coal and Steel Community) Treaty. Generally, a Meroni doctrine is then relied on to define the limits to agencification (cf. section IV 5.3.1). Despite virtually everyone relying on Meroni,212 there is no consensus on its actual meaning and how it affects the process of agencification. This may (at least in part) be explained by the apparent lack of original analysis of this judgment.

For this reason, the present section will analyse the Meroni ruling itself, the context in which Meroni was ruled, and the opinion of AG Roemer. This should allow a proper understanding of the motives of the Court to rule as it did, which in turn will inform the assessment of the different contemporary interpretations of the ruling.

Since more than a generation has passed since the verdict in Meroni, the ana-lysis of the scholars of that time will be presented and analysed in a separate sec-tion from the present- day legal opinions. Although the importance of Meroni for the legal issue of delegation was already evident to the previous authors, once the

212 In the framework of this research no contemporary author could be identified who princi-pally rejects Meroni’s applicability to the process of agencification, although Chiti has claimed that ‘the Community legitimacy of the setting up of the European agencies should not be assessed in the light of the Meroni doctrine’. See Edoardo Chiti, ‘Beyond Meroni: The Community Legitimacy of the Provisions establishing the European Agencies’, in Della Cananea (ed), European Regulatory Agencies, Paris, Éditions Rive Droite, 2005, p 80. Subsequently, possibly because of the Tralli and Alliance for Natural Health cases (discussed later), he took a more moderate view, merely questioning Meroni’s relevance for agencies: see Edoardo Chiti, ‘An Important Part of the EU’s Institutional Machinery: Features, Problems and Perspectives of European Agencies’, (2009) 46 CMLRev 5, pp 1420– 4.


process of agencification was kicked off, the specific relevance of Meroni for this process started influencing the interpretation of that jurisprudence. Further, the positions of the three political institutions will also be analysed to try and identify an EU Meroni doctrine.

After presenting the status questionis before the recent Short- selling case, an at-tempt will be made to reconstruct a Meroni doctrine based on the actual ruling. This pure Meroni doctrine will then be evaluated to determine (i) whether and (ii) how it may be transposed to the EU context. Since Meroni has not lain dormant since 1958, the Court’s subsequent jurisprudence where reference is made to Meroni will also be taken into account. The debate on Meroni has of course also been significantly affected by the Short- selling case, in which Meroni was explicitly raised for the first time to assess the empowerment of an EU agency. A contemporary understanding of Meroni will therefore also depend on Short- selling.

5.1 The original Meroni ruling

The original Meroni ruling should be situated in the context of the High Authority’s efforts in the early days of the Common Market in steel to equalize the prices of imported scrap metal and scrap metal originating from within the Community.213 Article 53 ECSC provided that the High Authority could auth-orize (under Article 53(a) ECSC) or institute (under Article 53(b) ECSC) finan-cial mechanisms for the accomplishment of the ECSC’s objectives. To this end the High Authority had adopted Decision 22/ 54 instituting such a mechanism.214 Article 1(2) of the Decision entrusted the execution of the scheme to the Office and the Caisse, which had their respective seats in Brussels (hence the so- called Brussels Agencies). Articles 2 and 3 of the Decision provided that all undertak-ings using ferrous scrap had to pay contributions which would be determined by the Caisse. Article 4 determined that in case of defaults or late payments the Caisse would ask the High Authority to adopt an enforceable decision vis- à- vis the undertaking concerned. Articles 8 and 9 of the Decision provided that the High Authority would have an observer on the boards of both Brussels Agencies and that this observer could reserve all decisions to the approval of the High Authority.

213 As Diebold explains, the main problem lay in Italy’s lack of natural resources, which made it heavily dependent upon imported scrap for its steel production. Making steel from scrap iron re-quires less coal than making steel from crude iron, making it interesting for countries such as Italy to resort to scrap iron. The establishment of the Common Market would have resulted in Italian undertakings diverting their sources for scrap metal to the northern ECSC members, causing price hikes in those countries and giving incentives to speculators and the formation of buyer cartels. See William Diebold, The Schuman Plan: A Study in Economic Cooperation 1950– 1959, New York, Praeger, 1959, p 291. For a further economic analysis of the price equalization scheme for scrap metal see Louis Lister, Europe’s Coal and Steel Community: An Experiment in Economic Union, New York, Twentieth Century Fund, 1960, pp 64– 7.

214 Décision (CECA) 22/ 54 de la Haute Autorité, J.O. 1954 4 286.


The choice to entrust the management of the financial mechanism to these two bodies was the logical result of the drafting of Article 53 ECSC. Before adopting Decision 22/ 54, based on Article 53(b), the High Authority had previously auth-orized through Decision 33/ 53215 a financial mechanism for the same purposes under Article 53(a). A number of enterprises in the common steel market had set up the Office and the Caisse as cooperative societies under Belgian law to this end.216 However, this form of self- regulation had not performed adequately,217 forcing the High Authority to intervene and to acquire greater control over the functioning of the Office and the Caisse.218 The result was the mandatory financial arrangement under Decision 22/ 54 obliging all undertakings to participate in the scheme.219

5.1.1 The facts and law in the Meroni caseIn his opinion AG Roemer set out the different stages which culminated in the individual decision of imposing an individual contribution on an undertaking.220 The first stage was the general decision of the High Authority.221 The second stage involved the Caisse fixing the ‘contribution rate’ and the third stage involved the Caisse determining the contribution for each undertaking. According to the facts in Case 9/ 56,222 Meroni, a steel producer, had refused to pay the contribution calculated by the Caisse. The latter had therefore requested the High Authority to force Meroni to pay the contribution, which the High Authority did on 24 October 1956. According to Meroni this decision was vitiated by several proced-ural irregularities, inter alia an inadequate statement of reasons and no possibility for the undertaking concerned to submit its observations to the High Authority. As regard the decisions of the Caisse, Meroni argued that these were inadequately reasoned, that the Caisse did not properly and timely inform the undertakings,

215 Décision (CECA) 33/ 53 de la Haute Autorité, J.O. 1953 8 137.216 The statutes of these two bodies were published in the Belgian Official Journal: see Annexe

au Moniteur Belge/ Bijlage tot het Belgisch Staatsblad of 18- 19 May 1953, nos 11164 & 11165, pp 3113– 17.

217 For the reasons, see Dirk Spierenburg and Raymond Poidevin, Histoire de la Haute au-torité de la Communauté européenne du charbon et de l’acier: une expérience supranationale, Bruxelles, Bruylant, 1993, pp 161– 3.

218 See the changes to the statutes of the Brussels Agencies required by Décision 33/ 53.219 This did not mean that all undertakings were under an obligation to become a member of

the Office or Caisse, but that they were under an obligation to pay contributions on the use of scrap220 Opinion of AG Roemer in Cases 9/ 56 & 10/ 56, Meroni & Co. v. High Authority, [1957- 1958]

ECR 177, p 181.221 Apart from the already mentioned Decision 22/ 54 the High Authority had also adopted a

general Decision 14/ 55 which further strengthened the High Authority’s position in the Brussels Agencies. See Décision (CECA) 14/ 55 de la Haute Autorité, J.O. 1955 8 685.

222 Reference is often made to the Meroni cases, 9/ 56 as well as 10/ 56. Both cases involved the same issues, except for the problem of the Caisse relying on its own assessment of an undertaking’s use of scrap metal, which only figured in case 9/ 56. Since this issue led the Court to rule that a delegation cannot be presumed, which is part of the Meroni doctrine (discussed later), only case 9/ 56 will be referred to in the present study.


that the Caisse committed errors of fact in determining the contribution rate, and that the Caisse could not have based its decision on its own assessment of the purchases of scrap metal made by Meroni.

Meroni did not raise the issue of the delegation of powers from the High Authority to the Caisse as a formal plea in law, but it did note that ‘in the mind of the High Authority the Brussels accounts are unassailable and almost sacrosanct and are certainly of greater weight and authority than are decisions proper, which can always be contested before the Court of Justice’.223 The Court and the AG therefore also scrutinized the delegation of powers.

5.1.2 The legal problem posed by the delegation at issueLooking back with the benefit of hindsight, one would be inclined to say that the delegation was bound to be found problematic. However, the overview of the scheme presented above and further scrutiny of the general decisions and the statutes of the Brussels Agencies does not readily show any problematic delegation of powers.

Article 1(2) of Decision 22/ 54 provided that the functioning of the mech-anism was entrusted to the Brussels Agencies under the responsibility of the High Authority.224 Article 8 of Decision 22/ 54 also provided that the High Authority’s observer would attend all meetings of both the Boards and the General Assemblies of the Brussels Agencies, and further provided that he ‘should immediately transmit to the High Authority the deliberations of these bodies and inform the High Authority of any question which requires its con-sideration pursuant to Article 9’. That latter Article provided that the observer could reserve all final decisions to the High Authority and that in the absence of unanimity, any issue would be decided by the High Authority. In 1955 the statutes of the Office and Caisse were amended and new (identical) Articles 11 and 14 were incorporated, providing that in the absence of the High Authority’s observer, all decisions of the Boards should be submitted to the High Authority, who could still amend the decisions.225 The preamble to Decision 14/ 55 fur-ther provided and confirmed that the High Authority was responsible for the financial mechanism and that it should therefore be able to intervene in every circumstance.226

Taking these different provisions together, it indeed seems that the scheme, the actual day- to- day operation of which was left to the Brussels Agencies, remained under the responsibility of the High Authority, inter alia because the High Authority retained a form of hierarchical control over the Brussels Agencies.

223 See Case 9/ 56, Meroni & Co. v. High Authority, [1957- 1958] ECR 133, p 146.224 Décision (CECA) 22/ 54, J.O. 1954 4 286.225 See Annexe au Moniteur Belge/ Bijlage tot het Belgisch Staatsblad of 29 July 1955, nos.

21688 & 21691, p 1573 and 1578.226 Décision (CECA) 14/ 55, J.O. 1955 8 685.


5.1.3 The opinion of AG RoemerThe AG dealt with the different pleas of the applicant along the three stages he identified earlier (cf. section IV 5.1.1).

As regards the enforceable decision of the High Authority, the AG concluded that the decision was indeed insufficiently reasoned. In its decision, the High Authority had only ordered Meroni to pay the contribution, without setting out the method of calculating the contribution or how the total sum was composed. The AG further emphasized that under the Treaty the High Authority was always under an obligation to motivate its acts itself and that it therefore could not hide behind the Caisse.

As regards the second stage, the AG concluded that fixing the contribution rate, a competence of the Caisse, was indeed the cornerstone of the whole equalization scheme. The AG found that the decisions of the Caisse did not meet the standards which were mandatory, under the Treaty, for those of the High Authority. The AG then changed the approach by asking the question of whether the High Authority could have delegated such a power to the Caisse, noting that ‘the decisive element is whether the guarantees of legal protection to be found in the Treaty also exist in the case of a delegation of powers’227 and finding that the High Authority had ignored these guarantees.228

Another delegation issue vitiating the general decision of the High Authority was the possibility for the Caisse to base its decisions on its own assessment of an undertaking’s use of scrap metal. Here the AG noted that the procedure followed by the Caisse to make determinations on its own assessment was not provided for under the High Authority’s general decision, and that it was also irregular.229

Concerning the first stage, that is, the legality of the High Authority’s general decisions, the AG was very brief— rejecting Meroni’s pleas or not analysing them in depth— since he had already found that the decisions were irregular because of the delegation they contained allowing the illegal decision of the Caisse (under the second stage) to be adopted.

5.1.4 The ruling of the CourtThe Court ended up with much the same conclusion as the AG by taking a different route.

5.1.4.1 The failure to state reasons and the Caisse’s own assessmentsAs regards Meroni’s pleas related to the failure to state reasons, the Court found that the High Authority had indeed insufficiently reasoned its enforceable dec-ision and that the reasoning of the Caisse’s decisions could not cover this ir-regularity.230 The Court ruled that the High Authority’s decision should also be

227 Opinion of AG Roemer in Cases 9/ 56 and 10/ 56, n 220, p 194. 228 Ibid, p 196.229 Ibid, p 191. 230 Case 9/ 56, n 223, p 142.


annulled because it did not mention that the sum claimed was based on the Caisse’s own assessment of Meroni’s use of scrap metal and because no legal base was men-tioned giving the Caisse this competence.231 The Court also found Meroni’s pleas concerning the lack of information from the High Authority or Caisse related to the methods used to calculate the contributions well founded.232

5.1.4.2 The question of delegationIn its more modern jurisprudence, the Court often does not enter into all of the grounds submitted by the applicant if it reaches the conclusion that at least one of them is successful, but in Meroni the Court continued its scrutiny. Perhaps the Court found itself obliged to enter into the issue of the delegation of powers by the High Authority to the Caisse, and it did so by citing Meroni in its application and reply:

[T] he applicant complains that the High Authority has delegated to the Brussels agencies powers conferred upon it by the Treaty, without subjecting their exercise to the conditions which the Treaty would have required if those powers had been exercised directly by it.

The applicant also complains that the High Authority has created ‘a situation in which the large- and medium- sized industries predominate over those of limited financial means, which have to obtain their supplies on the internal markets’, in other words that it has, by its Decision No 14/ 55, delegated powers to agencies ill- qualified to exercise them.

Those two complaints refer to the delegation of powers which General Decision No 14/ 55 granted to the Brussels agencies. The first complaint is concerned with the manner in which the powers were delegated, the second with the actual principle of delegation.233

5.1.4.2.1 Whether a delegation of powers had occurredTo address these issues, the Court found it necessary first to determine whether a delegation of powers was at issue or whether the scheme only granted the Brussels Agencies ‘the power to draw up resolutions the application of which belongs to the High Authority, the latter retaining full responsibility’.234 The Court found that some elements in the legal framework led to the conclusion that no delegation was at issue, referring to much the same provisions as in the analysis presented previ-ously (cf. section IV 5.1.2). Other elements, according to the Court, militated in favour of a delegation,235 but the deciding factor for the Court should be searched for in the position adopted by the High Authority in its defence:

The High Authority adopts the data furnished by the Brussels agencies without being able to add anything thereto. Any other specific explanations would mean unauthorized interference in another body’s powers […] The contested decision does no more than reproduce the result of the application by those agencies of the equalization rate to the applicant. Thus if it were to be admitted that the error of which it complains can con-stitute a misuse of powers, that misuse of powers was committed during deliberations of the equalization agencies which the High Authority can no longer contest in view of the

231 Ibid, p 143. 232 Ibid, pp 144– 5. 233 Ibid, pp 146– 7.234 Ibid, p 147. 235 Case 9/ 56, n 223, p 148.


fact that its representative on the Brussels agencies did not reserve the final decision to the High Authority.236

This position taken by the High Authority was very problematic and probably in itself forced the Court to conclude that the High Authority had indeed delegated its powers.237 For one, the High Authority disowned its competence and duty to interfere in the Brussels Agencies’ functioning. Secondly, the High Authority ig-nored that it became responsible for the Brussels Agencies’ decisions in absence of a reservation made by its representative. Lastly, the High Authority further denied that the decisions of the Brussels Agencies could be imputed to it, which would effectively leave the applicant without any means of redress. Having confirmed the delegation, the Court proceeded by checking whether the delegation was in conformity with the Treaty.

5.1.4.2.2 Whether the Treaty allowed delegationIn verifying this issue, the Court remarked that had the High Authority exercised the powers in question itself, these would have been subject to the Treaty rules. The Court then concluded that the general decision 14/ 55 had not made the ex-ercise of the delegated powers subject to any of these restrictions and in effect gave ‘the Brussels agencies more extensive powers than those which the High Authority [held] from the Treaty’.238 Because of this, the delegation contained in Decision 14/ 55 infringed the Treaty.

The Court proceeded by examining the plea related to the Caisse’s use of its own estimates concerning the use of scrap metal by Meroni. The Court did not exclude such a power for the authorities governing a scheme such as the one at issue, but noted that the general decision did not provide for such a power and that the procedure for making such estimates should be subject to precise rules which enable subsequent review and exclude arbitrary decisions. Generally the Court observed that ‘[a] delegation of powers cannot be presumed and even when empowered to delegate its powers the delegating authority must take an express decision transferring them’.239

The Court then arrived at the issue of the extent of the delegation. Meroni claimed that delegation was precluded by the Treaty because the latter had spe-cifically empowered the High Authority and did not explicitly allow delegations. The Court recalled the general decision’s legal base, explaining that the Treaty actually did allow for the involvement of private enterprises to accomplish the objectives of Article 3 ECSC. On the latter Article, the Court remarked that it referred to eight different objectives and that it was not certain that they could be simultaneously attained. As a result, in cases of conflict between the objectives the

236 Ibid, p 148.237 Askenazy also noted that the problem foremost lay with the High Authority’s statements in

its mémoire en défense rather than with the general decision: see H. Askenazy, ‘Les Arrêtés de la Cour de Justice de Luxembourg Concernant la Péréquation des Ferailles’, (1958) L’Echo des Mines et de la Métallurgie 7, p 431.

238 Case 9/ 56, n 223, p 150. 239 Ibid, p 151.


High Authority ought to weigh them against each other.240 The Court concluded that this required a real discretion and it used this criterion to make a distinction between two types of delegation. According to the Court, if purely executive powers are delegated, their use may be reviewed in the light of objective criteria laid down by the delegating authority. On the other hand, a delegation of discre-tionary powers allows the delegate authority to execute an actual economic policy, replacing ‘the choices of the delegator by the choices of the delegate, [bringing] about an actual transfer of responsibility’.241 The Court concluded therefrom that only clearly defined executive powers could be delegated, and this entirely subject to the supervision of the High Authority.

The Court further proceeded to recall the first sentence of Article 3 ECSC, which lays down that the ECSC’s objectives are binding upon the ECSC institu-tions, who need to further these ‘[w] ithin the framework of their respective powers and responsibilities and in the common interest’. According to the Court, ‘[f]rom that provision there can be seen in the balance of powers which is characteristic of the institutional structure of the Community a fundamental guarantee granted by the Treaty in particular to the undertakings and associations of undertakings to which it applies’.242 The Court then linked this with its earlier observations and concluded that delegating discretionary powers would render the guarantee of the balance of powers ineffective, hence infringing the Treaty.

5.1.4.2.3 Whether the delegation at issue respected the Treaty limitsFinally, the Court ascertained whether the delegated powers in casu were discre-tionary or executive in nature. The Court found that the High Authority, together with the Council, had indeed drawn up general guidelines, but these were rather vague, so that the proposals which the Office drafted for the Caisse were not ‘the result of mere accountancy procedures based on objective criteria laid down by the High Authority’.243 The Court also referred to the actual practice of the equal-ization scheme and noted that the High Authority had already been required to intervene because the Office and Caisse could not come to an agreement: ‘[O] n two occasions, by decisions nos 9/ 56 and 34/ 56, the High Authority has itself adopted, in the place and stead of the Brussels agencies, decisions which imply the exercise of a discretionary power.’244

The Court further tested whether the delegation of these discretionary powers could not be saved because of the oversight of the High Authority’s observer in the meetings of the Office and Caisse. At this point the High Authority’s own observa-tions came back to haunt it, as the Court recalled that the High Authority ‘adopts the data furnished by the Brussels Agencies without being able to add anything thereto’.245 The Court finally concluded that because of this, the delegation did not meet the Treaty requirements, and that the contested decision should also for this reason be annulled.

240 Ibid, p 152. 241 Ibid. 242 Ibid. 243 Ibid, p 153.244 Ibid, p 154. 245 Ibid.


5.2 Contemporary reception of the Meroni ruling

Kovar was one of the first authors to emphasize the importance of Meroni for the general issue of delegation, all the while pointing to the peculiarity of the case, since the delegation at issue involved ‘des organismes subsidiaires (et qui plus est de droit privé)’.246 For the present study, the fact that powers were delegated to ‘subsidiary organs’247 is less important than their private law nature. According to Kovar, the Court applied the rules ‘particulièrement défavorable à la délégation de pouvoirs’,248 stressing the rigour in the Court’s decision.249 This critique was developed further by Maas, who deplored the ruling from an administrative point of view, proposing a legal base in the Treaties allowing subsidiary bodies to be established similarly to Article 40 of the Benelux Treaty (current Article 30 of the new Benelux Treaty).250 Such a legal base had also been proposed by Pescatore251 and it would provide for the possibility to delegate powers in conformity with the Treaties, since Maas found that delegations to third parties were possible but del-egations to bodies established by the Community itself were not.252 Pescatore was quite visionary when he proposed, during the negotiations on the Merger Treaty, that the future single Commission should be the body steering and orienting the Community.253 The implementation of the common policies should then be en-trusted either to the national administrations or to European services, offices, or agencies. Noël and Amphoux echoed Maas and Pescatore when they noted that:

the restrictions on the Commission’s power to delegate, which are the result of a rigid institutional structure, pose problems the acuteness of which will only be revealed in the future for the future single Commission. Even more than the enlargement of its material competence, it is the multiplication of its management tasks which may create difficulties, in so far as their volume increases together with the development of the Communities […] In the longer term the fusion of the Treaties should be an opportunity for a profound study on delegation problems in order to determine the increased flexibility […] which the single Treaty may bring.254

According to Maas, Meroni was also applicable in the EEC, given the require-ment to uphold the legal protection offered by the Treaties following a delegation

246 Robert Kovar, Le pouvoir réglementaire de la Communauté européenne du charbon et de l’acier, Paris, Librairie générale de droit et de jurisprudence, 1964, p 151.

247 By this Kovar means organs which are not foreseen in the Treaties (which also applies to the EU agencies), as opposed to individual members of the High Authority and the Member States.

248 Kovar, n 246, p 155. 249 Ibid, p 157.250 Herman Maas, ‘Delegatie van bevoegdheden in de Europese Gemeenschappen’, (1967) 15

SEW 1, p 17.251 Pierre Pescatore, ‘La fusion des institutions et des pouvoirs’, in Dehousse (ed), La fusion

des Communautés européennes: colloque organisé à Liège les 28, 29 et 30 avril 1965, Liège, Université de Liège, 1965, p 49.

252 Maas, n 250, pp 16– 18. Maas agreed with Everling (discussed later) that even Article 235 EEC (current Article 352 TFEU) could not be used for this.

253 Pescatore, n 251, p 49.254 Emile Noël and Jean Amphoux, ‘Les Commissions’, in Ganshof van der Meersch,

Waelbroeck, and Plouvier (eds), Droit des Communautés européennes, Bruxelles, Larcier, 1969, pp 186– 7. Translated from French by the author.


of powers.255 Olmi also made no distinction between the EEC and the ECSC when concluding from Meroni that the Communities orient themselves towards a ‘modèle de constitution rigide’.256 The conclusions which these authors draw from Meroni is that delegation is severely restricted, in that only non- discretionary powers may be delegated to subsidiary bodies.257 On possible backdoors to this rule, Olmi remarked that a delegation of discretionary powers could not be saved ‘en attribuant formellement le pouvoir de décision à l’une des deux institutions [red. Conseil ou Commission] et en prévoyant une délégation de ce pouvoir à l’organe subsidiaire’.258

Everling critically observed that ‘the Commission may only be genuinely re-lieved when real decision- making powers can be delegated to subsidiary organs’.259 According to Everling, Meroni effectively excluded this possibility, only allowing such powers to be exercised by bodies foreseen in the Treaties, since the High Contracting Parties had transferred their sovereignty on the premise that only those bodies would exercise Hoheitskompetenzen.260 Lauwaars sided with Everling on this point and remarked that Kovar and Maas overlooked the fact that the Court did not simply rule on the possibility of delegation in general, but rather on whether the delegation at issue was in conformity with Article 53 ECSC, which, according to Lauwaars, permitted delegation.261 Similarly, Meier interpreted the ruling from an international law perspective and emphasized that the possibility for an international organization to create new subsidiary bodies opens the possi-bility of destroying the delicate build of that organization. The rule to be deduced

255 Herman Maas, ‘La Commission administrative pour la sécurité sociale des travailleurs mi-grants’, (1966) 2 CDE 4, p 351; Herman Maas, ‘The Administrative Commission for the Social Security of Migrant Workers’, (1966) 4 CMLRev 1, p 57. Reuter also found Meroni relevant for the EEC. See Reuter, n 18, p 773.

256 Giancarlo Olmi, ‘Les règlements de Conseil de la C.E.E. en date du 4 avril 1962 établissant une organisation commune des marchés agricoles’, (1962) 10 SEW 6, pp 286– 7.

257 See also Emile Noël and Henri Etienne, ‘Quelques considérations sur la déconcentration et la délégation du pouvoir de décision dans la Communauté Economique Européenne’, (1967) RMC 100, p 128.

258 Giancarlo Olmi, ‘L’Agriculture’, in Ganshof van der Meersch, Waelbroeck, and Plouvier (eds), Droit des Communautés européennes, Bruxelles, Larcier, 1969, pp 703– 4.

259 Everling, n 14, p 41. Translated from German by the author.260 Ibid, p 42. The development of the comitology system in the 1960s indeed triggered a reac-

tion from at least the German Bundesrat, which took the position that the Council and Commission could only delegate advisory, but not decision- making, powers to bodies which are not foreseen under the Treaties. See Georg Holch, ‘Der Bundesrat zur Rechtsetzung der Europäischen Gemeinschaften’, (1969) 4 Europarecht 3, p 228.

261 Richard Lauwaars, Lawfulness and Legal Force of Community Decisions, Leiden, Sijthoff, 1973, p 206. However, according to Kapteyn, Article 53 ECSC did not explicitly permit delega-tion: see Paul Kapteyn, ‘De uitvoering van het Verdrag en de Verordeningen: Gemeenschapsbestuur en nationaal bestuur’, in Van Panhuys, Kapteyn, and Donner (eds), De rechtsorde der Europese Gemeenschappen tussen het internationale en het nationale recht, Deventer, Kluwer, 1966, p 47. This issue is no longer discussed in the present academic debate, although Neergaard has noted that ‘[b] y the word “authorise”, the relevant legal basis actually came very close to explicitly providing that the making of the financial arrangements could be conferred on someone else.’ See Neergaard, n 46, p 615.


from Meroni, according to Meier, is that Article 3 ECSC262 precludes the delega-tion of discretionary powers to bodies not foreseen in the Treaties because such a delegation would violate the guarantee offered by the balance of powers. Meier emphasized the Member States’ interest in upholding the Treaty- defined struc-ture of the ECSC and claimed that the Court also linked its decision in Meroni with this specific interest,263 ignoring the fact that the Court only explicitly stated that the guarantee was provided to private parties. Daubler took a different pos-ition from all foregoing authors by finding that Meroni was a specific ECSC case that could not be transposed to the EEC.264

Legal opinion at the time was therefore already quite critical on Meroni and its effects for the subsequent development of the Communities. In general the Meroni jurisprudence was seen as too strict and the opening left by the Court to allow for powers to be delegated of limited use, as hinted at by Everling. Maas and Pescatore concluded therefrom that efficient forms of delegation necessitated a Treaty change. Still, this consensus was based as much on Meroni as on the authors’ strict interpretation of the ruling, possibly because of the paralyzing effect Meroni had had on the institutional development of the Communities (cf. section IV 5.3).265

5.3 Present- day reception and valuation of the Meroni ruling

Although the previous section shows that the importance of Meroni was not over-looked by legal scholars at the time of the ruling itself, Meroni really came into prominence decades later when the question on delegation resurfaced following agencification. As a result, the rather arbitrary dividing line between the pre-ceding and the present section is drawn at the year 1973, when the EMCF was established, the first body to meet the agency definition (cf. section I 1.1).

It may thus be said that from 1973 onwards Meroni acquired a new and up-graded importance, since most authors started referring to it to frame the new institutional development of agencification. Following the establishment of the EMCF, two eminent scholars in European law, Louis and Ehlermann, linked this institutional innovation back to Meroni. Ehlermann concluded that ‘[f] rom a legal

262 Meier also looked beyond the ECSC Treaty and referred to the analogous Article 3 EEC, implying he also found Meroni relevant in the EEC (and EU) context.

263 Meier, n 2, pp 19– 20. See also in this vein Reuter, who noted that ‘En effet, l’équilibre in-stitutionnel de la Communauté reposant sur l’étroite symbiose qui doit exister entre la Commission […] et les Etats membres […], il n’est pas possible, sans fausser le système, que la Commission devi-enne elle- même un organe de décision entièrement indépendant’. See Reuter, n 18, p 780.

264 Wolfgang Daubler, ‘Die Delegation von Rechtsetzungsbefugnissen im Recht der EWG’, (1966) Deutsches Verwaltungsblatt, p 663.

265 See also Ehlermann’s conclusion to this effect in 1971: Claus- Dieter Ehlermann, ‘Institutionelle Probleme im Bereich der Durchführung des abgeleiteten Gemeinschaftsrechts’, (1971) 6 Europarecht 3, p 259. Ehlermann himself did not concur with this general consensus, however, and correctly pointed to the fundamental difference, under the Treaties and notably the then Article 155 EEC, between the establishment of bodies by the Council or by the Commission. See ibid, p 260.


perspective, the Fund’s establishment [is] highly remarkable: the Community has successfully concluded an operation in which few would have believed only a year ago. The Commission and Council have opened a door which had seemed com-pletely closed to most jurists’.266 The EMCF’s establishment was indeed highly remarkable, given that the delegation of powers contained therein was not in conformity with the principles laid down in Meroni. However, both Ehlermann and Louis invoked the unique nature of the EMCF to justify this violation: the EMCF was to be considered as an embryonic European central bank and, since all national central banks within the EEC had a form of independence from the national legislative and executive powers, the illegality could be saved.267

The case of the EMCF might well be a shorthand for the subsequent academic debate on agencification in the EU: in legal doctrine authors seem to cling to Meroni to establish a workable framework. After having (implicitly) concluded that Meroni does not allow this type of institutional innovation they search for ways to neutralize this tension but, in dialectical terms, the synthesis which they propose to reconcile the thesis and antithesis is not always very convincing (cf. section IV 5.4.2).

During the 1960s and early 1970s Meroni had been viewed as paralyzing the further institutional development of the Communities, although there were al-ready important differences in opinion on Meroni at that time (cf. section IV 5.2). The first steps taken by the institutions in the 1970s, starting with the establish-ment of the EMCF, changed all this, however. Lauwaars for instance was origin-ally of the opinion that Meroni only allowed delegation when the Treaty foresaw delegation, just as Article 53 ECSC did,268 and that the powers of the EMCF should have been entrusted to the Commission, which could delegate the tasks of the Fund to one of its DGs.269 The first wave of agencification led Lauwaars to change his opinion, allowing for delegation to third parties if this is necessary for the attainment of one of the objectives of the Community, which of course im-plied a mandatory recourse to Article 235 EEC.270

In one of the first studies on agencification in 1981, Hilf downplayed the rele-vance of Meroni for the elaboration of the tertiary organizational structure of the Community, noting that the real critique of the Court in Meroni was directed at the lack of sufficient oversight exercised by the High Authority and further noting that Meroni was a specific case concerning bodies under private law.271 With the advent of the second wave of agencification, Lenaerts’ seminal article

266 Claus- Dieter Ehlermann, ‘Die Errichtung des Europäischen Fonds für Währungspolitische Zusammenarbeit’, (1973) 8 Europarecht 3, p 208. Translated from German by the author.

267 Ibid, 201; Jean- Victor Louis, ‘Le Fonds Européen de Cooperation Monetaire’, (1973) 9 CDE 3, p 293. Both authors were followed in this by Lauwaars: see Lauwaars, n 15, pp 286– 7.

268 See the differences in opinion on this, referred to in n 261.269 Richard Lauwaars, ‘Art. 235 als Grundlage für die flankierenden Politiken im Rahmen

der Wirtschafts- und Währungsunion’, (1976) 11 Europarecht 2, p 117.270 Lauwaars, n 15, p 372.271 Hilf, n 10, p 133.


on delegation of powers reaffirmed Meroni’s relevance for agencification,272 and numerous authors have since relied on his analysis.273 Some dissenting voices not-withstanding, it was generally accepted, before Short- selling, that Meroni is not only transposable to the EEC/ EU Treaties but furthermore applies to the current EU agencies. This leaves the questions of which rule(s) may be deduced from Meroni and which apply to agencies. A lot of authors simply observe that agencifi-cation should develop within the limits of the Meroni doctrine, without clarifying this doctrine274 or testing whether the agencies respect this doctrine. As a result these questions merit special attention.

5.3.1 Multiple Meroni doctrinesThe large number of authors writing on Meroni in relation to EU agencies makes its impossible to deal with each of them individually. The different Meroni doc-trines will therefore be presented based on which elements are retained. After all, notwithstanding a broad consensus on the existence of ‘the’ Meroni doctrine and the observation that the Court has drawn clear limits in that case,275 there is no consensus on which elements constitute the doctrine.

The most succinct version of the Meroni doctrine is probably the one used inter alia by Busuioc according to which the ‘Meroni doctrine […] prohibits the

272 Koen Lenaerts, ‘Regulating the Regulatory Process: “Delegation of Powers” in the European Community’, (1993) 18 ELRev 1, pp 23– 49.

273 Among those who refer directly to Lenaerts, see: Michelle Everson, ‘Independent Agencies: Hierarchy Beaters?’, (1995) 1 ELJ 2; Robert Uerpmann, ‘Mittelbare Gemeinschaftsverwaltung durch gemeinschaftsgeschaffene juristische Personen des öffentliches Rechts’, (2000) 125 AöR 4, p 561; Michelle Everson and Giandomenico Majone, ‘Réforme institutionelle: agences in-dépendantes, surveillance, coordination et contrôle procédural’, in De Schutter, Lebesis, and Paterson (eds), La Gouvernance dans l’Union Européenne, Luxembourg, Office des publications officielles des Communautés européennes, 2001, pp 153– 4; Daniel Kelemen, ‘The Politics of ‘Eurocratic’ Structure and the New European Agencies’, (2002) 25 WEP 4, pp 99– 100; Sharon Frank, A New Model for European Medical Device Regulation: A Comparative Legal Analysis in the EU and the USA, Groningen, Europa Law Publishing, 2003, p 125; Vos, n 111, p 130; Ronald van Ooik, ‘The Growing Importance of Agencies in the EU: Shifting Governance and the Institutional Balance’, in Curtin and Wessel (eds), Good Governance and the European Union: Reflections on Concepts, Institutions and Substance, Antwerpen, Intersentia, 2005, p 149; Stefan Griller and Andreas Orator, ‘Everything under Control? The “Way Forward” for European Agencies in the Footsteps of the Meroni Doctrine’, (2010) 35 ELRev 1, pp 3– 35; Herwig Hofmann, Gerard Rowe, and Alexander Türk, Administrative Law and Policy of the European Union, Oxford, OUP, 2011, p 242; Geradin and Petit, n 113, p 165; Gráinne de Búrca, ‘The Institutional Development of the EU: A Constitutional Analysis’, in Craig and de Búrca (eds), The Evolution of EU Law, Oxford, OUP, 1999, p 77; Robert Schütze, ‘From Rome to Lisbon: “Executive Federalism” in the (New) European Union’, (2010) 47 CMLRev. 5, p 1423; Pierre Schammo, EU Prospectus Law: New Perspectives on Regulatory Competition in Securities Markets, Cambridge, CUP, 2011, p 31.

274 See for instance Gregor Schusterschitz, ‘European Agencies as Subjects of International Law’, (2004) 1 International Organizations Law Review 1, p 164; Govin Permanand and Ellen Vos, ‘EU Regulatory Agencies and Health Protection’, in Mossialos (ed), Health Systems Governance in Europe: The Role of European Union Law and Policy, New York, CUP, 2010, p 139.

275 See inter alia Charlotte Gaitanides, ‘Kontrolle unabhängiger Institutionen der Europäischen Union’, (2013) 61 Jahrbuch des öffentlichen Rechts der Gegenwart, p 217.


delegation of discretionary powers’.276 Other authors add a description of the delegate and delegating authority since ‘the institutions may not delegate discre-tionary powers to bodies outside the Treaties’.277 According to inter alia Craig, the Meroni doctrine prescribes that an authority cannot delegate powers other than its own (nemo plus iuris transferre potest quam ipse habet) and that an authority can only delegate non- discretionary powers.278 Apart from these two widely cited elements, other authors add further rules to this Meroni doctrine, such as the req-uirement that the (executive) powers delegated should remain subject to a strict review by the delegating authority,279 the requirement that a delegation cannot be

276 Madalina Busuioc, ‘Accountability, Control and Independence: The Case of European Agencies’, (2009) 15 ELJ 5, p 601. Similarly concise but already different in scope, Thatcher and Coen write that ‘the “Meroni” doctrine of non- delegation […] prevents the Commission from dele-gating rulemaking power’. See Mark Thatcher and David Coen, ‘Reshaping European Regulatory Space: An Evolutionary Analysis’, (2008) 31 WEP 4. See also Chiti, n 212 (CMLRev), pp 1420– 1. Similarly: Christian Joerges and Jürgen Neyer, ‘From Intergovernmental Bargaining to Deliberative Political Processes: The Constitutionalisation of Comitology’, (1997) 3 ELJ 3, p 284; Berend Jan Drijber, ‘De Hoge Autoriteit en de “Brusselse instellingen”’, in Beukers, Van Harten, and Prechal (eds), Het Recht Van De Europese Unie In 50 Klassieke Arresten, Den Haag, Boom Juridische Uitgevers, 2010, p 10; Loïc Grard, ‘Le Contrôle des Actes des Agences de Régulation: Analyse Comparée’, in Peraldi Leneuf and Normand (eds), La légistique dans le système de l’Union européenne: quelle nouvelle approche, Bruxelles, Bruylant, 2012, p 138; Constantin Stephanou, ‘La Nouvelle Gouvernance Européenne: Bilan et Perspectives’, (2006) 42 CDE 5– 6, pp 619– 20.

Scott for his part interprets the Court’s ruling rather exotically as prohibiting the EU’s legis-lative institutions from empowering others to adopt legislation and prohibiting the Commission from delegating a decision- making power conferred on it by the Treaty or by legislation except for the limited purposes of implementation. See Colin Scott, ‘Agencies for European Regulatory Governance: A Regimes Approach’, in Geradin, Petit, and Muñoz (eds), Regulation through Agencies in the EU, Cheltenham, Edward Elgar, 2005, p 69.

277 von Bogdandy, Arndt, and Bast, n 119, p 130. Similarly Gehring and Krapohl see the Meroni doctrine as closing ‘the door for the extensive delegation of formal decision- making compe-tencies to entities not empowered by the European Treaties’. See Thomas Gehring and Sebastian Krapohl, ‘Supranational Regulatory Agencies between Independence and Control: the EMEA and the Authorization of Pharmaceuticals in the European Single Market’, (2007) 14 JEPP 2, p 215. Further see Saskia Lavrijssen and Leigh Hancher, ‘Networks on Track: From European Regulatory Networks to European Regulatory “Network Agencies”’, (2009) 36 LIEI 1, p 38.

278 Paul Craig, EU Administrative Law, Oxford, OUP, 2006, pp 160– 2. Similarly: Renaud Dehousse, ‘Misfits: EU Law and the Transformation of European Governance’, in Joerges and Dehousse (eds), Good Governance in Europe’s Integrated Market, Oxford, OUP, 2002, p 220; Damien Geradin, ‘The Development of European Regulatory Agencies: What the EU Should Learn from the American Experience’, (2004) 11 CJEUL 1, pp 9– 10; Geradin and Petit, n 113, p 148; Johannes Saurer, ‘Accountability of Supernational Administration: The Case of European Union Agencies’, (2009) 24 American University International Law Review 3, p 434; Jacqueline Dutheil de la Rochère, ‘EU Regulatory Agencies: What Future do They Have?’, in Bulterman, Hancher, McDonnell, and Sevenster (eds), Views of European Law from the Mountain, Alphen aan den Rijn, Kluwer Law International, 2009, pp 362– 3; Olivier De Schutter, ‘Mainstreaming Human Rights in the European Union’, in Alston and De Schutter (eds), Monitoring Fundamental Rights in the EU: The Contribution of the Fundamental Rights Agency, Oxford, Hart Publishing, 2005, p 41; Hubertus Rosenberg, ‘No Ace to Win the Trick— the Proposed ACER and Its Influence on EC Competition Law’, (2008) 29 European Competition Law Review 9, p 519.

Gatto also mentions these two and further adds that the requirements related to the institutional balance and the continued supervision by the Commission flow from this. See Alexandra Gatto, ‘Governance in the European Union: A Legal Perspective’, (2006) 12 CJEUL 2, pp 506– 7.

279 Herwig Hofmann and Alessandro Morini, ‘Constitutional Aspects of the Pluralisation of the EU Executive through “Agencification” ’, (2012) 37 ELRev 4, p 423. Similarly see Marc Blanquet,


implied but must be made expressly,280 or the requirement that the institutional balance may not be altered.281

Schütze is one of those who adds the element of the balance of powers and the guarantee which it protects because ‘[t] o delegate “a discretionary power” to “bodies other than those which the Treaty has established” would render that guarantee ineffective’.282 However, whether the notion of balance of powers is part of the doctrine or exogenous but underlying thereto is less clear.283 Lenaerts and Verhoeven note that the Meroni doctrine precludes the delegation of discretionary powers because of the balance of powers.284 On the other hand, Joerges and Neyer

‘Agences de l’union et gouvernance européenne’, in Couzinet (ed), Les Agences de l’Union européenne, Toulouse, PUSST, 2002, p 64; Xenophon Yataganas, ‘Delegation of Regulatory Authority in the European Union The Relevance of the American Model of Independent Agencies’, (2001) Jean Monnet Working Papers 3; Lauwaars, n 15, p 371; Aimery Clerbaux, ‘Compétences et pouvoirs des autorités européennes de surveillance’, (2011) Bank- en Financieel Recht— Droit bancaire et financier VI, p 312.

Van den bergh words this differently as ‘[t] he ultimate decision making power must be pre-served for the delegating entity which has to monitor the implementation of the delegated executive power by the agency’. See Caroline Van den bergh, ‘The Relationship Between Sector Specific Regulation and Competition Law in the Energy Sector— Living Apart Together?’, in Delvaux, Hunt, and Talus (eds), EU Energy Law and Policy Issues, Antwerp, Intersentia, 2012, p 203.

Other authors, such as Karpenstein, Uerpmann, Clément- Wilz, Groenleer, and Van Cleynenbreugel, also mention this requirement but leave out the nemo plus iuris requirement: see Peter Karpenstein, ‘Die Entwicklung des Gemeinschaftsrechts’, (1975) 10 Europarecht 4, p 358; Uerpmann, n 273, p 570; Laure Clément- Wilz, ‘Le contrôle des agences de l’union européenne’, in De Grove Valdeyron (dir), Mélanges en l’ honneur du professeur Joël Molinier liber amicorum, Paris, LGDJ, 2012, p 130; Martijn Groenleer, ‘The European Commission and Agencies’, in Spence (ed), The European Commission, London, John Harper, 2006, p 161; Pieter Van Cleynenbreugel, ‘Europees Financieel Toezicht tussen droom en daad: Enkele alternatieve suggesties in het licht van de financiële crisis’, (2009– 2010) 46 Jura Falconis 2, p 394.

280 Louis, n 267, p 290; Asan Lefterov, ‘How Feasible is the Proposal for Establishing a New European System of Financial Supervisors?’, (2011) 38 LIEI 1, pp 49– 50; Marco Zinzani, Market Integration Through ‘Network Governance’: The Role of European Agencies and Networks of Regulators, Antwerp, Intersentia, 2012, pp 56– 7; Berger, n 13, p 77.

Koenig further adds that ‘decisions by the agencies must not lack supporting reasons indispens-able for the exercise of judicial review’ but this may be subsumed under the nemo plus iuris rule. See Christian Koenig, Sascha Loetz, and Sonja Fechtner, ‘Do We Really Need a European Agency for Market Regulation?’, (2008) 43 Intereconomics 4, p 230.

AG Léger may be mentioned as well in this list, even though he splits up the nemo plus iuris rule into the prohibition of delegating powers other than those of the delegating authority itself and the require-ment that the exercise of powers should be subject to the same conditions had the powers not been delegated. See Opinion of AG Léger in Case C- 301/ 02 P, Tralli v. ECB, [2005] ECR I- 4071, para 31.

281 Peter Bartodziej, Reform der EG- Wettbewerbsaufsicht und Gemeinschaftsrecht: eine Studie zu Vorbildern, Möglichkeiten und primärrechtlichen Gestaltungsgrenzen für ein Europäisches Kartellamt, Baden- Baden, Nomos, 1994, p 304.

282 Schütze, n 147, p 235. Voermans sees the Meroni doctrine similarly, although he defines the category of delegatees as those bodies ‘other than the Commission’ rather than bodies outside the Treaties. See Wim Voermans, ‘Delegeren is een kwestie van vertrouwen’, (2010) 25 RegelMaat 4, p 166. Further see Daniel Kelemen, ‘The Politics of Eurocracy: Building a New European State?’, in Jabko and Parsons (eds), The State of the European Union: With US or Against US? European Trends in American Perspective, Oxford, OUP, 2004, Vol 7, p 186.

283 See also Frank, for whom the Meroni doctrine is based on the institutional balance but at the same time the prohibition on distorting the institutional balance is one of the rules making up the doctrine: n 273.

284 Koen Lenaerts and Amaryllis Verhoeven, ‘Institutional Balance as a Guarantee for dem-ocracy in EU Governance’, in Joerges and Dehousse (eds), Good Governance in Europe’s Integrated


remark that the ‘Meroni doctrine must be read together with the request to respect the “institutional balance” between Community organs’.285 According to Griller and Orator, the Meroni doctrine encompasses the following four rules: (i) nemo plus iuris, (ii) a delegation cannot be presumed, (iii) only clearly defined execu-tive powers may be delegated, and (iv) the preceding three rules are ‘related’ to the balance of powers.286 Schneider takes over these first three requirements but is clearer on the last point as he remarks that the Court’s ruling was based on the principle of balance of powers.287 Similarly but not identically, Zinzani notes that the notion of institutional balance reinforced the reasoning for limiting the types of delegatable powers.288 Having noted this point of divergence, this part of the debate is left for a subsequent section (cf. section IV 5.6.2.2.3).

Not all authors include the requirement prescribing that a delegation of powers should be made expressly. According to Everson, the four limits which flow from Meroni are that (i) only non- discretionary powers may be delegated, (ii) subject to the nemo plus iuris rule; further that (iii) the Commission should retain oversight and remain responsible; and lastly (iv) a delegation may not disturb the EU’s bal-ance of powers.289 Majone et al propose a similar Meroni doctrine and write, as regards the nature of the powers which may be delegated, that a delegation may only relate to the preparation and performance of executive acts,290 rather than the adoption of executive acts. As a result this interpretation would also result in a prohibition on outsourcing the preparatory work for legislation to agencies, rather than a prohibition on delegating the (formal) power to adopt a binding decision implying discretionary choices.291

More recently Tridimas deduced a new rule from Meroni: the delegation should be necessary for the performance of the tasks which the Treaty assigned to the

Market, Oxford, OUP, 2002, p 37. Similarly, see Daniel Riedel, Die Gemeinschaftszulassung für Luftfahrtgerät, Berlin, Duncker & Humblot, 2006, pp 272– 3; Michaela Wittinger, ‘Europäische Satelliten: Anmerkungen zum Europäischen Agentur(un)wesen und zur Vereinbarkeit Europäischer Agenturen mit dem Gemeinschaftsrecht’, (2008) 42 Europarecht 5, p 618.

285 Joerges and Neyer, n 276, p 284.286 Griller and Orator, n 273, pp 16– 17.287 Jens- Peter Schneider, ‘A Common Framework for Decentralized EU Agencies and the

Meroni Doctrine’, (2009) 61 Administrative Law Review Special Issue, pp 35– 6. See also Andoura and Timmerman, who note that the Court referred to the balance of powers ‘to justify [its] restrictive position’. See Sami Andoura and Peter Timmerman, ‘Governance of the EU: The Reform Debate on European Agencies Reignited’, (19 October 2008) EPIN Working Paper, pp 12– 13.

288 Zinzani, n 280, p 56.289 Everson, n 273, p 192. David proposes a similar Meroni doctrine, although she does not refer

to the Commission as the delegating authority but more generally to (any) übertragende Organ: see Antje David, Inspektionen im Europäischen Verwaltungsrecht, Berlin, Duncker & Humblot, 2003, p 187. Vos essentially proposes a similar Meroni doctrine rendered in six rather than four rules. See Vos, n 45, p 201. Ott also proposes a similar Meroni doctrine but explicitly adds continued judi-cial review by the Court as a fifth requirement. See Andrea Ott, ‘EU Regulatory Agencies in EU External Relations’, (2008) 13 European Foreign Affairs Review 4, pp 2– 3.

290 Michelle Everson, Giandomenico Majone, Les Metcalfe, and Adriaan Schout, The Role of Specialised Agencies in Decentralising EU Governance— Report Presented to the Commission, September 1999, p 53.

291 This may also already be seen in Everson, n 273.


delegating authority.292 Lenaerts also already hinted at this in his seminal article in 1993,293 and it is further in line with Schreiber’s findings referred to in the sec-tion on the principle of conferred powers.294

This short browse through the contemporary literature295 on the issue of dele-gation and the Meroni doctrine shows how multi- interpretable the ruling is. In all, seven limits and two applicability conditions have been identified:296 (i) the nemo plus iuris rule,297 (ii) the rule that delegations should be explicitly provided for, (iii) the rule that only executive, non- discretionary powers may be conferred, (iv) the rule that the delegating authority should continue to supervise the delegate, (v) the rule prescribing that review by the Court should be maintained,298 (vi) the rule that the balance of powers or the institutional balance may not be upset, and (vii) the rule that the delegation at issue must be necessary for the performance of the tasks assigned to the delegating authority. As regards the conditions defining when Meroni applies, there are some authors who restrict Meroni’s validity dep-ending on the (i) delegate authority.299 Likewise, some authors restrict Meroni’s validity depending on the (ii) delegating authority.300

Despite the disagreement on the rules to be brought under the Meroni doctrine, a consensus does exist on one requirement, that is, the prohibition on delegating

292 Takis Tridimas, ‘Financial Supervision and Agency Power: Reflections on ESMA’, in Nic Shuibhne and Gormley (eds), From Single Market to Economic Union: Essays in Memory of John A. Usher, Oxford, OUP, 2012, p 60.

293 Lenaerts, n 272, p 42. 294 See n 9.295 It is not claimed this overview is exhaustive as regards the number of authors which have

written on the subject, but the overview should be rather exhaustive as to the different rules which have been deduced from Meroni.

296 These seven limits are already a deduction from the legal doctrine; in reality even more limits could be identified, since some authors split up a single limit into multiple ones: see for instance Frank, n 273, pp 123– 4.

297 Vos, see n 289, also refers to the conditions which would have applied to the exercise of the power in question had it not been delegated but rather exercised by the delegating authority, but because such a rule forms part of the nemo plus iuris rule it is not referred to separately here.

298 As will be shown presently, this rule could also be subsumed under the nemo plus iuris rule even if some authors refer to it as separate from that rule. See Ott, n 287; Vark Helfritz, Verselbständigte Verwaltungseinheiten der Europäischen Union, Berlin, Weissensee Verlag, 2000, p 158.

299 A lot of authors refer to these entities as ‘agencies’, ‘bodies’ or ‘decentralized bodies’: see inter alia Groenleer, n 279, p 161; Zinzani, n 280, p 55; Tridimas, n 292, p 60.

Möllers is more precise when he defines the delegate authorities as bodies other than foreseen by the Treaties, see Christophe Möllers, ‘Durchführung des Gemeinschaftsrechts’, (2002) 37 Europarecht 4, p 495. This is different from bodies outside the Treaties of which Ohler speaks, see Christoph Ohler, ‘Finanzmarktregulierung und - aufsicht’, in Ruffert (ed), Europäisches Sektorales Wirtschaftsrecht, Baden- Baden, Nomos, 2013, p 656. Similarly see Schusterschitz who speaks of ‘bodies outside the ECT’: Schusterschitz, n 274, p 164. Furthermore, bodies ‘not foreseen by the Treaties’ and bodies ‘outside the Treaties’ are still different from the bodies ‘not created by the Treaties’ or ‘organismes autres que les institutions instaurées par les traités’ referred to by Kelemen and Tarrant and Stephanou. See Daniel Kelemen and Andrew Tarrant, ‘The Political Foundations of the Eurocracy’, (2011) 34 WEP 5, pp 943– 4 at footnote 7; Stephanou, n 276, p 619. As noted previously, Voermans speaks of bodies other than the Commission: see n 282.

300 According to Berger, the Court’s reference to Article 3 ECSC has the effect that its Meroni doctrine applies to all institutions. See Berger, n 13, p 77. See also Busuioc, n 276, p 601; Groenleer, n 279, p 161. David speaks of any delegating authority: see n 289. Other authors such as Frank and Gatto confine this to the Commission: see n 296 and Gatto, n 278, p 506.


discretionary powers. However, even if the Meroni doctrine boils down to this single rule, and assuming Meroni applies to agencies, agencification remains prob-lematic. Still, there appear to be many Meroni doctrines and as a result there is no real consensus on what Meroni actually means for agencification today.

Whether because Meroni seems to prevent agencification (in a qualitative sense) or simply because of analytical rigour, the mainstream interpretations of the Meroni ruling have also been the subject of critique. This critique mainly aims to argue that Meroni does not preclude further agencification, either because Meroni is irrelevant for the EU agencies or by presenting an alternative reading of Meroni. These dissenting voices will therefore be presented in the following section.

5.4 EU agencies under Meroni: Does it make sense?

Applying Meroni to agencies requires this implicit assumption to be addressed. In the years following Meroni this question did not really pose itself because the ruling had severely limited the institutional development of the Communities.301 Following the agencification in the 1970s, however, this new institutional dev-elopment could be analysed in the light of the Court’s jurisprudence, but the conclusions were not comforting. Some authors argued for a modification of the Treaties overruling Meroni.302 Others invited the Court to overrule Meroni, taking account of the new context. As early as 1979 Priebe argued that ‘as inte-gration continues, the practical necessity of transferring powers will become more prominent. The recent development makes this clear. The Court, in the event of a new decision, cannot shut its eyes for the new problems which have popped up in relation to the Community administration and which were unforeseen in 1958.’303

One could of course also argue that Meroni simply does not apply to EU agen-cies or that Meroni may be interpreted more flexibly— a Meroni- light, so to say— which may allow further agencification. These arguments will be looked at in turn. The specific arguments on a Meroni- light that draw from the arguments refuting Meroni’s relevance for agencies will also be included in the first section.

5.4.1 Meroni’s irrelevance for EU agenciesA preliminary remark for the present section as regards the authors cited is in place here. Although their arguments will be used to argue that Meroni may not apply to agencies, it is not claimed that the authors referred to also take this position. Dehousse for instance seems to make a case against applying Meroni to agencies without however drawing such an explicit conclusion from his arguments.304

301 Ehlermann, n 266, p 198; Maurice Torelli, ‘Les ‘Habilitations’ de la Commission des Communautés Européennes’, (1969) RMC 127, p 471.

302 Apart from the proposals made earlier by Maas and Pescatore (cf. section IV 5.2) see Geradin, n 278, p 16; van Ooik, n 273, p 132; Vos, n 111, p 132; Rosenberg, n 278, p 520.

303 Priebe, n 27, p 114. Translated from German by the author.304 See the argumentation presented below, from Dehousse, n 278.


5.4.1.1 From ECSC to EUAlthough legal doctrine at the time of the Meroni ruling saw no problem in trans-posing the ruling from its ECSC context to an EEC context, with the exception of Daubler,305 the naturalness of this was only questioned again during the third wave of agencification by Dehousse and Geradin.306 Dehousse notes that Meroni is generally considered to be valid within the wider Community legal order, but that this view neglects the important differences between the institutional set-tings of the ECSC and the E(E)C/ EU.307 Dehousse thereby juxtaposes the traité loi of the ECSC with the traité cadre of the EEC.308 Under the ECSC Treaty, the High Authority had important regulatory powers, whereas the EEC’s objec-tives had to be attained gradually through legislative decisions. The thrust of the argument therefore is that the position of the High Authority is fundamentally different from that of the Commission, which only has regulatory powers in com-petition policy. Because of this, Meroni could not be transposed to the present- day EU. Priebe, on the other hand, while acknowledging this difference between the ECSC and EEC Treaties, still rejected such an argument, since the Court in Meroni relied on general principles which therefore also applied to the EEC. As a result, even if the EEC’s traité cadre allowed for more integration than the ECSC, it still did not allow for a substantial ‘restructuring of the Treaty’.309

Allegedly the argument by Dehousse and Geradin came under pressure some years after it was made, when the Court referred to Meroni in two cases under the EC Treaty.310 These cases, Alliance for Natural Health and Tralli, are dealt with in a subsequent section, the relevant question here being whether the Court’s reliance on Meroni in those cases amounted to a rejection of Dehousse’s and Geradin’s ar-gument. On this, it should be noted that the Meroni doctrine expresses both general principles related to delegation and specific principles in the context of the ECSC. As a result, the Meroni doctrine is no monolith that can only be ac-cepted or rejected as a whole. This is something Zwart also seems to ignore when he generally states that the limits established by the Court in Meroni do not apply to the establishment of agencies, since agencies are established by the legislator

305 See n 264.306 Geradin largely relies on Dehousse: see Geradin, n 278, p 10. Although Hilf also made crit-

ical remarks, he did not explicitly question Meroni’s relevance for agencies. Rosenberg concurs with and refers to Geradin and Dehousse on this issue: see Rosenberg, n 278, pp 519– 20.

307 Dehousse, n 278, p 221.308 Similarly, Berger notes that the Rahmenvertrag of the EEC should be interpreted more flex-

ibly than the ECSC Treaty. See Berger, n 13, p 89. Yet, ‘more flexibly’ would point towards a Meroni- light (cf. section IV 5.4.2) rather than to a rejection of Meroni.

309 Priebe, n 27, pp 102 and 113– 14.310 Inter alia the following authors rely on Alliance for Natural Health and Tralli to argue the

relevance of Meroni for EU agencies: Pierre Schammo, ‘The European Securities and Markets Authority: Lifting the Veil on the Allocation of Powers’, (2011) 48 CMLRev. 6, p 1895; Koenig, Loetz and Fechtner, n 280, pp 230– 1; Saurer, n 278, pp 435– 6; Edoardo Chiti and Ramses Wessel, ‘The Emergence of International Agencies in the Global Administrative Space: Autonomous Actors or State Servants?’, in Collins and White (eds), International Organizations and the Idea of Autonomy, New York, Routledge, 2011, p 149; Chiti, n 212 (CMLRev), p 1421; Griller and Orator, n 273, p 20; Craig, n 278, pp 161– 2; Zinzani, n 280, p 55.


and do not receive their powers following a sous- délégation by the Commission.311 In reality, some of the limits in Meroni seem generally applicable in any modern state based on the rule of law. It should be clear that the distinction between the ECSC and the E(E)C/ EU is therefore irrelevant for such limits. However, this is not necessarily the case for those limits which the Court imposed from a reading of the ECSC Treaty itself. As such, different authors have relied on the different contexts under primary law between Meroni and the current agencies to support their argument for a Meroni- light (cf. section IV 5.4.2).

5.4.1.1.1 Recycling the argument for a Meroni- lightBartodziej and Schneider also used the above arguments to argue for a Meroni- light.312 In his argument Schneider notes that in the EU Treaties, unlike in the ECSC Treaty and specifically Article 53, there are few provisions explicitly confer-ring administrative powers to the Commission and that ‘in most cases, implemen-tation competencies are allocated to the Commission by secondary law in accord-ance with Art. 202 ECT’.313 The relevant provision in that Article has now been replaced by Article 291 TFEU (and Article 290 TFEU). Because the EU Treaties have dealt with the implementation problem in a general way, instead of specific-ally for each policy field, Schneider is of the opinion that concrete administrative arrangements may be made by the EU legislator. In a way the Court has accepted this broad reasoning in its jurisprudence on the comitology system, when the ques-tion was raised as to whether these comitology procedures did not impinge on the Commission’s competences.314 However, allowing in principle that ‘administrative arrangements’ may be made by the EU legislator does not answer the question of which arrangements are justified and which arrangements are excluded.

Because of the specific scope of his study (cf. section IV 5.4.2.2), Bartodziej notes that the current Articles 103 and 352 TFEU do not lend themselves as easily as Article 53 ECSC to infer a non- delegation rule from them. Because of this, and in the absence of other specific provisions on delegation, the limits thereto should be sought in another ‘allgemeinen Grundsatz’ of EU law, that is, the institutional balance.315 Again, these observations appear correct, but they do not in them-selves provide support for a Meroni- light. Whether or not a Meroni- light can be construed will depend exactly on how EU agencies and the conferral of powers to them should be viewed under the notion of institutional balance (cf. section IV 7). Lastly, it may be noted that the Court in Short- selling does not seem to have been inspired by the argument under this section.

311 Tom Zwart, ‘La Poursuite du Père Meroni ou pourquoi Les Agences Pourraient Jouer un Rôle plus en Vue dans l’Union Européenne’, in Dutheil de la Rochère (ed), L’Exécution du Droit de l’Union, entre Mécanismes Communautaires et Droits Nationaux, Bruxelles, Bruylant, 2009, p 166.

312 Bartodziej, n 281; Schneider, n 287.313 Schneider, n 287, p 38.314 The so- called management and regulatory procedures, the forerunners to the current exam-

ination procedure, were sanctioned by the Court in Köster and Tedeschi: see Case 25/ 70, Köster, [1970] ECR 1161; Case 5/ 77, Tedeschi, [1977] ECR 1555.

315 Bartodziej, n 281, p 309.


5.4.1.2 A Europeanization rather than a delegation of powersAgain it was Dehousse and Geradin, but also Yataganas, who argued that Meroni’s relevance for the EU agencies may be doubted because what is currently at stake in agencification is not so much a delegation of powers at EU level, but the ‘Europeanization’ of powers from the Member States to an agency.316 Dehousse links this with the previous argument in that the ECSC Treaty gave the High Authority specific powers and in Meroni the High Authority had del-egated (some of) these powers. The agencies however exercise tasks which were previously fulfilled by the national administrations. The result is that there is no delegation from the Commission to an agency, but the conferral of powers from the Member States to the EU, which should not be governed by the prin-ciples applicable to delegation:317 consequently ‘the institutional system is not concerned’.318 This argument in itself is insufficient, as Dehousse and Geradin do not make explicit one important premise on which it is based. Yataganas on the other hand, relying on Triantafyllou, does refer to this premise.319 This is the question of whether the Treaties lay down an exhaustive list as regards the EU institutions and whether the current Articles 290 and 291 TFEU lay down a rule on delegation which is exclusive in nature or whether the rule allows for exceptions.320

The question of who exercises which competences in the integrated EU legal order indeed has both a horizontal and a vertical dimension. From a vertical per-spective the (simplified) choice is between the EU and the Member States. From a horizontal perspective, once a legal base is established for the EU to exercise the competence in casu, the question is how this competence will be performed by the different EU actors. Dehousse and Geradin seem to suggest that once it is estab-lished that the competences have (gradually) shifted vertically from the national to the EU level, the horizontal dimension becomes irrelevant,321 but this is of course

316 See also Wolfgang Weiss, ‘Agencies versus Networks: From Divide to Convergence in the Administrative Governance in the EU’, (2009) 61 Administrative Law Review Special Issue, p 50. Hancher and Larouche also argue that the BEREC’s powers could be enhanced without upsetting Meroni precisely because this enhancement would be to the detriment of the national authorities rather than to that of the Commission. See Leigh Hancher and Pierre Larouche, ‘The Coming of Age of EU Regulation of Network Industries and Services of General Economic Interest’, in Craig and de Búrca (eds), The Evolution of EU Law, Oxford, OUP, 2011, p 777.

317 Dehousse, n 278, p 221; Geradin, n 278, p 10. Rosenberg concurs with and refers to Geradin and Dehousse on this point as well: see Rosenberg, n 278, p 520. In the same vein, Hofmann and Morini note that if the empowerment of EU agencies does not result from a (sub- )delegation, then the ‘limits to the creation of agencies and their powers arise from the constitu-tional principles such as conferral, proportionality and subsidiarity’. See Hofmann and Morini, n 279, p 434.

318 Yataganas, n 279. 319 Ibid.320 Dimitris Triantafyllou, Les compétences d’attribution au domaine de la loi: étude sur les

fondements juridiques de l’action administrative en droit communautaire, Bruxelles, Bruylant, 1997, p 306. That Articles 290 and 291 TFEU do not form a closed system has now been confirmed in Short- selling.

321 Previously, the change brought by the Lisbon Treaty to Article 7 EC was noted. However, even following that change, the horizontal dimension to this problem has not become irrelevant. On its relevance see also AG Bot n 751.


not the case. How much leeway exists to endow agencies with certain powers will indeed depend on how the questions raised by Triantafyllou are answered. This will be done in the chapter devoted to the institutional balance.

A rather exotic argument, which has not found much support in legal doctrine, was proposed by Ehring, who argued that the conditions on which Meroni relies are lacking when bodies are established under the current Article 352 TFEU. Admittedly, because of Ehring’s reliance on Article 352 TFEU, his argument would also have fitted under the previous section. Because Meroni only referred to existing competences attributed by the Treaty to a delegating authority— whereas Article 352 TFEU is used when the competences provided by the Treaty are in-sufficient to achieve the objectives of the EU— and because it is impossible to identify an authority in which these competences are originally vested, Ehring suggests that Meroni cannot be applied.322 In his own words: ‘Both the limits to existing competences as well as their transfer from the bodies in which they were originally vested are lacking for the application of the [Meroni] principles.’323 Again because some rules in Meroni seem generally applicable, this argument cannot be accepted. The second element of Ehring’s argument is interesting, how-ever, and shows how the lack of precisely attributed competences (linking com-petences to bodies or institutions) opens the scope for the institutional balance to govern empowerments of agencies.

Finally, following Short- selling, Van Cleynenbreugel made an argument similar to that of Ehring but based on Article 114 TFEU. When this Article is used to create agencies which essentially play a supporting role for national authorities (such as ECHA) or which support the implementation of EU law by stepping in exceptionally (such as the ESAs and ACER), Van Cleynenbreugel argues that neither Meroni (nor Romano) are relevant for the empowerment of these agencies since what are at issue are ‘constitutional delegations upwards from the Member States’.324 In this, Van Cleynenbreugel already makes clear that the delegation at issue does not meet the general characteristics of a normal delegation (see section IV 5.7.2.2). The delegation is constitutional in the sense that ‘[t] he establishment of EU agencies […] concerns the delegation of Member States’ implementing powers under EU law to an independent supranational authority by the Council and the European Parliament in accordance with an acknowledged Treaty legisla-tive procedure’.325 For this, Van Cleynenbreugel also invokes Article 291 TFEU, which emphasizes that the Member States are the default actors responsible for implementation. On this point there is of course a certain tension in the argument presented by Van Cleynenbreugel. If Article 291(1) TFEU is read as an emphasis

322 Chiti makes a similar argument when stating: ‘differently from what [was] presupposed by [Meroni], the powers conferred to the European agencies are not univocally provided by the Treaty as pertaining to the supranational authorities.’ See Chiti, n 212 (Della Cananea), p 79.

323 Ehring, n 15, p 790. Translated from German by the author.324 Pieter Van Cleynenbreugel, ‘Meroni circumvented? Article 114 TFEU and EU

Regulatory Agencies’, (2014) 21 MJECL 1, p 81 (emphasis in original).325 Ibid, p 81 (emphasis added).


on the Member States as primary loci of implementing powers in the EU, it is remarkable that Article 114 TFEU would implicitly allow the Member States (de-ciding only with QMV in the Council) to ‘delegate’ such powers away from the national and to the EU level. The argument that the EU legislator can delegate powers from the national to the EU level would rather incidentally suggest that Article 291(1) TFEU grants implementing powers to the Member States instead of confirming the retained competences in this area.326 Just as may be noted in the analysis of the CLS (see section IV 5.4.3.2), the rejection of Meroni by Van Cleynenbreugel is then followed by a recycling of much of what can be found in Meroni, together with institutional balance considerations, in a set of newly pro-posed constitutional boundaries.327

For the reasons set out in the preceding paragraph, the arguments presenting agencification as ‘Europeanization’ rather than ‘delegation’ cannot as such rule out the relevance of Meroni for EU agencies.328

5.4.1.3 Private bodies or vertragsfremde Organe?As early as 1964, Kovar stressed that Meroni was a case of delegation to bodies under private law.329 Yataganas further remarked that ‘the strictness of the Meroni judgment is quite admissible and pertinent for the delegation of such powers to private legal persons’,330 hinting that this is not the case when dealing with a dele-gation of powers to agencies.331

Unfortunately in the legal doctrine on Meroni previously presented, few other authors dedicate special attention to this issue.332 When for instance German authors speak of vertragsfremde Organe, it is not clear whether they mean bodies not foreseen under the Treaties or bodies outside the Treaties.333 The presentation of the facts of the Meroni case (cf. section IV 5.1) further shows why the distinc-tion between private and public bodies is important: delegation to private parties increases the risk of conflicts of interest, which were also apparent in Meroni (dis-cussed presently).

But again the private– public distinction is only relevant to a certain extent in light of the general principles on delegation in Meroni. The simple observation of the Court in Short- selling that the bodies in Meroni ‘were entities governed by private law, whereas ESMA is a European Union entity, created by the EU legis-lature’334 is therefore unhelpful. That EU agencies are different from the bodies

326 See the discussion mentioned in Chapter I, n 266.327 Van Cleynenbreugel, n 324, pp 83– 8.328 See also in this vein Schneider, n 287, pp 37– 8. 329 See n 246.330 Yataganas, n 279.331 See also Triantafyllou, n 320, pp 306– 7.332 For some that do, see Jürgen Kühling, ‘Die Zukunft des Europäischen Agentur(un)

wesens— oder: Wer hat Angst vor Meroni?’, (2008) 19 EuZW 5, p 129; Koenig, Loetz and Fechtner, n 280, p 231; Merijn Chamon, ‘EU Agencies: Does the Meroni Doctrine Make Sense?’, (2010) 17 MJECL 3, pp 297– 8.

333 Of course this issue is not confined to the authors writing in German.334 Case C- 270/ 12, n 92, para 43.


at issue in Meroni is clear for anyone to see; the real question is how this affects the way in which the rules originally laid down in Meroni may be applied to EU agencies. Unfortunately, the Court in Short- selling did not dwell on this issue.

5.4.1.3.1 Recycling the argument for a Meroni- lightThis argument would however also allow a construction of a Meroni- light. In a comparative study of administrative reform in Germany and France, Schmidt- Aßmann and Dagron note that despite both systems’ differences, the legitimacy of the administration ultimately goes back to the people.335 Delegating powers to bodies under private law evidently puts this relationship at risk, making it gener-ally impossible to confer essential government functions on bodies under private law.336

Probably inspired by similar considerations, Koenig et al note that in Meroni ‘[t] he High Authority entrusted certain tasks falling within its responsibility to two private agencies set up under Belgian private law, which were therefore not integrated into the authority organisation of the Community. The delega-tion of powers to a European agency that is embedded in the administration of the Community should [therefore] be possible under less rigorous conditions.’337 Unfortunately Koenig et al do not elaborate on this point, and neither do they dif-ferentiate between the conditions which are unassailable regardless of the nature of the delegate and those which may be eased. Zwart also merely notes that dele-gation to public bodies should be more generous because they are subject to the guarantees provided by public law.338

Although few other authors that argue for a Meroni- light refer to this important difference between the EU agencies and the Office and the Caisse, it may be said that this element is implicitly present in the argument of those authors who pro-pose that the Meroni doctrine may be softened by emphasizing different forms of control over EU agencies. The fact that EU agencies are public law bodies evidently broadens the scope of possible control mechanisms which may be estab-lished. In addition, public law bodies are also less prone to agency capture.339 The Meroni- light established by the Court in Short- selling may have been inspired by

335 Eberhard Schmidt- Assmann and Stéphanie Dagron, ‘Deutsches und französisches Verwaltungsrecht im Vergleich ihrer Ordnungsideen’, (2007) 67 ZaöRV 2, pp 426– 33.

336 Frederik Vandendriessche, Publieke en private rechtspersonen: naar een graduele, meer-duidige en evolutieve benadering van het onderscheid in de wetgeving en de rechtspraak, Brugge, Die Keure, 2004, p 184.

337 Koenig, Loetz and Fechtner, n 280, p 231.338 Zwart, n 311, p 167.339 To put it correctly, agency capture is a risk for public bodies regulating (an) economic

sector(s). Dal Bó defines agency capture as ‘the process through which special interests affect state intervention in any of its forms’. See Ernesto Dal Bó, ‘Regulatory Capture: A Review’, (2006) 22 Oxford Review of Economic Policy 2, p 203. This is slightly different for forms of self- regulation such as the compulsory scheme at issue in Meroni, where the risk is that self- regulation is conducive to cartel- like agreements. Furthermore a self- regulatory regime may also be perceived as furthering industry rather than public interest. See Tony Prosser, ‘Self- regulation, Co- regulation and the Audio- Visual Media Services Directive’, (2008) 31 Journal of Consumer Policy 1, p 103.


such considerations but, as was noted, the Court did not elaborate on the import-ance of this difference.

A caveat may also be noted in legal doctrine, perhaps because the Court itself in Meroni did not find it necessary to point to this (even if AG Roemer did): since the equalization scheme had originally been established as a voluntary scheme under Article 53a ECSC, the Office and the Caisse were cooperative societies dominated by the biggest steel companies.340 Under the compulsory scheme, the High Authority had therefore given these undertakings important powers through which they could affect their own competitive positions and that of their competitors in the steel market. It is also for this reason that the applicant, Meroni, was so adamant about the ‘untransparent’ way in which the Caisse had made its calculations. If there ever was a legal construction conducive to conflicts of interest, it would indeed be the compulsory scheme established by the High Authority through Decision 14/ 55.

5.4.1.4 ConclusionThe three arguments just presented have been invoked in legal doctrine to argue that Meroni may not be all that relevant to EU agencies. While some are worth pursuing, the question of Meroni’s relevance for agencies should not be presented as having a binary answer. In addition, the Court in Short- selling has now expressly confirmed Meroni’s relevance, closing the door to these arguments. However, since they may still shed light on a Meroni- light, used by the Court in Short- selling (cf. section IV 5.8.1.3), the Meroni- light variants of two of the argu-ments were elaborated. Further Meroni- light arguments will be evaluated in the following section.

5.4.2 A Meroni- light for the EU agenciesThat some of the arguments in support of a Meroni- light may appear rather op-portunistic is evidenced by Kelemen, who noted that ‘[a] s the Commission’s polit-ical willingness to delegate rulemaking powers increases, its emphasis on Meroni is likely to decrease’.341 According to Kelemen it is the Commission which holds back further agencification but once the political institutions agree on the estab-lishment of an agency, the Court will not likely intervene.342 Kelemen makes a distinction in this regard between the limits on delegation and the ongoing con-trol following delegation, arguing that the Court will focus on the latter rather than on the former.343

340 This may be juxtaposed with the AEC’s Board which was composed of functionaries of the Commission (cf. section I 1.2).

341 Kelemen, n 282, p 186.342 The analysis of the Court’s ruling in Short- selling (cf. section IV 5.8.1.3) indeed confirms

Kelemen’s prediction.343 Kelemen, n 273, p 99.


Wittinger proposes a de lege ferenda argument noting that the law should also reflect legal reality and that the Union needs to rely more on agencies to be able to fulfil its objectives.344 According to Wittinger, the result would be that ‘the development of a distinct level of administration in the Community, through the agencies in the framework of a mitigated Meroni Doctrine, [is] compatible with Community law’.345 Wary of more dogmatic critique on this position, she further remarks that ‘this is not the same as giving in to the “de facto situation”, but it does take account of the praxis of law in the context of a possible interpret-ation’.346 Although one could concur with this observation, the question remains of how much weight should be given to it when interpreting Meroni.

5.4.2.1 Focusing on the institutional balanceA different technique with which to overcome Meroni’s strict limits is empha-sizing the Court’s reliance on the balance of powers. According to Vos, ‘the Court’s objections to agencies in the Meroni cases can be reduced to one main ob-jection: distortion of the institutional balance’.347 In this Vos equates the ‘balance of powers’ with the ‘institutional balance’, something on which there is indeed a general consensus in legal doctrine. Whether this equation is correct will be dealt with in a subsequent section (cf. section IV 5.6.2.2.3). What is important for now is that the Meroni doctrine is being reduced here to the obligation to respect the institutional balance. This is an important analytical step because it effect-ively makes everything what the Court said about delegation of powers in Meroni irrelevant. What instead becomes important when considering a ‘delegation’ of powers is how the conferral would affect the institutional balance; only if the bal-ance is upset will such a conferral be ultra vires.

A legitimate question here is whether the Meroni doctrine may indeed be re-duced to the balance of powers. The argument proposed by Vos only concentrates on the requirement that discretionary powers cannot be delegated. Vos relies on a lot of the elements previously cited which distinguish Meroni from the agencies today, to downplay the importance of the prohibition on delegating discretionary powers:348 the change from a traité loi to a traité cadre, the Brussels Agencies’ pri-vate law nature, and the distinction between ‘executive’ and ‘discretionary powers’ having been blurred by the Court in its subsequent jurisprudence, more specif-ically Opinion 1/ 76 (cf. section IV 5.5.1.1). The latter element in particular leads Vos to conclude that ‘the significance of the Meroni doctrine for the delegation of powers to agencies does not lie in the distinction between a delegation of “mere ex-ecutive” and a delegation of “discretionary” powers […] Rather, attention should be focused on the principle of the institutional balance of powers.’349 Regardless

344 Wittinger, n 284, pp 619– 20.345 Ibid, p 620. Translated from German by the author.346 Ibid (emphasis in original). Translated from German by the author.347 Vos, n 111, p 131. Similarly see, Lavrijssen and Hancher, n 277, p 38; Zinzani, n 280,

pp 55– 6; Chiti, n 212 (CMLRev), p 1423.348 Vos, n 45, pp 202– 3. 349 Ibid, p 203.


of the fact that the reasoning proposed by Vos ignores elements in Meroni other than the prohibition of delegating discretionary powers, it does not seem entirely convincing. The implicit criticism on the Court’s reasoning in Opinion 1/ 76 can indeed be subscribed to, but the remarkable qualification of the laying- up fund’s powers by the Court in that Opinion cannot serve as a basis to conclude that the Court’s distinction in Meroni between executive and discretionary powers has become irrelevant, which has recently also been confirmed in Short- selling.

Still, assuming one could reduce Meroni to the institutional balance, a second problem arises: the meaning and content of the principle of institutional balance are not clear, and the authors presenting this argument have not attempted to clarify them either. In fact there are other authors who claim that the principle of institutional balance is not a principle at all (cf. section IV 7.2). What is important for the moment is that Vos’ proposition opens up two possible arguments under-pinning a Meroni- light. The first is to claim that the effects of agencification on the institutional balance are positive, meaning agencification should be allowed and even encouraged. The second is to claim that the institutional balance only needs to be rebalanced to allow further agencification, and that this rebalancing may be done outside primary law.

5.4.2.1.1 ‘Strengthening’ the institutional balance through agencificationSome authors, such as Blanquet, have argued that conferring discretionary powers to agencies actually strengthens the institutional balance.350 As a result such dele-gations would be in conformity with Meroni. These kinds of arguments suffer from one serious defect, since they first require a profound understanding of the content, purpose, and exigencies of the institutional balance. Without getting into these questions it is impossible to conclude that agencification does not harm the institutional balance, let alone that it would be beneficial to that balance. Blanquet notes that identifying the profound meaning of Meroni paradoxically leads to a more flexible application of the doctrine,351 since the politicization of the Commission means that delegations to agencies respect rather than infringe the institutional balance intended by the Treaty authors.352 Secondly, Blanquet claims that the institutional balance is also a vertical principle applicable be-tween the institutions and the Member States. The fact that agencies are often accompanied by the establishment of networks of national authorities then re-inforces this vertical dimension.353 Apart from the question of whether there is indeed a vertical dimension to the institutional balance (cf. section IV 7.2.3.2),

350 See also Majone who argues that a dynamic institutional balance ‘has to do less with pre-serving the competences and privileges of the different institutions than with enhancing their loyal cooperation in the law- making process’, adding that the EU agencies assist this process, thus strengthening the institutional balance. See Giandomenico Majone, ‘Delegation of Regulatory Powers in a Mixed Polity’, (2002) 8 ELJ 3, pp 330– 1.

351 Blanquet, n 279, p 66.352 This argument is based on one of the oft- cited reasons for agencification, ie to help relieve the

Commission of technical tasks, allowing it to focus on its core mandate (cf. section III 2).353 Blanquet, n 279, p 66.


the conclusion that the simple fact of a network being established leads to some kind of reinforcing of the Member States is too simplistic. The fact that networks are established and national authorities being integrated in a European structure precisely raises the question what the effects of agencification on the national ad-ministrations are,354 but does not answer it.

5.4.2.1.2 Rebalancing the institutional balanceThe argument that the institutional balance may be rebalanced is different from the preceding one, since it admits that the ‘current’ institutional balance does not allow further (qualitative) agencification. Authors such as Vos have claimed that

the institutional balance of powers will not be upset so long as a shift of power is ac-companied by reinforcing or rebalancing the existing institutions […] This means that already under the current Treaty provisions, discretionary powers could be delegated to agencies provided that this be accompanied by a strengthening or rebalancing of the existing institutions and functions.355

The last sentence in particular is remarkable in light of the Court’s most clear ob-servations on the institutional balance in Chernobyl, in which it remarked that ‘the institutional balance [is] laid down in the Treaties’.356 If the only (legal) obstacle to further agencification is indeed the institutional balance, overcoming Meroni is quite straightforward (albeit not easy): it involves amending the Treaties and thereby redefining the institutional balance, which a number of authors indeed see as the only possible solution.357

Vos’ argument precisely argues that such a Treaty revision is not necessary since her argument presumes the existence of a dynamic institutional balance. Whether the institutional balance could be a dynamic principle will be dealt with later. For now it suffices to note that a dynamic institutional balance is not apparent from the case law of the Court.358 If the institutional balance is laid down in the Treaties, it would appear that it can only be redefined through a Treaty revision. The mechanisms which Vos proposes to rebalance the institutional balance are a strengthening of Commission supervision and ensuring judicial review by the Court. These two mechanisms correspond to two elements of the Meroni doctrine (cf. section IV 5.3.1), implying that the rejection of a Meroni doctrine in favour of the principle of institutional balance requires a corollary of rebalancing the insti-tutional balance by relying on (parts of) the Meroni doctrine.359 This idea is still

354 On this, see Merijn Chamon, ‘The Influence of ‘Regulatory Agencies’ on Pluralism in European Administrative Law’, (2012) 5 REALaw 2, pp 61– 91.

355 Vos, n 111, pp 131– 2.356 Case 70/ 88, Parliament v. Council, [1990] ECR I- 2041, para 26. 357 See n 302.358 See also the observations by AG Van Gerven in Chernobyl, who found that only the Treaty

authors could readjust the institutional balance. See Opinion of AG Van Gerven in Case C- 70/ 88, Parliament v. Council, [1990] ECR I- 2041, para 6.

359 Vos, n 45, pp 241– 7. An embryonic version of this argument may further be seen with Kapteyn and VerLoren Van Themaat, who note in relation to Meroni: ‘The significance of this ECSC caselaw for the question of delegation of powers in relation to the EC Treaty lies in the principle of the bal-ance of powers, part of which is the supervision of legality by the Court; this balance may not be


worth pursuing since ‘control’ by the High Authority over the Brussels Agencies was indeed a fundamental aspect of the Meroni ruling.360 Lavrijsen and Hancher also suggested that a rebalancing could consist of preserving the prerogatives of the legislator and the Commission and making EU agencies ‘politically and le-gally accountable in a manner comparable to the Commission’.361 This indeed seems to present the core of a rebalancing exercise: introducing mechanisms to the benefit of both the EU legislator and the Commission and ensuring that the net effect of agencification is neutral— the same ‘level of control’ is guaranteed for both agencies and Commission.362

Griller and Orator take a next step, attempting to propose such mechanisms. Which level of steering and control is adequate should depend on the powers conferred on the agencies, according to the authors, whereas the overall result should be an adequate level of legitimacy and institutional balance.363 Here the problematic nature of this argument, or the reference to the institutional balance, is already apparent. The authors present ‘institutional balance’ as a variable de-pendent on steering and control mechanisms, whereas the Court appears to see it as an independent concept written in the Treaties on which steering and control mechanisms are based.

Griller and Orator propose to safeguard the legislator’s prerogatives by intro-ducing the mechanisms which Article 290 TFEU now foresees for the delegated acts (callback and veto options) or by relying on joint decisions of agencies and the Commission.364 The Commission’s prerogatives would be secured by restricting agencification to exceptional cases and by providing the Commission with veto

disturbed. The question arises whether this caselaw forms the basis for a complete prohibition of the delegation of discretionary powers in the functioning of the EC. In the ECSC system, which is char-acterized by strong powers conferred on the Commission by the Treaty (which, however, the Treaty narrowly defines), the strict test for permissible delegation is justified because if the Commission were to transfer its powers to other bodies this would be tantamount to a carte blanche delegation. The EC system often confers very broad powers on the Council and the emphasis of the test lies on the clear and precise delimitation of the powers transferred, on the resulting restriction of free decision making powers and not on the exclusion of all freedom.’ See Paul Kapteyn, Pieter VerLoren Van Themaat, and Laurence Gormley, Introduction to the Law of the European Communities: From Maastricht to Amsterdam, London, Kluwer Law International, 1998, p 245.

360 See also Berger, n 13, p 32.361 Lavrijssen and Hancher, n 277, p 38. See also Griller and Orator, n 273, p 27. Zwart

proposes something similar when he notes that under Meroni, ‘la délégation de pouvoirs discré-tionnaires n’est pas interdite intrinsèquement, mais seulement dans la mesure où cela contrarie la balance institutionnelle au sein de la Communauté. Ainsi, une telle délégation a des chances d’être autorisée dans des domaines qui n’exigent pas qu’un organe fasse la balance entre les objectifs poli-tiques et économiques’. See Zwart, n 311, p 166. This assumes that the legislator will have balanced all political and economic objectives, leaving no balancing power for the administration.

362 Similarly, see Vincent Correia, ‘La Coopération de l’Agence Européenne de la Sécurité Aérienne des États Membres et des Industriels au Sein du Système Européen de la Sécurité Aérienne’, (2010) 46 CDE 5- 6, pp 599– 601.

363 Griller and Orator, n 273, p 27.364 A similar suggestion is made by Rosenberg, according to whom no discretionary choices

are made by EU agencies as long as the Commission has the power to amend their decisions: see Rosenberg, n 278, p 520. However, it is not entirely clear how the political accountability of agen-cies is promoted by having them adopt joint decisions together with the Commission.


rights. Griller and Orator see fewer problems as regards legal scrutiny since the Treaty of Lisbon already took a ‘step in the right direction’ in this area. Lastly, political accountability could be ensured by making agencies answerable to the Parliament just like the European Central Bank (ECB) and enhancing surveil-lance mechanisms of the Parliament, Council, and Commission. Under these conditions, Griller and Orator propose discretionary powers could be conferred on EU agencies.365 Although their more elaborate reasoning increases the appeal of this argument, there are still two principal counter- arguments.366 Firstly, by reinterpreting the Meroni doctrine, Griller and Orator have shifted the attention away from what the Court actually said in Meroni, that is, that ‘[a] delegation of powers […] which implies a wide margin of discretion […] cannot be considered as compatible with the requirements of the Treaty’.367 If one takes the position that Meroni applies to EU agencies, it is difficult to circumvent this unambiguous statement. Furthermore, as was noted previously, this element of the Meroni doc-trine is generally perceived as its core. Going to these lengths to argue that the core of the Meroni doctrine should not be upheld, but that Meroni is still ‘good law’ and should still be applied to EU agencies, raises the question of why one would look to Meroni in the first place to frame the process of agencification. Secondly, the mechanisms proposed by Griller and Orator, if they are to be effective, pre-suppose that the Commission, Council, and Parliament have sufficient knowledge to exercise their veto and/ or callback powers. This may relate difficultly with one important ‘driver of agencification’, that is, the absence of expertise within the Commission and the need for it at EU level.

The arguments presented in favour of rebalancing the institutional balance should thus be rejected if the institutional balance indeed has primary law status. These arguments also seem to trifle with the Court’s ruling in Meroni, to the extent that they refer, in a first instance, to Meroni to make legal sense of agenci-fication, only to fundamentally alter the meaning of that ruling in a second step. Lastly, it may be noted that the Court itself in Short- selling did not follow this route to come to a Meroni- light, since in that case the Court can be faulted with completely ignoring the institutional balance (cf. section IV 7.2.4.1.3).

5.4.2.2 The meaning of ‘discretionary power’ in MeroniOne of the difficulties related to actually applying a Meroni doctrine in practice is that the Court did not draw its dividing line between an acceptable delegation of powers and an unacceptable delegation with chirurgical precision.368 In its ruling the Court juxtaposes executive with discretionary powers and clarifies that the use of the latter presupposes balancing different policy objectives allowing a body to

365 Griller and Orator, n 273, pp 27– 9.366 Chamon, n 332, p 292.367 Case 9/ 56, n 223, p 154.368 In a way this is quite similar to the notion of ‘essential elements’ introduced by the Court in

Köster and now enshrined in primary law in Article 290 TFEU. See Merijn Chamon, (2013) 50 CMLRev 3, pp 849– 60.


formulate economic policy. The difficulty in drawing this dividing line has there-fore also inspired different authors to argue for a less restrictive application of Meroni.

Curtin and van Ooik note in this regard that Meroni does not prevent further agencification so long as the agencies concerned are bound to a strict mandate, citing the OHIM as an example.369 Whether ‘a strict mandate’ may be taken as a proxy for ‘the exercise of strictly executive powers’ is debatable,370 but the argu-ment by Curtin and van Ooik makes clear that there is a great deal of flexibility in how one could define ‘discretionary powers’. Schneider argues as such when he notes that legislative discretion is something different from the administrative discretion which allows an administration to make ‘context- adequate decisions of no actual political importance’.371 Priebe makes a similar distinction between non- delegatable competences of a political nature and the powers of administra-tive discretion which should be delegatable but which still go beyond the notion of executive powers as presented by the Court in Meroni. Such powers are not confined to ‘strictly executive powers’ or having to do with ‘mere accountancy procedures’.372 Bartodziej should be situated on the same line, since he finds that the delegation of powers implying only an ‘unpolitical administrative discretion’ is in conformity with Meroni. Unlike Schneider, Bartodziej does support his argu-ment by referring to the Meroni ruling where the Court did not simply juxtapose executive with discretionary powers, but with powers which could be exercised as one sees fit, implying a large discretion and allowing for the adoption of a genuine economic policy.373 Taking these three elements together, Bartodziej suggests that in theory an independent Cartel Office (the topic of his study) could be established in line with Meroni because such an office only has to apply competition rules without having to balance different Treaty objectives, noting that no legislative powers could be delegated.374 Bartodziej thus suggests that the competences which cannot be delegated under Meroni are legislative competences, further relying on the differences between the ECSC and the E(E)C Treaties to mitigate the Meroni doctrine.

But can the prohibition on delegating discretionary powers be reinterpreted as a prohibition on delegating legislative powers? According to Bartodziej, this was

369 Deirdre Curtin and Ronald van Ooik, Revamping the European Union’s Enforcement Systems with a View to Eastern Enlargement, WRR Working Documents, The Hague, October 2000, W 110, 62. In this regard Zwart notes that even if Meroni applies to the agencies, the legis-lator could still establish genuinely regulatory agencies, in line with Meroni, if the US example on Congress’ intelligible standard (cf. section V 1.3) is followed. Read together with Zwart’s statement cited in n 361, this would mean that agencies could adopt legislation (in a substantive sense) which does not imply the balancing of more than one political or economic objective.

370 See Chamon, n 332, p 293.371 Schneider, n 287, p 38. Schneider and Bartodziej both seem to hold that it is possible to

separate technical from political matters. Although not in the context of delegation, AG Léger made a similar distinction between ‘political discretion’ and ‘technical discretion’. See Opinion of AG Léger in Case C- 40/ 03 P, Rica Foods NV v. Commission, [2005] ECR I- 6814, paras 45– 6.

372 Priebe, n 27, p 112.373 Bartodziej, n 281, p 308. 374 Ibid, p 309.


already apparent in Meroni and the Court further confirmed this in its opinion on the draft Agreement establishing a European laying- up fund. As will further be dealt with, the Court’s observations on the nature of the powers of the laying- up fund were rather problematic (cf. section IV 5.5.1.1). Bartodziej’s assertion that the Court actually envisaged ‘legislative powers’ rather than ‘discretionary powers’ in Meroni then seems difficult to accept, given that the Court used another criterion to draw this dividing line in Köster,375 that is, the distinction between essential and non- essential elements. Bringing together Meroni and Köster,376 Bartodziej’s point of view would further also implicate that ‘pursuing an economic policy’ (referred to in Meroni) can only be done through legislative action and that there is no policy- making when spelling out the non- essential elements of a body of rules.377

Lastly, in the jurisprudence of the Court, the germs of a re- interpretation of discretionary choices in relation to EU agencies could be noted before Short- selling. In Schräder, the Court of First Instance repeated its traditional obser-vations on the discretion which the EU institutions may exercise, but now in relation to the CPVO: ‘It is clear from the case law of the Court of Justice that where a Community authority is called upon, in the performance of its duties, to make complex assessments, it enjoys a wide measure of discretion, the ex-ercise of which is subject to limited judicial review.’378 In a subsequent case related to the ECHA, the General Court further noted that ‘[as regards the proportionality of the contested decision,] it must be acknowledged that the ECHA has a broad discretion in a sphere which entails political, economic and social choices on its part, and in which it is called upon to undertake complex assessments’.379

The present section has shown that there is scope in devising a Meroni- light by clarifying what powers should fall in and outside the category of discretionary powers. This issue will be returned to later (see section IV 5.6.2.2.2). The Court itself in Short- selling has indeed followed this route, upholding Meroni and at the same time concluding that the powers of the ESMA at issue were merely ex-ecutive, or rather ‘precisely delineated’. However, the Court did not lay down a clear new rule (such as inter alia proposed by Bartodziej) to identify ‘discretionary powers’, and neither did it refer to the General Court’s existing jurisprudence on agency discretion.

375 Case 25/ 70, n 314.376 Following Meroni, but prior to Köster, Rabe suggested that the limits to the delegations under

Article 155 EEC could be similar to those set out by the Court in Meroni: see Hans- Jürgen Rabe, Das Verordnungsrecht der Europäischen Wirtschaftsgemeinschaft, Hamburg, Kommissionsverlag Ludwig Appel, 1963, p 109 at footnote 37.

377 However, in Biocides, AG Cruz Villalón suggested that the Commission could have limited policy- making capacity when adopting delegated acts: see Opinion of AG Cruz Villalón in Case C- 427/ 12, Commission v. Parliament and Council, ECLI:EU:C:2013:871, para 76.

378 Case T- 187/ 06, Schräder v. CPVO, [2008] ECR II- 3151, para 59.379 Case T- 96/ 10, Rütgers Germany GmbH e.a. v. ECHA, ECLI:EU:T:2013:109, para 134 (em-

phasis added). Related to the EASA, see also Case T- 102/ 13, Heli- Flight GmbH & Co. KG v. EASA, ECLI:EU:T:2014:1064, paras 74 and 89– 90.


5.4.2.3 ConclusionDevising a Meroni- light has been quite a popular endeavour in legal doctrine. Drawing on Meroni, a sanctioning by the Court of the process of agencification is found, and by construing a light version the resulting framework for agencifica-tion becomes so generous that there hardly seem any real limits left. However, although these arguments all have a certain appeal, they are often insufficiently convincing. Indeed, it does not simply suffice to point to the differences between the Brussels Agencies in Meroni and the current- day EU agencies to argue that Meroni should not be applied as strictly to the latter. Others point to the at times unclear wording of the Court’s ruling in Meroni, suggesting alternative inter-pretations to concepts such as ‘discretionary powers’ or emphasizing the Court’s reliance on the balance of powers, subsequently reinterpreted as the vague notion of institutional balance. All in all, most arguments were found not to be convin-cing enough, but the unclear delimitation in the (simplistic) dichotomy between executive and discretionary powers does provide an opportunity for further clari-fication. In clarifying, account could then also be taken of the differences in legal context between the EU and the ECSC. Indeed, the Court in Short- selling has foremost relied on the unclear distinction between executive and discretionary powers to come to a Meroni- light.

5.4.3 The institutions’ understanding of MeroniThis debate in academic doctrine on the relation between Meroni and the agen-cies raises the question of what the EU institutions’ view on that relation was pre- Short- selling. However, as noted in the chapter on political limits, the institu-tions’ primary concern in the process of agencification has been to secure their own interests rather than ensuring the process is in line with primary law or the Court’s jurisprudence. It should therefore be noted from the outset that it is diffi-cult to identify a single coherent view for each institution, even if the three main institutions do have in common that they found Meroni to be applicable to the EU agencies.

5.4.3.1 The CommissionIn its draft interinstitutional agreement of 2005,380 the Commission went into detail by noting that agencies could be granted the power to adopt individual decisions but that this power should be ‘limited to applying the rules of secondary legislation to specific cases, in accordance with the institutional system and the case law of the Court of Justice’,381 referring to Meroni (and Romano). Following a negative approach, the Commission further clarified what agencies could not do: (i) adopt general regulatory measures, (ii) exercise decision- making powers

380 See also Working Group 3a of the European Commission, June 2001, ‘Report on Establishing a Framework for Decision- making Regulatory Agencies’, SG- 8597/ 01- EN, p 11.

381 European Commission, COM (2005) 59 final, p 5.


requiring them to arbitrate between conflicting public interests or to exercise pol-itical discretion, and (iii) exercise tasks with respect to which the Treaties have conferred direct decision- making powers to the Commission.382 Already from these three ‘red lines’ one may discern the Commission’s attempt to use Meroni to safeguard its own institutional interests.

During legislative negotiations, the Commission has in the past rejected a number of amendments of the European Parliament on the ground that they would be in violation of Meroni. Still the Commission’s rejections are rarely mo-tivated beyond the observation that granting ‘this or that power’ is prohibited by Meroni.383 The negotiations on the third energy package, which saw the establish-ment of the ACER, are a good illustration of this. As to what would become the Directive 2009/ 72, the Parliament had proposed to grant the ACER the power to adopt guidelines for the Member States and to act as the competent authority for regional markets, in lieu of the national authorities.384 The Commission noted that, in line with Meroni, the ACER should not be allowed to adopt (binding) guidelines but rather only (non- binding) guidance and that it would be contrary to Meroni to have the ACER replace the national authorities as competent auth-ority,385 presumably because a competent authority should be empowered to adopt regulatory measures.

As regards the proposal to establish the ACER itself, the Parliament had inter alia suggested amendments to the effect that the ACER (i) would be tasked to issue decisions addressed to the TSOs on all technical matters pertaining to the good functioning of the internal energy market; (ii) would receive a delegated power from the Commission to take suspensive decisions; (iii) would be empow-ered to fine TSOs; (iv) would adopt the network codes itself; and (v) would take binding decisions on access and use of connected transmissions systems involving several Member States if the authorities of those Member States cannot come to a joint agreement themselves.386 On these five amendments, the Commission remarked that pursuant to Meroni, (i) the ACER’s powers should be laid down in detail rather than simply granting it power to issue opinions ‘on all technical matters’; (ii) the ACER could not receive such a power from the Commission since the Commission does not itself possess it; (iii) only the Commission can adopt

382 Ibid, pp 11– 12.383 In general the Commission often invokes Meroni to argue for or against different policy

options. See eg European Commission, SEC (2010) 1126, p 54; European Commission, SEC (2009) 837, p 43. In relation to the ESMA, see European Commission, SWD (2012) 22 final; European Commission, SEC (2010) 678, pp 13 and 30– 1. In relation to the overhaul of the data protection rules, the Commission contemplated the establishment of a European Data Protection Agency but found it problematic in the light of Meroni: see European Commission, SEC (2012) 72 final, p 72.

384 See amendments 11 and 55 proposed by the Parliament. For the latter’s Resolution, see OJ 2009 C 286E/ 106.

385 European Commission, 16 July 2008, ‘Communication on the action taken on the opinions and resolutions adopted by the Parliament at the June I and II 2008 part- sessions’, SP (2008) 4439.

386 See amendments 19, 26, 30, and 42 proposed by the Parliament. For the latter’s Resolution, see Legislative Resolution, OJ 2009 C 286E/ 149.


fines; (iv) the Commission should adopt network codes; and (v) the ACER could not exercise discretionary powers.387

The proposed amendments did not make it to the final regulation, apart from the ACER’s power to step in when national authorities cannot agree, which is pro-vided in Article 8(1) of the ACER Regulation. Interestingly, only a couple of years following the Commission’s rejection, relying on Meroni, of granting the ACER such powers, the Commission apparently changed its mind when it proposed itself that the ESAs be empowered to take decisions in lieu of national authorities (when the latter cannot agree among themselves) and agreed (but did not propose itself) to grant ESMA the power to impose sanctions on market operators. This example shows how (inevitably) the Commission’s understanding of Meroni is subject to evolution.

5.4.3.2 The CouncilThe Council(’s) (Legal Service’s) understanding of Meroni has also evolved over the years. In its 1985 opinion on the OHIM proposal, the CLS remarked that Meroni as such did not apply since no ‘true delegation’ of powers was at issue. Instead new powers, which were not vested in any of the institutions, were created and conferred on the OHIM. This is indeed in line with the argument presented by Ehring (cf. section IV 5.4.1.2). At the same time, the Legal Service noted that Meroni did not simply incarnate a number of rules applicable to ‘true delegations’. As a result, and unlike under Ehring’s argument, part of the Meroni doctrine also applied to conferrals (rather than delegations) of powers. Because of this Meroni remained relevant for empowering an agency such as the OHIM, as it required that (i) effective legal protection needed to be secured and (ii) the structure and institutional balance of the Community could not be distorted.388

As regards the second element, the CLS rightly noted that it should first be ascertained whether the powers at issue should not be conferred on one of the institutions provided by the Treaties and, if not, whether they may be conferred on a body other than the existing institutions.389 Remarkably, the CLS did not refer a single time to the core of Meroni, that is, the prohibition to confer discretionary powers (cf. section IV 5.3.1). When the Legal Service further verified whether the powers at issue should not be accorded to existing institutions, it noted that Köster did not compel the Council to empower the Commission under Article 155 EEC, even if Romano could be read as imposing such an obligation. As a result, the CLS found that the powers at issue could be conferred on the OHIM.390 Following the SEA, however, Treaty revisions have strengthened the Commission’s executive position.391 Ultimately, the Treaty of Lisbon has divested the legislator, in Article 291 TFEU, of the power to determine itself the appropriateness of empowering

387 European Commission, n 385, pp 31– 6.388 Legal Service of the Council of the European Communities, 6 June 1985, Doc 5837/ 85, p 6.389 Ibid, p 7. 390 Ibid, pp 7– 8.391 See Merijn Chamon, ‘EU Agencies between “Meroni” and “Romano” or the Devil and the

Deep Blue Sea’, (2011) 48 CMLRev 4, pp 1066– 8.


the Commission. Instead, Article 291 TFEU has ‘objectivized’ this decision (cf. section IV 5.7.2.1).392

Recently, in one of its opinions on the Single Resolution Mechanism (SRM) for the Banking Union, the CLS changed course. In its opinion on the delegation of powers to the SRB it summarized the Meroni jurisprudence as follows:

i) no delegation can be presumed […]; ii) a delegation of powers cannot be excluded even in the absence of a specific basis for it in the Treaty; iii) any delegation of powers where the conferred powers are broader than those of the delegating authority is unlawful; iv) a delegation involving “discretionary power implying a wide margin of discretion which may, according to the use which is made of it, make possible the execution of actual eco-nomic policy” would imply an illegal transfer of responsibility by substituting the choices of the delegator by those of the delegate and by altering the balance of powers thus doing away with the guarantee granted by the Treaty to undertakings; and v) powers to carry out assessments under own authority should be subject to precise rules in order to avoid arbitrary results and to make review of those assessments possible.393

Unlike in its opinion on the OHIM, the CLS did not make a clear distinction between a conferral and a ‘true delegation’; also unlike in its OHIM opinion it further primarily focused on the prohibition to delegate discretionary powers, construing the latter narrowly, perhaps influenced by the opinion of the AG in Short- selling. In addition, whereas in its OHIM opinion it raised the question of whether the OHIM’s proposed powers should not be granted to any of the exist-ing institutions, it ignored this question in the opinion on the SRM. Evidently, had the Legal Service entered into the question, it could no longer have concluded with the same ease that granting implementing powers to the Commission is purely facultative, in view of the changes brought to Article 155 EEC by subse-quent Treaties.

5.4.3.3 The ParliamentIt is difficult to identify a clear position on the Meroni jurisprudence for the Parliament, if only because of its very diverse composition and the changing lead-ing committee drafting the resolutions for the plenary. In addition, there is the fact that, just like the other institutions, the Parliament is foremost concerned with its own institutional interest. Lastly, it would further seem that the Parliament is still outplayed by the Commission and Council when it comes to expertise, also in relation to agencies.

For instance, in the explanatory memorandum for the 2008 Resolution on a strategy for the future settlement of the institutional aspects of Regulatory Agencies,394 the latter are confused with the US regulatory agencies. The memo-randum provided that EU agencies ‘are assigned competences which resemble

392 See also Merijn Chamon, ‘Comitologie onder het Verdrag van Lissabon’, (2013) 61 SEW 2, p 68; Claude Blumann, ‘Le système normatif de l’Union européenne vingt an après le traité de Maastricht’, (2012) 19 RAE 2, p 251.

393 Legal Service of the Council of the European Union, 7 October 2013, Doc. 14547/ 13.394 See the chapter III on political limits, n 167.


not only the regulatory role played by the administration […] but also the role of the judicial authorities in imposing penalties’.395 In reality, at that time no single EU agency could impose penalties. As regards the agencies’ powers, it was simply noted that agencies should only adopt individual decisions and that ‘[t] he Court of Justice has issued a judgment on this subject (the Meroni doctrine, Case 98/ 80, Romano)’.396 In the same procedure, the Committee on Economic and Monetary Affairs confused the issue of empowering an agency with that of establishing an agency by suggesting that ‘all agencies […] must be established in accordance with the Meroni judgement’.397

Some confusion also seems to exist on the qualification of EU bodies as EU agencies. When the Commission suggested merging the European Network and Information Security Agency (ENISA) with the proposed European Electronic Communications Market Authority (ECMA), the Budgetary Control Committee remarked: ‘It is unnecessary to create a new agency. Instead, a body such as an enhanced ERG embedded in EU law should be established […] Article 95 […] gives ample grounds to give more powers to the ERG by giving it legal personality and therefore independence.’398 However, how the committee saw the difference between an enhanced European Regulators Group (ERG) with legal personality and an EU agency is unclear. Turning to Meroni, the Parliament deplored that it prevented the establishment of a strong ACER.399

In other areas, however, the Parliament pleaded for a more rigorous applica-tion of Meroni, as was the case when it discussed the rules on access to docu-ments, pre- Regulation 1049/ 2001. Here the Parliament’s Committee on Petitions invoked Meroni to ensure the ‘supreme purview’ of the Court as a minimum condition.400

Another example may be found in one of the 1993 Parliamentary resolutions mentioned previously, where the Parliament explicitly referred to the Meroni doc-trine to ensure that the Commission would remain responsible for the agencies before the Parliament.401

395 See the explanatory memorandum annexed to the motion for a European Parliament Resolution under procedure 2008/ 2103(INI).

396 See ibid.397 See suggestion n° 10 made by the Economic and Monetary Affairs Committee under pro-

cedure 2008/ 2103(INI).398 See Amendment 1 proposed by the Budgetary Control Committee to the Committee on

Industry, Research and Energy in first reading in procedure 2007/ 249(COD).399 See the explanatory memorandum annexed to the draft European Parliament Resolution

under procedure 2007/ 197(COD).400 Article 191a EEC (now Article 15 TFEU) at that time only referred to access to the docu-

ments of the Parliament, Council, and Commission. See Opinion by the Committee on Legal Affairs and Citizens’ Rights for the Committee on Petitions on the special report from the European Ombudsman to the European Parliament following the own- initiative inquiry into public access to documents. The final Resolution adopted by the Parliament did not take over this detailed assess-ment but did provide (in point 13) that the Parliament ‘[f ] inds that Articles 164, 173 and 175 of the EC Treaty, in particular, the principle of interinstitutional balance and the caselaw of the European Court of Justice should be regarded as providing a legal framework’.

401 See section III 4.3. Similarly, even if it did not explicitly invoke Meroni, see Résolution, J.O. 1968 C 108/ 37, point 10.


5.4.3.4 Concluding remarksWhile the three main institutions do hold Meroni to be applicable to the EU agencies, a clear and consistent understanding of that jurisprudence on their part cannot be identified. Since political and institutional praxis has informed the institutions’ understanding of Meroni, rather than the other way around, that understanding has been found to be subject to evolution. In any case it is clear from legislative practice itself that the institutions have applied a Meroni- light.

5.4.4 ConclusionDoes it make sense to evaluate EU agencies under Meroni? The arguments raised in legal doctrine against this have not been found to be totally convincing. In part this is because the rules constituting the Meroni doctrine are often erroneously taken as a monolith. Other arguments seem difficult to support because they identify the institutional balance as the core of Meroni, whereas the analysis of the ruling and the AG’s Opinion point to the more concrete notion of control as the primary concern of the Court.

The institutions themselves have not been found helpful in this debate since they have adapted their understanding of Meroni to existing practice and/ or to the benefit of their own institutional interest.

5.5 Meroni in the subsequent case law of the Court

To identify an EU Meroni doctrine, the subsequent references to this case in the Court’s jurisprudence should also be looked into. Generally four types of cases may be identified. Evidently one group of cases is related to the problem raised in Meroni. Following the Court’s judgment, Meroni and other companies tried to rely on the ruling to get out of the obligation to pay their contributions. In a second group of cases some procedural questions left open in Meroni were raised.402 In a third group the Court recalled its observation in Meroni on the Commission’s discretion to balance the different objectives of the ECSC (and

402 These questions were foremost concerned with the objection of illegality, since Article 36 ECSC provided this was only open to undertakings in the case of individual decisions imposing fi-nancial sanctions. In Meroni, however, the Court had ruled that Article 36 ECSC expressed a general principle. Because of this Boulouis remarked in 1958 that ‘[l] ’exception d’illégalité forme la partie la plus remarquable des décisions sur les affaires 9 et 10- 56’. See Jean Boulouis, ‘Jurisprudence de la Cour de Justice de la Communauté Européenne du Charbon et de l’Acier’, (1958) 4 AFDI, p 317. Subsequently Meroni was invoked by Germany in Case 3/ 59, Germany v. High Authority, [1960] ECR 53. The Court however ruled that privileged parties could not avail themselves of this remedy as they should challenge the contested general act directly. Although Article 184 EEC referred to the alleged illegality of a regulation, the Court relied on Meroni to conclude that this should be understood as any act of general application: see Case 92/ 78, Simmenthal v. Commission, [1979] ECR 777, para 39. In Arizona Chemicals e.a. the General Court referred to Meroni to note that a successful plea of illegality could not result in the annulment of the general act but at the most of that of the contested act: see Case T- 369/ 03 R, Arizona Chemical BV e.a. v. Commission, [2004] ECR II- 205, para 66.


later EEC).403 Only the last group of cases in which Meroni was referred to in the context of a delegation of powers will be commented upon further.

5.5.1 Meroni in the context of delegationThe academic attention to the Meroni ruling stands in contrast with the Court’s own reliance on that case in its subsequent delegation jurisprudence. This may have indeed confirmed authors such as Dehousse and Geradin (cf. section IV 5.4.1) in their belief that Meroni was not relevant to the current process of agencification.

5.5.1.1 Opinion 1/ 76 on the draft agreement establishing a laying- up fundThe first major case which lent itself to an application of the Meroni doctrine was the Court’s Opinion on the draft Agreement establishing a European laying- up fund.404 The Commission had queried the Court on the powers which would be granted to the fund.405 The Agreement would have been a mixed agreement with Switzerland and the fund’s purpose was to give out subsidies for the temporary and voluntary laying- up of excess transport capacity in the Ruhr and Moselle basins. Its income would have been procured from contributions levied on the vessels using the inland waterways. The mechanism proposed to stabilize the transport market was therefore quite comparable to the mechanism underlying the equal-ization scheme in Meroni. However, whereas under the latter scheme the Caisse had merely received general guidelines from the Council and High Authority, the Fund would have found itself more restricted initially. As the Court noted in its analysis of the agreement,

[t] he provisions of the Statute state precisely and in detail which categories of vessel are subject to the proposed system. In addition they determine the contribution to be levied and the conditions of payment and fix the basic rate for contributions for the first year […] The basic rate and the adjustment coefficients [for subsequent years] may be altered by the Supervisory Board within the prescribed conditions.406

Article 10(3)– (5) of the Statute provided that the Supervisory Board could unani-mously decide on such alterations.407 Following Article 27 of the Statute the Supervisory Board would have been composed of representatives of the parties whereby the European Commission would not have held voting rights on behalf of the EEC.

The Commission argued that the Agreement would be compatible with the Treaties. Firstly, the Commission claimed that since the Member States could

403 See inter alia Case T- 243/ 94, British Steel plc v. Commission, [1997] ECR II- 1887, para 105.404 Following the negative opinion, the project of a laying- up fund was abandoned. See

Stanislas Adam, La procédure d’avis devant la Cour de justice de l’Union européenne, Bruxelles, Bruylant, 2011, p 407.

405 Opinion 1/ 76, re the draft Agreement establishing a European laying- up fund for inland waterway vessels, [1977] ECR 741, p 747.

406 Ibid, pp 744– 5 (emphasis added).407 See European Commission Proposal for a Council Regulation, OJ 1976 C 208/ 2.


establish and empower an international organization such as the Fund, so should the Communities, including ‘the possibility of granting to organs set up within the framework of an international agreement the power to take decisions which are directly applicable within the legal system of the Community’.408 Secondly, the Commission claimed that the EEC institutions ‘will be able to exercise indirectly an appropriate influence on the action of the Fund because of the membership of the Supervisory Board of the Fund and the possibility of appeal to the Council of the Community with regard to questions of special interest’.409 The latter possibility was not provided for by the Statute but was to be found in the proposal for a regulation concluding the agreement. Article 5 thereof provided that the Council would adopt positions on the matters of principle to be decided by the Supervisory Board, to which the (EEC) members of the Supervisory Board were obliged to adhere. Of course this fundamentally changed the way in which the Board would take its decisions, and the Swiss delegation had objected to this provision in the draft regulation.410 The Council rejected the Commission’s reliance on public international law, arguing that the question was of a constitutional nature and the answer thereto should therefore be found in the Treaty.411

Interestingly, only the Danish government explicitly relied on the Meroni ruling. According to Denmark, because ‘[t] he organ to which the powers are transferred must remain under the control and the supervision of the authority in which the original power was vested’, the conditions for acknowledging the neces-sity of the delegation should ‘become more rigorous as the degree of autonomy of the [delegate] authority increases’.412 Referring to Meroni, the Danish government ruled out the possibility of delegating discretionary powers to the Fund.

However, the Court’s answer ended in an anti- climax. Even though the Supervisory Board had clearly been given discretionary powers, the Court ruled that ‘the provisions of the statute define and limit the powers which the latter grants to the organs of the fund so clearly and precisely that in this case they are only executive powers’.413 Although the Court had already concluded that the agreement was incompatible with the Treaties because of the Fund’s institu-tional setup, the brevity in its answer to the question on the powers of the Fund remains unfortunate. The Court’s answer was furthermore unsound. By unani-mous decision, the Supervisory Board could alter the basic rate and the adjust-ment coefficients. Even if the Fund could not remedy structural imbalances or set a permanent minimum level of freight rates in periods of slack demand, it is hard to see how these powers could not be seen as discretionary in nature.

408 Opinion 1/ 76, n 405, p 748. 409 Ibid. 410 Ibid, p 754.411 The Council implicitly relied on the institutional balance. See ibid, p 751.412 Ibid, p 752.413 The Court further noted that ‘subsequent amendments by decision of the Supervisory Board

must either remain within certain limits or result from a unanimous decision’. Ibid, para 16. It is clear that the second part of this observation is completely out of place, since it suggests that the discretionary nature of the powers conferred would depend on whether decisions are made pursuant to (simple) majority or unanimity.


As a result many authors were puzzled. On the face of it, the Court seemed to follow and confirm its Meroni jurisprudence by stressing that the Fund’s powers were only executive in nature. Some ignored the Court’s dubious qualification of the powers of the Fund,414 but Kapteyn expressed doubts on whether the Fund’s powers would merely have been executive and concluded that the Court had departed from its Meroni jurisprudence. At the same time Kapteyn warned against making rash conclusions for delegations within the EU, since the Court had consistently referred to the transfer rather than dele-gation of powers.415 Weis was even more critical and rejected the Court’s con-clusion on the ‘executive’ nature of the Fund’s powers, further deploring that the Court itself did not refer to Meroni.416 According to Weis, the Court clearly relaxed the criteria it had laid down in Meroni.417 Given the Court’s insufficient reasoning, Weis noted that the only conclusion to be made from Opinion 1/ 76 is that the boundaries between executive and discretionary powers had shifted.418

Whether Opinion 1/ 76 confirmed or relaxed Meroni is not the only contro-versy in legal doctrine. Relying on Kapteyn, Lauwaars distinguished the internal from the external delegation, inter alia because the latter allegedly did not raise the question of supervision by the EU institutions.419 This last point is debatable, since the Court gave a negative opinion precisely because the institutions were not (adequately) represented in the Fund. According to Hartley the distinction between internal and external ‘delegations’ also has consequences for the nature of the powers which could be delegated or transferred; he notes that ‘[o] ne would […] expect that the competence of the Community to delegate powers to a supra-national body would be at least as extensive as its ability to delegate internally’.420 According to Lauwaars the reason for this is that an external delegation would be governed by principles of international public law.421 However, this argument cannot be accepted.422 Although the EU’s external relations are evidently (also)

414 Griller and Orator, n 273; Frank, n 273, p 124.415 Paul Kapteyn, ‘Het Advies 1/ 76 van het Europese Hof van Justitie, de externe bevoegdheid

van de Gemeenschap en haar deelneming aan een Europees oplegfonds voor de binnenscheepvaart’, (1978) 26 SEW 5/ 6, pp 365– 6. Similarly, see Schneider, n 287, p 39.

416 Hubert Weis, ‘Anmerkung zum Gutachten 1/ 76’, (1977) 12 Europarecht 3, pp 283– 4. Philip also found that the solution adopted by the Court was ‘loin d’être évidente et certaine’ and therefore would have merited greater elaboration. See Christian Philip, ‘A Propos de l’Avis 1/ 76 de la Cour de Justice des Communautés Européennes: Réflexions concernant le champ de la com-pétence externe de la Communauté’, (1978) RMC 214, p 61.

417 Hubert Weis, ‘Außervertragliche Institutionen der Gemeinschaft’, (1980) 15 Europarecht 3, pp 283– 4. Fischer- Appelt also sees Opinion 1/ 76 as easing the Meroni ruling: see Fischer- Appelt, n 8, p 108.

418 Weis, n 417, p 284.419 Lauwaars, n 15, p 373.420 Trevor Hartley, ‘Opinion 1/ 76 of April 26 1977, Laying- up Fund for Inland Waterway

Vessels’, (1977) 2 ELRev 2, p 279.421 Lauwaars, n 15, pp 384– 5.422 Weis also rejects the view, referring to Lauwaars, that the easing of the Meroni criteria may

be explained by reference to the ‘völkerrechtlichen Besonderheiten des Stillegungsfonds’. See Weis, n 417, p 284.


governed by rules of international public law, the EU’s external action should also conform with the basic Treaties themselves, as evidenced by the function of the opinion procedure itself.

As a result, Hartley’s conclusion should in fact be reversed: precisely because the EU would no longer be able to take unilateral action and because external actors cannot be put under the supervision of EU institutions, the rules on ex-ternal delegation would have to be applied more strictly. Again some of the rules set out in the Meroni ruling regarding internal delegation are generally applic-able in any Rechtsstaat. Changing the context from an internal delegation to an external delegation would therefore not mean that all these rules become irrele-vant.423 Reasoning along the lines set out in the observations of the Danish gov-ernment, a proportionality test would indeed have to be applied very strictly: only if an international agreement is the only course of action to realize a given policy objective could the transfer of powers to an international body be justified.424 However, this would still leave open the question of how adequate supervision by the delegate authority could be organized.

What then to make of Opinion 1/ 76 and its consequences for the problem of delegation of powers within the EU? Unfortunately the conclusion is that the Opinion does not bring much to the table. Arguably, one could conclude that the Opinion overruled Meroni and also allows for further agencification. Upon closer scrutiny, however, only the Court’s reference to ‘executive powers’ links up with Meroni, while it did not rely itself on that ruling, although it was invited by the Danish government to do so. Furthermore, as Kapteyn noted, the Court consist-ently referred to the notion of a ‘transfer’ rather than a ‘delegation’ of powers as in Meroni. Lastly, the Court’s observation that the powers in question were ‘merely executive’ was also inadequately reasoned. The Opinion therefore does not offer a sound basis for making conclusions on Meroni or the delegation of powers within the EU.

Finally, this raises the question of whether Opinion 1/ 76 should actually be categorized under the Court’s delegation jurisprudence. It is clear that the con-trol by the delegating authority which the Court emphasized in Meroni is highly problematic for external bodies under international law, much more so than for external bodies under national private law. Although Gautier believes that the control by the delegating authority could take the form of its participation in the organs of the international organization,425 participation does not readily amount to control. As will be noted later (see section IV 5.7), the transfer of powers to international organizations is not qualified by the Court in terms of delegation,

423 See also Henry Schermers, International Institutional Law, Leiden, Sijthoff, 1972, Vol 1, pp 152– 4.

424 Bülow makes the same argument and (correctly) reverses the argument made by Hartley. See Erich Bülow, ‘Die Gutachtentätigkeit des EuGH’, in Grewe (ed), Europäische Gerichtsbarkeit und nationale und nationale Verfassungsgerichtsbarkeit: Festschrift zum 70.Geburtstag von Hans Kutscher, Baden- Baden, Nomos, 1981, p 75.

425 Yves Gautier, ‘Organisations Internationales’, in Gavalda and Kovar (eds), Répertoire de droit communautaire, Paris, Dalloz, 1996, Vol. III, p 13.


and instead it foremost stresses the continued autonomy of the EU as a red line which cannot be crossed.

5.5.1.2 Reliance of Advocates General on MeroniUnlike the Court itself, the Advocates General (AG) have relied more extensively on Meroni in their analyses. In the PVC case on appeal, one of the questions raised was whether the Commission as a collegial body could, after having adopted one authentic language version of a decision, delegate the power to adopt the authentic language versions in the remaining official languages to the Commissioner for Competition. Advocate General Van Gerven in his conclusion set out a frame-work, guided by the Court’s established jurisprudence.

In the first place the AG relied on Akzo, in which the Court first found that the delegation to an individual Commissioner is possible if it does not divest the college of competences.426 Secondly, the Court noted that a system whereby measures of administration may be adopted by a single member ‘appears neces-sary, having regard to the considerable increase in the number of decisions which the Commission is required to adopt, to enable it to perform its duties’.427

In PVC, AG Van Gerven concluded that these delegations are lawful if the Commission itself retains ultimate responsibility and if there is no negative effect on legal protection.428 The AG continued by setting out a conceptual framework on delegation involving three levels of decision- making. The first level involves the exercise of power at the political level, the second comprises measures of management and administration, and on a third level there are the measures of practical execution.429 According to the AG the power to take the decisions at the first level could not be delegated; thereby referring to the Meroni ruling in a footnote.430

AG Van Gerven’s conceptual framework is interesting since it allows for more gradation in assessing powers than the rather simplistic dichotomy between ex-ecutive and discretionary powers or the equally difficult metaphor of a ‘sliding scale’.431 By referring to Meroni the AG rightly alludes to the fact that the rules on internal Commission delegations should not be isolated from general rules on delegation.

The Ismeri case resulted from the scandals related to the MEDA programme (Mésures d’accompagnement financières et techniques) in the 1990s. In its special report 1/ 96432 the Court of Auditors had revealed that Ismeri, a consultancy firm, had been guilty of malversation, whereafter Ismeri challenged its naming by the Court of Auditors.433

426 Case 5/ 85, AKZO Chemie v. Commission, [1986] ECR 2585, para 36.427 Ibid, para 37.428 Opinion of AG Van Gerven in Case C- 137/ 92 P, Commission v. BASF AG e.a., [1994] ECR

I- 2555, para 41.429 Ibid, para 42. 430 Ibid, footnote 99 at p I- 2593.431 See for instance Schneider, n 287, p 38.432 Special Report 1/ 96 of the Court of Auditors, OJ 1996 C 240/ 1.433 Case T- 277/ 97, Ismeri Europa Srl v. Court of Auditors, [1999] ECR II- 1825.


In the case on appeal, AG Colomer, relying on Meroni, remarked that under the nemo plus iuris rule the Court of Auditors could name not only officials of the institutions but also those that administer EU funds on behalf of the institu-tions.434 Inghelram noted in this regard that applying Meroni leads not only to confirming the Court of Auditor’s competence but also to an obligation on the part of the Commission to act as an intermediary,435 resulting from the require-ment that the delegating authority should retain a form of supervision over the delegate authority (cf. section IV 5.3.1).

The LIFE case related to a question on the interpretation of the Comitology Decision,436 more specifically whether the legislator was bound by the criteria laid down in Article 2 when deciding on the precise comitology procedure to be prescribed. In his opinion AG Geelhoed set out the legal context and also referred to the Court’s settled case law on the Commission’s implementing powers. In passing he also remarked that ‘case law has also laid down criteria for the lawful transfer of implementing powers to agencies’,437 relying in a footnote on Meroni. The AG’s postulation was unfortunate, since this digression was unnecessary and lacked any reasoning. In addition it should be noted that it is not entirely clear which type of body the AG had in mind when he referred to ‘agencies’.

5.5.1.3 The General Court’s reliance on MeroniThe first time the General Court relied on Meroni was in DIR International Film. In the facts leading to the case the applicants had requested funding from the MEDIA programme.438 The decision on the implementation of the MEDIA pro-gramme identified the Commission as the responsible authority but also provided that the European Film Distribution Office (EFDO), a private association, should undertake action to promote the cross- frontier distribution of European films. Subsequently, the Commission had granted the EFDO implementing powers. In casu the Commission had requested the EFDO to put on hold the decision on the applicant’s request for subsidies, following which the applicant started proceed-ings before the Court.

In assessing the legal context, the Court referred to Article 7(1) of the MEDIA implementing decision which provided that ‘[t] he Commission shall be respon-sible for implementing the programme’ and read this provision together with the Meroni ruling. Noting that it follows from the latter that discretionary powers

434 Opinion of AG Ruiz- Jarabo Colomer in Case C- 315/ 99 P, Ismeri Europa Srl v. Court of Auditors, [2001] ECR I- 5281, para 78. The Court did not deal with this issue. See Case C- 315/ 99 P, Ismeri Europa Srl v. Court of Auditors, [2001] ECR I- 5281, paras 41– 2.

435 Jan Inghelram, ‘L’Arrêt Ismeri: Quelles Conséquences pour la Cour des Comptes Européenne?’, (2001) 37 CDE 5– 6, pp 721– 2.

436 See Decision (EC) 1999/ 468 of the Council, OJ 1999 L 184/ 23.437 Opinion of AG Geelhoed in Case C- 378/ 00, Commission v. Parliament and Council, [2003]

ECR I- 937, para 59.438 See Decision (EEC) 90/ 685 of the Council, OJ 1990, L 380/ 37. The MEDIA programme has

now been succeeded by the Creative Europe programme.


could not be conferred, the Court concluded that decisions by the EFDO relying on the exercise of discretionary powers could only be taken following prior agree-ment by the Commission. This had indeed been standard practice in the EFDO’s decisions on MEDIA funding.439 The Court’s reference to Meroni may have been superfluous, but it was correct nonetheless in light of the legal context in casu. After all, EFDO, just like the Brussels Agencies, was a body under private law. However, differently from the Brussels Agencies, the EFDO was only involved in implementing a pre- defined programme which had a fixed budget and com-plementary (instead of possibly conflicting) objectives. Of course the EFDO did exercise a discretionary power, because the demand for subsidies was greater than the supply, but this discretionary power was different from the discretionary power at issue in Meroni. In the words of Schneider, the EFDO only had an ad-ministrative discretion.440 The Court’s reliance on Meroni in its assessment of the EFDO’s tasks would then suggest such administrative discretion also comes under the notion of discretionary powers mentioned in Meroni, dimming the chances for a Meroni- light based on a restrictive reading of the notion of ‘discre-tionary powers’ (cf. section IV 5.4.2.2).

In Rothley e.a. a number of MEPs challenged the establishment and em-powerment by the Commission of the European Anti- Fraud Office (OLAF), since it allegedly went beyond the Commission’s competence to decide on its internal organization. In addition they argued that the Commission had violated Meroni: since the Commission did not have the competence to carry out investi-gations in relation to the Parliament and its members, it could not delegate this power to OLAF. Further in breach of Meroni, the applicants found that OLAF would exercise discretionary powers in order to organize and conduct investiga-tions.441 This case would indeed have been interesting to clarify the scope of Meroni as well as the qualification of ‘discretionary powers’ but the Court of First Instance (CFI), as well as the Court in the case on appeal, dismissed the action as inadmissible.442

In FMC Chemical the applicants challenged an opinion of the EFSA. One of the problems which the Court therefore had to solve was whether the opinion was indeed a challengeable act, producing binding effects. The European Crop Protection Association, intervening in support of the applicants, argued that the Commission had delegated the power to adopt these ‘binding opinions’ and noted that a ‘delegated entity is subject to the same conditions for the exercise of its powers as those applicable to the institutions under the control of which it acts, so far as the review of the legality of its acts is concerned’,443

439 Joined Cases T- 369/ 94 & T- 85/ 95, DIR International Film e.a. v. Commission, [1998] ECR II- 357, para 52.

440 See text at n 371.441 Case T- 17/ 00 R, Rothley e.a. v. Parliament, [2000] ECR II- 2085, paras 67– 8.442 See Case T- 17/ 00, Rothley e.a. v. Parliament, [2002] ECR II- 579; Case C- 167/ 02 P, Rothley

e.a. v. Parliament, [2004] ECR I- 3149.443 Case T- 311/ 06, FMC Chemical and Arysta Lifesciences v. EFSA, [2008] ECR II- 88, para 41.


arguing that the EFSA opinion should be challengeable. The Court countered this argument by explicitly referring to Meroni, recalling that ‘powers cannot be presumed to have been delegated and that, even when empowered to dele-gate its powers, the delegating authority must take an express decision to that effect’.444

The last and perhaps most interesting reference to the Meroni ruling can be found in Inet Hellas. Following the introduction of the ‘.eu’ internet domain, the Commission had designated a non- profit organization (EURid) as the body res-ponsible for the management of the domain.445 The applicant, Inet Hellas, had already registered the second level ‘.co’ domains of the first level ‘.gr’ domain and requested a registration of a second level ‘.co’ domain of the first level ‘.eu’ domain with EURid. The latter refused the request because ‘.co’ is also the country code of the Republic of Colombia and the applicable legislation prohibits the registration of such subdomains.446 Inet Hellas brought the matter before the Commission but the latter replied that it could not review the decision of EURid, and that it in any case agreed with EURid’s decision in substance. The applicant then initiated proceedings before the Court and argued that it follows from Meroni that the powers delegated by the Commission should remain subject to its supervision and that the Commission therefore remains responsible for the acts (or inactions) of the delegate authority.447 The Court brushed aside this argument, noting that it was the legislative Regulation448 itself which ‘gives the registry the power to refuse the registration of a second level domain and that, contrary to the applicant’s assertion, it is not a power delegated to EURid by the Commission’,449 finding Meroni irrelevant in casu.

The Court’s finding that it was not the Commission delegating powers seems logical indeed, since the legislator had laid down the framework and had tasked the Commission to entrust the actual management of the system to a third party. Although the relationship between the Commission and the Registry could still be defined as a delegation (which the Court rejected), it is equally clear that the Commission did not have a choice in making this delegation and that the original delegation should be traced back to the legislator (which the Court did). However, the Court’s finding that ‘Meroni is devoid of any relevance in the present case’ is more debatable. As was already noted, part of Meroni is generally applicable to any grant of power. Since the Court seemed to accept that there was a delegation, it should therefore also accept Meroni’s (partial) relevance. In addition, by holding Meroni completely irrelevant, it seemingly also ignored Alliance for Natural Health (cf. section IV 5.5.1.4.2).

444 Ibid, para 60.445 Decision (EC) 2003/ 375 of the Commission, OJ 2003 L 128/ 29.446 Presumably this prohibition may be found in recital 19 of the preamble to Regulation

733/ 2002. The Commission elaborated this in its implementing regulation. See Article 8(2) of Regulation (EC) 874/ 2004 of the Commission, OJ 2004 L 162/ 40.

447 Case T- 107/ 06, Inet Hellas v. Commission, [2009] ECR II- 4591, para 41.448 Regulation (EC) 733/ 2002 of the European Parliament and of the Council, OJ 2002 L 113/ 1.449 Case T- 107/ 06, n 447, para 63.


5.5.1.4 The Court of Justice’s reliance on MeroniAs was mentioned above, the Court, in 2005, referred to Meroni in two delegation cases. Inter alia this case law led Griller and Orator to conclude that

[i] n spite of Meroni’s generally ‘bad press’, the argument that the doctrine is “outdated” seems to be difficult to uphold. The jurisprudence of the EU courts shows that a dele-gation of wide discretionary powers is clearly unlawful under primary law and that the creation of regulatory agencies which typically would dispose of such powers is violating the principle of institutional balance.450

However, before drawing such conclusions, it is worthwhile to look into these cases in more depth. Obviously, the recent Short- selling case should be mentioned here as well, but given its importance it will be dealt with in a separate section (cf. IV 5.8.1).

5.5.1.4.1 TralliIn Tralli the question was whether the ECB Governing Council had lawfully delegated the power to adopt staff rules to the ECB Executive Board. The ap-plicant argued that because the Governing Council was not mandated in the ECBS Statute to delegate this power, such a delegation was ultra vires. AG Léger relied on Meroni (i) to conclude that unless a delegation is expressly prohib-ited, it should be allowed but also (ii) to refute Tralli’s plea arguing that the Executive Board’s rules went beyond a mere implementation of the Conditions of Employment.451

The issue itself was not too hard to solve, since the Staff Rules clearly remained within the framework set by the Conditions of Employment.452 Remarkably enough, the AG also dealt with this issue by reference to Meroni, combining that jurisprudence with the Court’s jurisprudence on the Commission’s implementing powers in the CAP:

the Court has ruled that the delegation of a discretionary power, implying a wide margin of discretion, replaces the choices of the delegator by those of the delegate and thus brings about an actual transfer of responsibility. In certain fields, such as the common agricul-tural policy, the Court has held, however, that the concept of ‘implementation’ must be given a wide interpretation, having regard to the scheme of the Treaty and practical considerations.453

AG Léger did not really argue why or how these two strands of jurisprudence should be connected and his observation is another good example of the diffi-culty, already noted in relation to legal doctrine,454 in distinguishing Meroni from the Köster jurisprudence. In this regard Tridimas rightly points to different legal

450 Griller and Orator, n 273, p 21. Similarly see Simone Gabbi, ‘The European Food Safety Authority: Judicial Review by Community Courts’, (2009) REDC 1, p 175.

451 Opinion of AG Léger in Case C- 301/ 02 P, n 280, para 30.452 Case C- 301/ 02 P, Tralli v. ECB, [2005] ECR I- 4071, para 5.453 Opinion of AG Léger in Case C- 301/ 02 P, n 280, para 31. 454 See n 368.


contexts between a Meroni delegation and the powers of the Commission under the old Article 202 EC.455

The Court picked up on the AG’s reliance on Meroni and remarked that ‘if the Court’s reasoning in Meroni related to the delegation of powers […] to bodies established under private law […], a Community institution or body must be entitled to lay down a body of measures of an organisational nature, delegating powers to its own internal decision- making bodies’.456 The Court thus correctly remarked that the rules on delegation within an institution should at least be as favourable as those applicable to external delegations. However, the Court, unlike the AG, did not refer to the rule on discretionary powers in Meroni and merely observed that ‘the Staff Rules remain within the limits of the executive powers conferred on the Executive Board by […] the ECB Rules of Procedure’.457

5.5.1.4.2 Alliance for Natural HealthIn Alliance for Natural Health only the Court, but not the AG, referred to Meroni. In this case, the applicant challenged the validity of Directive 2002/ 46 on food supplements.458 The directive provides that food supplements may only be mar-keted if they are included in the directive’s annexes, which were to be amended by the Commission following a comitology procedure and a scientific opinion of the EFSA. The Court confirmed the legality of resorting to positive rather than negative lists of substances, further remarking:

Finally, it should be noted that, when the Community legislature wishes to delegate its power to amend aspects of the legislative act at issue, it must ensure that that power is clearly defined and that the exercise of the power is subject to strict review in the light of objective criteria (see, to that effect, Meroni v High Authority) because otherwise it may confer on the delegate a discretion which, in the case of legislation concerning the func-tioning of the internal market in goods, would be capable of impeding, excessively and without transparency, the free movement of the goods in question.

In this instance […] the recitals to Directive 2002/ 46 […] limit the Commission’s power to modify the lists through their reference to objective criteria connected exclu-sively with public health. They show that in this instance the Community legislature laid down the essential criteria to be applied in the matter when the powers thus delegated are exercised (see, to that effect, Case 25/ 70 Köster).459

It is difficult to understand why the Court referred to Meroni here, and a confusion of the Meroni and Köster case law may again be seen here. As was noted above, an equation of both rulings cannot be supported and Alliance for Natural Health shows why. The Court first seems to argue to the effect that when the Commission uses its powers to amend legislation, it could not receive ‘discretionary’ powers, while after-wards it applies its traditional case law which makes a distinction between essential

455 Takis Tridimas, ‘Community Agencies, Competition Law, and ECSB Iniatiatives on Securities Clearing and Settlement’, (2009) 28 YEL, pp 244– 5.

456 Case C- 301/ 02 P, n 452, para 42. 457 Ibid, para 48.458 Directive (EC) 2002/ 46 of the European Parliament and of the Council, OJ 2002 L 183/ 51.459 Joined cases C- 154/ 04 & C- 155/ 04, Alliance for Natural Health, [2005] ECR I- 6451, paras 90– 2.


and non- essential elements. In Germany v. Commission the Court had earlier held that the legislator ‘may delegate to the Commission general implementing power without having to specify the essential components of the delegated power’.460 However, the observation by the Court in Alliance for Natural Health seems to amount to a prohibition for the Commission to exercise discretionary powers when it imple-ments or amends legislation. Such a prohibition would not be in line with Germany v. Commission and other well- established case law in which the Court sanctioned a wide interpretation of the notion of ‘implementation’. In Rau the Court had held that

it follows from the context of the Treaty in which article 155[EEC] must be placed and also from practical requirements that the concept of implementation must be given a wide interpretation. Since only the Commission is in a position to keep track of agricultural market trends and to act quickly when necessary, the Council may confer on it wide powers of discretion and action in that sphere.461

In Afrikanische Frucht- Compagnie the CFI relied on this case law to reject the plea by the applicants in that the Council had unlawfully delegated its powers to the Commission and that it had failed ‘to reserve to itself any power of intervention or control’.462 Evidently the CFI did not see a problem in this, but in Alliance for Natural Health the Court suddenly emphasized the need for a strict review in the light of objective criteria when the legislature delegates implementing powers.

One way of making sense of the reference to Meroni is to read it analogously to the reasoning of the Court in Tralli, where the Court found that delegations within one body should be assessed at least as favourably as delegations between different bodies. Similarly, then, the Court could have meant in Alliance for Natural Health that since the empowerment of the Commission in casu would even respect the strict Meroni doctrine, it could never be problematic, since the rules governing Commission empowerments are more favourable.

According to Van Damme and Mok, the Court referred to Meroni to justify the powers delegated to the EFSA. However, this interpretation is far from ob-vious, since the Court only referred to a delegation by the legislature and the modification of the lists by the Commission and not to the scientific opinion of the EFSA or its effect on the procedure.463 Sticking to the actual wording of the ruling, the Court therefore identified the legislature as the ‘delegating authority’ but Van Damme concludes that the Court emphasized the ‘gen-eral competence and duty of the Commission to provide for suitable, trans-parent procedures before the EFSA’.464 This is a strange interpretation of the

460 Case C- 240/ 90, Germany v. Commission, [1992] ECR I- 5383, paras 36 and 41.461 Joined Cases 279/ 84, 280/ 84, 285/ 84 & 286/ 84, Walter Rau Lebensmittelwerke e.a.

v. Commission, [1987] ECR 1069, para 14 (emphasis added).462 Joined Cases T- 64/ 01 & T- 65/ 01, Afrikanische Frucht- Compagnie GmbH v. Council and

Commission, [2004] ECR II- 521, para 113.463 Cosimo is also of the opinion that the Court held the Commission, rather than the EFSA, to

be the delegate authority. See Eloisa Denia Cosimo, (2005) RAE 3, p 508.464 Thomas Van Damme, (2006) 33 LIEI 3, p 314. The annotation by Van Damme is an

English translation of that of Mok who makes the same observation, see Robert Mok, (2006) 54 SEW 3, p 124.


ruling because nowhere did the Court actually mention such a duty for the Commission. As a result, the Court’s reference to Meroni is generally seen as a validation of Meroni, rather than as proof that Meroni governs the process of agencification (cf. section IV 5.4.1.1).

The Court’s reference to Meroni in Tralli and Alliance for Natural Health can be seen as ‘unfortunate’,465 because it led many authors to conclude that Meroni is not only valid in the EU context but also applies to EU agencies. As regards the reliance on Meroni in its legal reasoning, one can criticize the Court in Alliance for Natural Health but not in Tralli where the Court rightly applied its previous jurisprudence, despite the more questionable suggestion by the AG (cf. section IV 5.5.1.4.1).

5.5.1.4.3 Jurisprudence following Short- sellingFinally, the two cases in which Spain invoked the Meroni doctrine to challenge the unitary patent package may be mentioned here as well. Of course, the Court’s decision in these cases dates from after its ruling in Short- selling but since the Court found the Meroni doctrine not to be applicable, it did not further elab-orate or refine the doctrine substantively. Instead, the Court’s ruling should be situated in the line of cases of Inet Hellas (cf. section IV 5.5.1.3). In the cases on the unitary patent package, Spain had argued that the EU legislator had violated Meroni by entrusting certain implementing powers to the EU Member States gathered in a select committee of the Administrative Council of the European Patent Organisation (EPO). The Court however held that Meroni was irrelevant since the legislator merely invited the Member States (not the institutions) to en-trust certain tasks to the EPO pursuant to the European Patent Convention itself, which allows the parties to enter into special agreements with each other and to grant further tasks to the EPO.466

5.5.2 ConclusionDrawing a coherent conclusion from the above- cited case law is not straightfor-ward. The Court’s Opinion 1/ 76 has been dealt with in a separate section be-cause of its importance and because it is exemplary for the difficulties Meroni still faces. Although the problem at hand allowed a further refining of Meroni, the Court dealt unconvincingly with the delegation aspect of the draft Agreement. Disagreement then arose in legal doctrine as to whether the Opinion confirmed or weakened Meroni and whether these two rulings dealt with the same legal problem in the first place.467

Similar questions also arose in cases without an international dimension. In Tralli the Court rightly found that delegations within an institution should be

465 Chamon, n 332, p 297.466 See Case C- 146/ 13, Spain v. Parliament and Council, ECLI:EU:C:2015:298, paras 84– 7;

Case C- 147/ 13, Spain v. Council, ECLI:EU:C:2015:299, paras 56– 63.467 See section IV 5.5.1.1.


approached more generously than other delegations. Likewise the Court’s mixing of the Meroni and Köster strands of its jurisprudence in Alliance for Natural Health (just as the AG did in Tralli) could be criticized since delegations to the Commission under Articles 155 EEC or 202 EC are actually prescribed by the Treaty.

The Court’s reluctance to rely on Meroni meant that the judgment lay dormant for some fifty years until it was relied upon twice in 2005. Although these refer-ences may be seen as a confirmation of Meroni’s continued relevance, a closer look at these rulings shows that the Court dealt with the alleged core of the Meroni doctrine in only one of these rulings and that it further only applied Meroni on delegations within an institution and between institutions. It would therefore be premature to conclude from these rulings that Meroni indeed governs the agen-cification process.

The reliance of the General Court and the Advocates General then show that it is wrong to reduce Meroni to its alleged core. The insights gathered from the preceding sections serve as a final building block in coming to a proper under-standing of the Meroni ruling.

5.6 Assessment of Meroni and deducing a doctrine

With the insights gathered from the previous sections, it is possible to come to a final position on Meroni itself and its repercussion for the process of agencifica-tion, pre- Short- selling.

5.6.1 Identifying the Meroni doctrineMeroni undoubtedly contains an inter- related set of rules which is generally ap-plicable to problems of delegation, that is, a delegation doctrine.

The Court in Meroni formulated the following rules:

1. ‘[T] he delegation […] could not confer upon the authority receiving the delegation powers different from those which the delegating authority itself received under the Treaty.’468

2. ‘A delegation of powers cannot be presumed and even when empowered to delegate its powers the delegating authority must take an express decision transferring them.’469

3. ‘[D] elegations of powers are only legitimate if the High Authority recog-nizes them to be necessary for the performance of the tasks set out in Article 3.’470 This observation by the Court is often overlooked in legal doctrine,471 but it holds the germ of a proportionality test.

468 Case 9/ 56, n 223, p 150. 469 Ibid, p 151. 470 Ibid.471 See however Tridimas, n 292.


4. A last rule results from the combination of two of the Court’s observations and provides that the exercise of delegated powers should be subject to review in the light of objective criteria laid down by the delegating authority. This fourth rule, emphasizing the control over the delegated authority, then encompasses two sub- rules. Firstly, it allows ‘a delegation of clearly defined executive powers, the exercise of which can be subject to strict review in the light of objective criteria determined by the delegating authority, [and which therefore] cannot appreciably alter the consequences involved in the exercise of the powers concerned.’472 Secondly, it prohibits an authority ‘[t] o delegate a discretionary power, by entrusting it to bodies other than those which the Treaty has established to effect and supervise the exercise of such power, each within the limits of its own authority, [since this] would render [the] guarantee [enshrined in the balance of powers] ineffective’.473

It should be emphasized that the two sub- rules are not each other’s mirror image. The excerpts show that the Court found the delegation of executive powers ac-ceptable only because such a delegation (still) allows the delegating authority to exercise a strict review in the light of its ‘objective criteria’. The Court’s observa-tion should then be read as an obligation for the delegating authority: executive powers may be delegated only because and in so far as they are still subject to adequate review. The rationale of the fourth rule is that the Treaty established institutions which effect and supervise the exercise of powers by the Community. Institutions may still delegate powers to other bodies, but only in so far that they still supervise the exercise of the (delegated) powers. This supervision forms part of the guarantee of the balance of powers and since discretionary powers cannot be supervised, their delegation violates that balance.

Not all rules identified in the literature above have been withheld. A number of authors referred to the continued legal scrutiny by the Court or the fact that the same conditions should apply to the exercise of powers as those had the delegation not occurred. However, since competences are not only positively but also nega-tively defined, these elements may be subsumed under the nemo plus iuris rule. Second, some authors proposed that the respect for the balance of powers/ insti-tutional balance was a rule in itself.474 In reality, the balance of powers underlies the fourth requirement just as the principle of conferred powers and the propor-tionality principle (partially) underlie the first and third requirements. Each of the four rules contains a unique element, meaning none of the rules can be subsumed under another. This is also why the Court’s observations on executive and discre-tionary powers have been presented as two separate consequences of one single rule, thereby properly reflecting the Court’s underlying reasoning.

The four rules show that the Meroni doctrine applies whenever an authority delegates powers which it received under the Treaty, necessarily implying that these authorities themselves are not only foreseen by the Treaty but have also

472 Case 9/ 56, n 223, p 152. 473 Ibid. 474 See section IV 5.3.1.


had powers transferred to them by the Treaty. In Hilf ’s categorization, the delegating authorities in Meroni are those of the primary organizational layer. Secondly, the doctrine applies to delegations to ‘bodies other than those which the Treaty has established’.475 Because of the general wording, there is no imme-diate reason to assume Meroni only applies to non- EU bodies.

Of course, the real debate is how Meroni should be transposed to the EU legal framework. This will be dealt with in the following section and assumes the rul-ing’s relevance is indeed not confined to the ECSC.

5.6.2 Meroni in the EU framework5.6.2.1 The first three rulesThe major question is how the ECSC Meroni doctrine could be transposed to the EU context. Since rules one and two seem general rules applicable to any delega-tion in any Rechtsstaat, they necessarily should apply in the EU context as well.

As regards rule number three, a transposition to the EU context results in taking into account the necessity of the delegation in the light of the objectives as now set out in Article 3 TEU. The rule must then be seen as a more specific expression of the principle of proportionality enshrined in Articles 3(6) and 5(4) TEU. When an authority delegates powers, it should argue why a delegation is appropriate and necessary for the achievement of the EU’s objectives. Interpreted in this light it is a confirmation of the fact that ‘delegation’ as a legal concept is the duly justified exception. Reworked in this way, few objections exist against in-cluding the third rule of the ECSC Meroni doctrine into an EU Meroni doctrine.

5.6.2.2 The fourth rule: Control and discretionary powersThe hardest nut to crack is rule number four, the one rule which partially fig-ures in every conception of the Meroni doctrine (cf. section IV 5.3.1): the need for the delegating authority to control the delegate authority from which flows a prohibition on delegating discretionary powers. As noted, simply applying this rule to agencification without any adaptation would be highly problematic, which explains why so many authors argue for a Meroni- light or even for the irrelevance of Meroni. The insights gathered from previous sections will not be repeated here, but they do form an important background to the argument.

5.6.2.2.1 Why a simple transposition to the EU context is unwarrantedThe change in legal context from a traité loi to a traité cadre is decisive in arguing for a fresh look at the fourth rule. If one recalls that the ECSC Treaty had a lim-ited purpose (and objectives), it is pertinent to argue that there is no or little need to establish bodies other than those for which the traité loi provided. As regards one of the drivers behind agencification (cf. section III 2.1), one could further not argue that the High Authority lacked the necessary expertise, and even if there

475 Case 9/ 56, n 223, p 152.


was a need for a certain independence from the core administration (something which did not play in Meroni) one could follow the example of Eurostat. Indeed the third rule may be understood as the Court hinting at this (cf. section IV 5.6.1).

All this changes however when one moves to the legal context of a traité cadre. Taking the limited institutional setup of the EU as a given, one can no longer argue that there is no or little need to delegate certain tasks. A risk may therefore be identified of an institutional deficit arising because of the widening gap be-tween the institutional capacities of the EU and the tasks which it is supposed to take up.476 In addition, the effet utile of the competence to delegate would require that a delegation should not be limited to ‘clearly defined executive powers’. As will be dealt with later (cf. section IV 5.7), part of the ratio behind a delegation is that it increases efficiency, but if discretionary powers may not be delegated this greatly reduces the efficiency gains which may be made through delegation.477 The fourth rule may then by relaxed by redefining what discretionary powers are and by cutting the link between purely executive powers and control, since con-trol is also possible when discretionary powers have been exercised.478

5.6.2.2.2 (Re- )Defining discretionary powersIn the sections on the nature of discretionary powers and the subsequent case law of the Court it was noted that the notion of discretionary powers has been equated, or at least confounded, in jurisprudence and legal doctrine with the notion of ‘essential elements’ laid down by the Court for the first time in Köster. It seems clear that the power to adopt ‘essential elements’ of legislation cannot be delegated. However, this minimum content of the category of ‘discretionary powers’ should not at the same time be its maximum content unless the delegate authority’s legitimacy under the Treaties matches that of the Commission.

Because the fourth rule depends on it, the notion of control should also guide how ‘discretionary’ is understood. The more distant (and thus harder to control) the delegate authority from the delegating authority, the more strictly this rule should be applied, as was also suggested by the Danish government in Opinion 1/ 76. The rule should therefore be relaxed when moving from bodies such as the laying- up fund,479 over the Office and Caisse,480 to bodies under secondary law,481 and lastly bodies under primary law.482 In addition, if EU agencies were to be empowered through genuine delegations (by the Commission) rather than through conferrals (by the legislator) (cf. section IV 5.7) the control over EU

476 Everson and Majone, n 273, pp 144– 50.477 See also the argument by Everling, section IV 5.2.478 Actually, this was also hinted at by the Court in Meroni itself: see the last paragraph of sec-

tion IV 5.1.4.2.3.479 Ie an external body in which a third country is involved, thereby preventing the EU from

unilaterally imposing its will.480 Ie an external body, on which the EU can still impose its will.481 Ie an EU body unforeseen in the Treaties.482 Ie an EU body foreseen in the Treaties.


agencies could be more readily assumed, translating into the powers delegatable to agencies.

For the purposes of identifying these powers, one could draw inspiration from the scheme set out by AG Van Gerven in the PVC case.483 In addition to the first level decisions of principle, the second level decisions of management, and the third level decisions of practical execution, one could include the decisions on the essential elements (reserved to the legislator at level 0) and (at level 1.5) the deci-sions implementing or concretizing the (legislator’s and Commission’s) decisions of principle in a single system. The higher- level decisions would have to be taken by bodies of the primary organizational layer,484 but lower- level decisions would be open to the agencies. In all senses, these decisions would still require ‘discre-tion’ to be exercised, but this is not problematic if in turn sufficient control is put in place. On this point the Court’s original reasoning, that is, that exercising control implies that only purely executive powers have been granted, should be rejected.

5.6.2.2.3 The underlying notion of controlBecause of its two sub- rules, the mere conclusion that no discretionary powers have been delegated does not completely satisfy the fourth rule, since the powers delegated should also be sufficiently controlled. In its ruling the Court derived this control from the balance of powers which was a guarantee to private parties. This reference has been taken up by some authors to link the process of agencifica-tion with the contemporary notion of institutional balance (cf. section IV 5.3.1). However, that notion cannot be derived automatically from Meroni.485 The main objection thereto is that the institutional balance was introduced, elaborated, and refined by the Court in its case law since 1958. The principle as it is construed today is not the same as when it was construed in the time of the Meroni ruling. However, most authors apply the modern interpretation of the principle in their analyses of the Meroni ruling. In reality, the Meroni judgment only forms part of the different constituent elements of this principle as it evolved later in the case law of the Court (cf. section IV 7.2.3). It is not self- evident, therefore, to apply the principle in its evolved interpretation to make sense of the Meroni judgment.

The main objection to applying the modern interpretation of the principle of institutional balance is that in its evolution from ‘balance of powers’ in Meroni to the modern- day ‘institutional balance’, a qualitative leap has occurred. As Jacqué points out, it was originally conceived as a substitute for the principle of the separation of powers of Montesquieu, the aim of which was to protect indi-viduals against the abuse of power.486 Jacqué observes that this protective aspect

483 See n 428.484 For example, in Commission v. Germany, where the Court noted that the decisions to address

a reasoned opinion to a Member State or to initiate proceedings under Article 258 TFEU cannot be delegated by the Commission to one of its members but must be adopted by the college. See Case C- 191/ 95, Commission v. Germany, [1998] ECR I- 5449, para 36.

485 Chamon, n 332, pp 295– 7.486 Jean- Paul Jacqué, ‘The Principle of Institutional Balance’, (2004) 41 CMLRev 2, p 384.


of the principle seems gradually to have been lost as other means of protection appeared,487 meaning that the interinstitutional relations, which are now the core of the institutional balance, were not on the mind of the Court in Meroni.

The importance of the notion of control in the Meroni ruling is further un-derlined by the opinion of AG Roemer. The AG was foremost preoccupied with the continued respect for the system of judicial protection. The possible solu-tion offered by the AG, that is, equating the acts of the Brussels Agencies with those of the High Authority, is a further indication of this. One can see in the fourth rule identified previously that the same preoccupation was also important in the Court’s ruling. However, in its ruling the Court did not choose the sol-ution offered by the AG. Instead the Court prescribed that these bodies may only exercise strictly executive powers, perhaps assuming that this would preclude problems under the Treaty’s system of legal protection from arising. This is not necessarily the case, however, as the Court also implicitly recognized a year later in the SNUPAT case. In SNUPAT the Court equated a notification by the Caisse with a decision of the High Authority.488 If the Court had ruled differently, such a notification would not have been challengeable, thus depriving the undertaking concerned of the protection offered by the Treaty.

Despite the different solutions worked out by the AG and the Court in Meroni, it appears that a key concern for both was that rights of private parties should not be affected by a delegation of powers. This is of course something very dif-ferent from how the institutional balance is conceptualized today (cf. section IV 7) and should make clear that the relevance of Meroni for any delegation issue is its insistence on both continued political control and judicial scrutiny. Ironically the latter has been one of the problems in the current process of agencification, because of the ad hoc way in which the agencies have been established and the powers increasingly conferred on them.489 This problem will form the subject of a separate chapter. The question of political control has been equally problematic, since neither the EU legislator nor the Commission can assume political respon-sibility for agency (in)action.

5.7 Delegating powers in the EU legal order

Before proceeding further, it is useful to try and clarify the central notion of dele-gation. An attempt at clarification seems in place since the notion was absent from

487 However, relatively recently, see Joined Opinion of AG Fennelly in Case C- 286/ 95 P and Joined Cases C- 287/ 95 P & C- 288/ 95 P, Commission v. ICI and Commission v. Solvay SA, [2000] ECR I- 2341, para 23.

488 Joined Cases 32/ 58 & 33/ 58, SNUPAT v. High Authority, [1959] ECR 127. Schindler criti-cized the Court on this point since in Meroni it laid down the nemo plus iuris rule, implying the High Authority had transferred competences, while in SNUPAT it held the High Authority ac-countable for the Brussels’ agencies’ acts. See Schindler, n 18, p 50. This issue flows from a more general disagreement, viz whether a delegation transfers responsibility or not: see section IV 5.7.1.

489 See for instance Johannes Saurer, ‘Individualrechtsschutz gegen das Handeln der Europäischen Agenturen’, (2010) 44 Europarecht 1, pp 51– 66.


the Treaties (before the Lisbon Treaty),490 and given the debate in legal doctrine as to whether certain transferrals of power could be qualified as delegations.491

5.7.1 Different concepts of delegationIn his study on delegation in Community law, Gautier remarks that the notion of delegation is in common use but that ‘nul n’est véritablement en mesure d’en proposer une définition univoque’,492 adding that even more confusion exists as regards the object of the delegation, that is, whether competences, powers, sig-natures, authority, puissance, sovereignty, functions, etc are being delegated.493 Focusing on the delegation of powers, he proposes the following definition: ‘La délégation de pouvoir est l’acte par lequel le délégant, titulaire de la compétence, se dessaisit d’une parcelle de ses pouvoirs au profit du délégataire.’494 Delegation then connects two authorities, but Barents opens the possibility of a triad: a dele-gation of powers means ‘that on the basis of a valid title, authority A obtains powers from authority B, either by virtue of a decision of authority B or by a dec-ision of a third, higher authority C’.495

According to Gautier a delegation does not result in an aliénation défini-tive of the delegating authority from the object which he delegates; instead he retains the essence of the object and only transfers the exercise thereof.496 To identify the possible object of a delegation, Gautier builds on Constantinesco’s presentation of the relation between competences and powers. According to Constantinesco, both are each other’s condition d’existence in the sense that ‘power’ is the materialization of ‘competence’. As a result, ‘competence’ without ‘power’ is ineffective and ‘power’ without ‘competence’ is illegal.497 Applied to the competences of the delegating authority, Gautier concludes that a dele-gating authority cannot delegate its competences, but only the power to exercise a competence.498

According to Gautier, if the representation underpinning the delegation is perfect, the delegate authority acts in the name and on behalf of the delegating authority.499 This should be juxtaposed with Triepel’s account of delegation in

490 To be precise, the concept of delegation was already used in primary law but only in the statutes of the ESCB and of the ECB and those of the EIB (relating to the internal functioning of these institutions), currently Protocols No 4 and 5.

491 See section IV 5.5.1.1.492 Yves Gautier, La délégation en droit communautaire, Strasbourg, Strasbourg 3, 1995, PhD

Thesis, p 4.493 Ibid, p 5. 494 Ibid, p 49.495 René Barents, The Autonomy of Community Law, The Hague, Kluwer Law International,

2004, Vol 45, p 224.496 Gautier, n 492, p 42. See also André Mast, ‘Delegatie en toewijzing van bevoegdheid’, in

Rechtsfaculteit (ed), Liber amicorum baron Louis Fredericq, Gent, Story- Scientia, 1966, Vol 2, p 738.497 Vlad Constantinesco, Compétences et pouvoirs dans les Communautés européennes: contri-

bution à l’ étude de la nature juridique des communautés, Paris, Librairie Générale de Droit et de Jurisprudence, 1974, pp 82– 3.

498 Gautier, n 492, p 47.499 Ibid, p 37. See also Lauwaars, n 261, pp 147– 8.


which the delegating authority acts in its own name and on its own account-ability.500 Both authors do however agree that the delegation results, on the part of the delegating authority, in the loss of (power to exercise) the delegated com-petence and on the unilateral character of a delegation.501

However, Berger notes that this classic view was dismissed by Barbey, who bel-ieves a competence or power cannot be dissociated from the authority in which it is vested. A delegation, according to Barbey, results in the creation of competence in the delegate authority.502 Three norms are then important for establishing an außerordentlichen Zuständigkeit (an extraordinary competence, ie a competence following delegation):503 a Regelnorm, which defines the originally competent au-thority; a Delegationsnorm, which is (at least) of the same rank as the Regelnorm and which sets out under which circumstances a lower rule may deviate from the rules laid down in the Regelnorm; and a delegierende Norm, which is of a lower rank than the Regelnorm and which, in accordance with the Delegationsnorm, organizes a specific delegation.504

Berger rightly notes that Barbey’s description of a delegation as a creation of competences is not simply a theoretical nicety because it means that a delegate authority can receive competences which the delegating authority did not possess, something which is not possible if the nemo plus iuris rule applies.505 Barbey’s delegation concept revolves around the außerordentlichen Zuständigkeit, which deviates from the normal structuring of competences. Such a deviation may be organized as long as the delegating act (the delegierende Norm) is adopted in ac-cordance with the Delegationsnorm.

An appraisal of Barbey’s delegation concept does leave one wondering whether the notion of ‘delegation’ may still be used. Evidently this will be the case when a Delegationsnorm provides that authority A may confer a specific competence to au-thority B, as was the case for the provisions in the German Grundgesetz on which Barbey based his study.506 However, this would be different if the Delegationsnorm is implicit or worded imprecisely, whereby conferrable competences are only ref-erred to generally and/ or whereby no specific delegate authority is designated. In

500 Berger, n 13, p 72. Gautier’s delegation whereby the delegate authority acts in the name and on behalf of the delegating authority is qualified as Mandat by Triepel: see Gunther Barbey, Rechtsübertragung und Delegation: eine Auseinandersetzung mit der Delegationslehre Heinrich TRIEPELs, Münster, Rechts- und Staatswissenchlaftliche Fakultät der Westfälischen Wilhelms- Universität, 1962, pp 53– 4.

501 Berger, n 13, p 72; Gautier, n 492, pp 39 and 47.502 This also means that under Barbey’s delegation concept, the delegating authority can per-

fectly remain competent following a delegation. See Barbey, n 500, pp 84– 5.503 As Barbey notes, the delegation is an exception to the ordinary allocation of powers, and

therefore extraordinary. See ibid, p 68. See also in this regard Gautier who notes that delegations are problematic precisely because they are an exception to the normal règles de compétence, Gautier, n 492, p 36.

504 Barbey, n 500, pp 76– 7.505 Berger, n 13, pp 72– 3. See also Barbey, n 500, p 91.506 These provisions were Article 66(3) GG iuncto Article 66(2), and (the original, ie before the

1962 amendment) Article 108(3) GG. See Grundgesetz für die Bundesrepublik Deutschland vom 23. Mai 1949, Bundesgesetzblatt 1949 1, pp 14– 15.


those cases one could also argue that the außerordentlichen Zuständigkeit is not created following a delegation, but that it is simply created under an (implicit) enabling clause.

5.7.2 In search of an EU delegation conceptIn their studies, Gautier, Berger, and Schindler507 conclude that applying their general insights on delegation to the EU legal order is not straightforward. Berger notes that since the Court in Meroni relied on the nemo plus iuris rule (cf. sec-tion IV 5.1.4.2.2), it seems to have followed the classic approach on delegation. However, according to Berger Meroni is still compatible with Barbey’s view on delegation,508 since in Meroni the High Authority was both the originally com-petent authority as well as the delegating authority, necessarily resulting in a legal situation compatible with the nemo plus iuris rule.

Similarly, Gautier observes that the Court has deviated from a general rule on delegation to the effect that there is no need for an explicit competence to delegate, only that a reasonable interpretation of the legal basis is not opposed to delegation.509 Furthermore the pre- Lisbon delegations by the Council to the Commission under the Articles 202 and 211 EC were atypical in the sense that (i) there is no hierarchical relation between the Council and the Commission, (ii) the Commission did not act on behalf of the Council,510 (iii) the Commission would have to agree to a withdrawal of the delegation (by proposing an amend-ment to the delegating act),511 and (iv) both Council and Commission remained competent to adopt the same acts.512 Harnier also concluded that the concept of delegation, as it emerges from the national legal orders, could not be applied to Article 155 EEC.513 Instead he describes the empowerment of the Commission as rechtssatzmäßige Zuweisung von Befugnisse (ie an allocation of competences in accordance with a rule of law).514 Referring to Barbey he observes that compe-tences cannot be dissociated from a competent authority and proposes that the Commission’s executive competences are originally created through but not dele-gated by the Council’s act.515

507 Schindler, n 18, pp 37– 60. 508 See ibid, pp 64– 6.509 Gautier, n 492, p 536. 510 Ibid, p 518.511 Ibid, p 466. However, Schindler rejected the idea that the Council would depend on the

Commission to withdraw a delegation under Article 155 EEC. See Schindler, n 18, p 178.512 Gautier, n 492, p 468. Triepel would characterize this as a konservierenden Delegation,

which he sees as an unechten Delegation (a false delegation). See Barbey, n 500, pp 59– 60.513 In addition to some of the elements noted by Gautier, Harnier noted that the Council cannot

reserve to itself the power to lay down all non- political elements of legislation since this falls under the ‘vom Vertrag festgestelte Domäne’ of the Commission and therefore it cannot be delegated to the Commission. See Otto Harnier, Kompetenzverteilung und Kompetenzübertragung zwischen Rat und Kommission unter Berücksichtigung der Einsetzung von Hilfsorganen im Recht der EWG, München, Frank, 1969, pp 67 and 77– 8. Here Harnier seems too strict in his position, especially under the wording of Article 155 EEC, but even today it is difficult to maintain that the legislator may only deal with the political (or essential) questions. Following Harnier’s study, the Court in Köster actually reversed this reasoning.

514 Ibid, pp 81– 2. 515 Ibid, p 83.


The agencification of the EU administration has further made the conceptual confusion complete. Berger rightly notes that the legislator has ‘delegated’ powers to EU agencies which it does not possess itself but which fall to the Commission ‘under the normal division of competences’,516 leaving Berger to conclude that legislative praxis has adopted Barbey’s notion of delegation.517 Similarly, Gautier notes that the agencies have not been delegated powers by the legislator,518 since ‘[p] our que la délégation soit une explication satisfaisante, il faudrait que le Conseil [leur] est confié des tâches que relèvent directement de lui’.519 As a result, and con-trary to an overwhelming consensus in legal doctrine, Gautier simply rejects the concept of delegation to describe the empowerment of EU agencies.520

The Treaty of Lisbon then added a new dimension to this discussion, since it introduced for the first time the notion of delegation in primary law in Article 290 TFEU. Barents noted that Article 290 TFEU is absurdly worded because it provides that the object of the delegation in Article 290 TFEU is the power to adopt delegated acts. According to Barents, this power cannot be delegated be-cause the legislator is not invested with it in the first place; instead, this power may only be attributed to the Commission.521 Still, the legislator now has the power to withdraw a delegation, something which was previously not possible, as noted by Gautier.522 Without further commenting on Article 290 TFEU, it may be noted for the purposes of the present study that the framework of Article 290 TFEU is strong and elaborate compared to those of other ‘delegation regimes’.

The fact that primary law only refers to delegation since the Lisbon Treaty makes it difficult to identify a delegation notion in EU law. Still the Court has used the notion of delegation to describe a variety of legal situations. In the English- language version of Köster,523 the Court held that the Council could delegate implementing powers to the Commission under Article 155 EEC, extending this to Article 145 EEC in Germany v. Commission.524 The Court also referred to the notion to ap-preciate ‘internal’ delegations within the Commission.525 In other cases the Court even referred to the powers of the Member States to implement EU law as ‘powers delegated to the Member States’.526 Although Lauwaars refers to Opinions 1/ 76 and

516 Berger, n 13, p 75. 517 Ibid, p 76.518 Gautier refers specifically to the EMCDDA but his conclusion may be generalized to all

agencies.519 Gautier, n 492, p 402.520 Under the same reasoning, Gautier would not contest that the executive agencies are empow-

ered through delegations.521 Barents, n 138, p 460. This is similar to Harnier’s remarks on Article 155 EEC.522 See n 511.523 Case 25/ 70, n 314, para 9. The other language versions do not refer to a delegation.524 Case C- 240/ 90, n 460.525 Case 48/ 69, Imperial Chemical Industries Ltd. v. Commission, [1972] ECR 619, para 14.526 Case 131/ 73, Criminal proceedings against Giulio et Adriano Grosoli, [1973] ECR 1555, para

5; Joined Cases 213 to 215/ 81, Norddeutsches Vieh- und Fleischkontor Herbert Will, [1982] ECR 3583, para 10. These cases are related to the CCP, which helps explain why the Court defined this power as delegated from the EU to the Member States. Interestingly, the Commission in ACOR also proposed to define powers of the Member States in the CAP, as powers delegated to them: see Case C- 416/ 01, ACOR, [2003] ECR I- 14083, para 36. Similarly as regards the administration of


1/ 91 when discussing delegations to international bodies,527 it should be noted that the Court itself did not use the notion of delegations in those opinions. While the Court did (extensively) rely on the notion of delegation in its Meroni ruling, it ref-erred to ‘empowerment’ in Romano.528 Gautier noted that although the Court has categorized these different situations under the single denominator of ‘delegation’, there is no single legal regime which applies to all of them.529

As noted, the Court also relied on Meroni to appreciate an internal delega-tion in Tralli. And here again the problematic character of Alliance for Natural Health should be noted, since the Court also seemingly made a link between Meroni and the delegations under Articles 202 and 211 EC. A combined reading of these judgments could be interpreted as giving rise to a single notion of delega-tion in EU law, but envisaging the effects this would have on the Commission’s powers under Articles 290 and 291 TFEU, this should be adamantly rejected. Fortunately, the Court has not further referred to Meroni in its case law on the implementing powers of the Commission.530

5.7.2.1 Multiple delegation regimes in the EU legal orderIt follows that a clear delegation concept is lacking in both primary law and juris-prudence. As noted previously, Berger did induce from legislative practice that the (ordinary) legislator has adopted Barbey’s delegation notion, but it should be clear that this question is not in the hands of the ordinary legislator. Applying Barbey’s reasoning to the EU legal order makes this clear. The Regelnorm, in Barbey’s study the Grundgesetz, is EU primary law. The delegierende Norm cor-responds to the secondary legislation from which Berger concludes that the EU legislator has adopted Barbey’s model. In that model, these two Norms should be linked by a Delegationsnorm, but in EU law, few such Delegationsnormen may be found. Specifically in relation to the process of agencification, the problem is that the Treaties are silent and therefore an express Delegationsnorm is lacking. This means an implicit Delegationsnorm would have to be found, but Barbey himself dismissed this possibility as irrelevant. According to him, although an implicit Delegationsnorm would in theory be possible if the delegated competences are also implicit competences, the fact that these competences would be implicit would mean a formal delegation is also unnecessary.531

Community aid in the CAP, AG Stix- Hackl in Industrias de Deshidratación Agrícola referred to the limited powers which had been delegated to the Member States: see Opinion of AG Stix- Hackl in Case C- 118/ 02, Industrias de Deshidratación Agrícola SA, [2004] ECR I- 3073, para 19.

527 Richard Lauwaars, ‘Hoofdstuk IV: De Institutionele Structuur’, in Kapteyn and VerLoren Van Themaat (eds), Het recht van de Europese Unie en van de Europese Gemeenschappen, Deventer, Kluwer, 2003, p 193.

528 Case 98/ 80, Romano, [1981] ECR 1241, para 20.529 Gautier, n 492, p 461. According to Gautier, the same rules applied in Opinion 1/ 76 as in

Meroni: see ibid, p 459. See also Gautier, n 425, pp 13– 14.530 See recently however, Opinion of AG Cruz Villalón in Case C- 427/ 12, ECLI:EU:C:2013:871,

para 76.531 Barbey, n 500, p 103.


In this regard it is interesting to note that Triepel would allow a delegation in the absence of clear provisions allowing or prohibiting a delegation. To Triepel, ‘the development of customary law under the pressures of practical necessity’532 would make a delegation permissible as long as the constitutional norm does not attribute special importance to the delegated power being exercised by the auth-ority in which it is originally vested.533 Of course, in the EU legal order one could rather forcefully argue that the EU Treaties indeed attach special importance to the position of the Commission as the authority responsible for ‘the application of the Treaties, and of measures adopted by the institutions pursuant to them’ and for exercising ‘coordinating, executive and management functions’.534 Does agencification then not undermine this position? As will be discussed later (see section IV 7.2.4.1.3), even if one could think so, the Court did not take issue with this in Short- selling.

As was briefly referred to previously, the problem in the EU legal order is that there are not that many Delegationsnormen to be found in primary law. Of course Article 53 ECSC, which was the legal basis of the equalization scheme, contained a Delegationsnorm,535 and Article 290 TFEU is another good example since it spe-cifies a number of conditions which the delegierende Norm should meet. This is different for the Commission’s power under Article 291 TFEU, which is more in line with the old Article 145 EEC. The Commission’s power in that Article is ori-ginal and not delegated because it is created by virtue of the finding that ‘uniform conditions for implementing legally binding Union acts are needed’. Similarly Article 282 EC may also be noted, which inter alia provided that ‘the Community shall be represented by the Commission’. In European Communities v. Région Bruxelles- Capitale the question was put to the Court how the Commission could then delegate its power of representation to other institutions. Although Article 282 EC was silent on this, the Court remarked that ‘[i] t was possible for the Commission to delegate that power through a mandate granted to the other in-stitutions “in matters relating to their respective operation”’.536 The Treaty of Lisbon has amended Article 282 EC, but it did not do so by providing for a Delegationsnorm. Instead, Article 335 TFEU now inter alia reads: ‘the Union shall be represented by the Commission. However, the Union shall be represented by each of the institutions, by virtue of their administrative autonomy, in matters relating to their respective operation’. This power cannot therefore be delegated by the Commission to the other institutions since it is an original power vested in those institutions.537

532 As cited in Barbey, p 97. Translated from German by the author.533 Barbey, n 500, p 97. 534 See Article 17 TEU.535 Schindler, n 18, p 72. See also n 261.536 Case C- 137/ 10, Communautés européennes v. Région de Bruxelles- Capitale, [2011] ECR I- 3515,

para 19.537 This in turn raises the question of whether the Commission requires a delegation before it can

represent the EU in matters relating to the operation of other institutions. In Europese Gemeenschap v. Otis NV e.a. this question was put to the Court but remained unanswered, since the Court ruled that the action in the main proceedings was governed by Article 282 EC and not Article 335 TFEU. See Case C- 199/ 11, Europese Gemeenschap v. Otis NV e.a., ECLI:EU:C:2012:684, para 35.


One could conclude that the traditional notion of delegation is apparent in the EU legal order in the Meroni- type cases.538 As regards the Commission’s powers under Article 291 TFEU the notion of delegation is ill suited, since the Commission is vested with an original power therein.539 This is also because of the evolution of the Commission’s implementing powers following subsequent Treaty revisions. Whereas the Court in Köster defined the use of Article 155 as op-tional,540 the Commission’s position was strengthened by the SEA and the Treaty of Lisbon. Following the SEA the question arose as to whether the implemen-tation of EEC law fell within the Commission’s own competence, and whether there was an obligation on the part of the Council to confer powers on the Commission. Although Gautier rejected that the Commission had a competence propre he did agree that the Council was under an obligation to confer powers on the Commission.541 In this regard, Blumann referred to the Commission’s nat-ural vocation to exercise the executive function which flowed from Article 155 EEC.542

Before the Lisbon Treaty, the Council (or legislator) had the power to decide whether implementing powers should be exercised at EU level. Under Article 291 TFEU, however, this question is taken out of the hands of the legislator and ob-jectivized.543 Although Craig rightly notes that the Council’s exclusion from the control on the Commission, under Article 291(3) TFEU, is foremost theoretical and that it is doubtful whether the Council will exercise this restraint in prac-tice,544 it should still be noted that the theoretical shift between Articles 202 EC and 291 TFEU is fundamental. Bianchi notes that pre- Lisbon it was more or less assumed that comitology was a form of control by the Council as the delegating authority and he concludes that comitology should thus undergo fun-damental changes post- Lisbon.545 Even if the Lisbon Treaty anchored the idea

538 See for instance the DIR International Film case, n 439.539 Amphoux also concluded this in relation to Article 155 EEC, observing that the Commission’s

powers were ‘des compétences nouvelles, nées en même temps que le droit dérivé à exécuter, […] et dont la Commission est titulaire en propre’. See Jean Amphoux, ‘La Commission’, in Megret, Waelbroeck and Louis (eds), Le droit de la Communauté économique européenne, Bruxelles, ULB, 1979, Vol 9, p 177.

540 Case 25/ 70, n 314, para 9.541 Gautier, n 492, p 384. Jacqué as well noted that ‘[l] a règle générale contient bien l’obligation

pour le Conseil de conférer ces compétences.’ See Jean- Paul Jacqué, ‘L’Acte unique européen’, (1986) 22 RTDE 4, p 595.

542 Claude Blumann, ‘La Comitologie: l’exercice de la fonction exécutive dans la Communauté européenne’, in Engel and Wessels (eds), From Luxembourg to Maastricht: institutional change in the European Community after the Single European Act, Bonn, Europa Union, 1992, p 96. Similarly, see Ellen Vos, ‘The Rise of Committees’, (1997) 3 ELJ 3, p 214; Paul Craig, EU Administrative Law, Oxford, OUP, 2012, p 115.

543 On this, Blumann notes: ‘[L’article 202 CE] semblait conférer au Conseil une sorte de pou-voir discrétionnaire lui permettant de choisir entre l’exécution par les États membres ou par les in-stitutions de l’Union […] Cette faculté ne paraît plus pouvoir être ouverte aujourd’hui’. See Claude Blumann, ‘Un nouveau départ pour la Comitologie. Le règlement no 182/ 2011 du 16 février 2011’, (2011) 47 CDE 1, p 25. See also n 392.

544 Craig, n 542, pp 132– 3.545 Daniele Bianchi, ‘La comitologie est morte! vive la comitologie!’, (2012) 48 RTDE 1, p 83.


of Vollzugsföderalismus in primary law,546 at EU level the Commission’s power was (formally) strengthened as the EU authority in which this power is princi-pally vested.547 The notion of delegation is therefore unsuited to describe the legal regime under Article 291 TFEU.

This is different for the regime under Article 290 TFEU, which uses a dele-gation terminology and which may be seen as a Delegationsnorm under Barbey’s notion of delegation. For reasons of efficiency, one of the principles underpinning the ratio of a delegation,548 the legislator may delegate part of his legislative com-petences as long as he himself has laid down the essential elements of the legisla-tion in question.

The empowerment of bodies under international public law is again a different matter. Previously it was already noted that Gautier, who rightly emphasizes the notion of control between delegating and delegate authority in a delegation, char-acterizes such an empowerment as being a delegation because the EU institutions participate in the governing body of the international organization.549 However, this may risk diluting the concept of control.

In this regard Sarooshi differentiates between three models of empower-ing international organizations, located on a single spectrum. The three models depend on three characteristics of conferrals: the revocability of powers, the con-trol over the international organization, and the exclusive or concurrent nature of the powers conferred.550 These result in the models of (i) agency relationship, (ii) delegation of powers, and (iii) transfer of powers.551 When the powers conferred are irrevocable and exclusively exercised by the international organization and the control a state can exercise over an international organization has become minimal, as is the case for the World Trade Organization (WTO) and the EU according to Saaroshi,552 one cannot speak in terms of delegation but rather in terms of a transfer of powers.553 Because Saaroshi does not precisely define the characteristics of a delegation, it is difficult to compare it to the delegation notions set out previously. Still Sarooshi’s study is insightful, since it shows that global or regional integration can achieve such levels that it no longer makes sense to speak in terms of delegation. The question of control, which the Court emphasized in Meroni, is especially highly problematic in multilateral forums such as the WTO. Where the laying- up fund at issue in Opinion 1/ 76 could have been controlled by

546 Chamon, n 392, p 67.547 Not going as far, Blumann notes that the Commission is the default authority and that the

legislator must properly motivate its decison to exceptionally empower the Council. See Blumann, n 543, p 26.

548 Gautier, n 492, p 32. 549 See n 425.550 Dan Sarooshi, International Organizations and Their Exercise of Sovereign Powers, Oxford,

OUP, 2005, p 29.551 Ibid.552 Gautier concluded that delegation could not aptly describe the vertical relationship and con-

ferral of powers from the Member States to the EU, although delegation could describe the con-ferral of powers from the Member States to the ECSC. See Gautier, n 492, p 85.

553 Sarooshi, n 550, pp 29– 32.


the EEC because there was only one third state as contracting party, this is not the case for the WTO. Describing the EC’s ratification of the WTO agreements as a delegation would then result in the conclusion that it is an illegitimate delega-tion, although membership of the WTO is an (indispensable) instrument in the EU’s Common Commercial Policy (CCP). In addition the Court has accepted the participation of the EU in international forums, equally accepting that the EU ‘may accept rules made by virtue of an agreement’ and drawing the red line that this may not lead to ‘chang[ing] the nature of the powers of the Community and of its institutions as conceived in the Treaty’,554 further stressing the EU’s continued autonomy.555 For these reasons, conferrals of powers to international organizations should not be qualified as delegations, even if in specific cases, such as the one in Opinion 1/ 76, such conferrals could possibly meet the requirements of a delegation.

5.7.2.2 May the empowerment of an EU agency be described as a delegation?Next to the regimes described in the previous section, the empowerment of EU agencies could be a final regime. Although the executive agencies fall outside the scope of this study, it may be noted that the framework regulation for these agen-cies556 could be conceptualized as a Delegationsnorm allowing the Commission to adopt delegierende Normen. Of course the problem here is that the Regulation is an instrument of secondary law which allows for an exception to a rule of primary law, whereas under Barbey’s model the Delegationsnorm should at least be of the same rank as the Regelnorm. Still it is interesting to note that this regulation is not an ordinary one but should instead be positioned alongside other regulations such as the ones on comitology,557 access to documents,558 and the protection of personal data.559 These organic instruments take up a special position within the body of secondary law because they govern the EU institutions’ and bodies’ activ-ities in a horizontal manner.560

As regards the EU agencies proper, the problems were already highlighted previ-ously. Triepel and Gautier’s traditional notions of delegation do not seem to apply,

554 Opinion 1/ 92, re the draft agreement relating to the creation of the European Economic Area, [1992] ECR I- 2821, para 41.

555 Opinion 1/ 00, re the proposed agreement on the establishment of a European Common Aviation Area, [2002] ECR I- 3493, paras 11– 14. See Opinion 1/ 92, n 554, para 41; Opinion 2/ 13, re Accession to the ECHR, ECLI:EU:C:2014:2454, para 183. See also Adam, n 404, pp 73– 83.

556 See Chapter I, n 48.557 Regulation (EU) 182/ 2011 of the European Parliament and of the Council, OJ 2011 L 55/ 13.558 Regulation (EC) 1049/ 2001 of the European Parliament and of the Council, OJ 2001 L 145/ 43.559 Regulation (EC) 45/ 2001 of the European Parliament and of the Council, OJ 2001 L 8/ 1.560 The then Director- General of the Commission’s Legal Service, Mr. Petite, proposed during

the Convention to enshrine these organic or constitutional laws in a hierarchy of norms, below primary law but above regular legislation. See Secretariat of the European Convention, Note sum-marising the meeting of Working Group IX (Simplification) on 17 October 2002, CONV 363/ 02, p 5. This special status was also recognized by the Court in LIFE, as noted by Schütze. See Case C- 378/ 00, Commission v. Parliament and Council, [2003] ECR I- 937, paras 39– 40; Robert Schütze, “Delegated” Legislation in the (New) European Union: A Constitutional Analysis’, (2011) 74 The Modern Law Review 5, p 678 at footnote 103.


since the legislator cannot delegate powers which are not its own,561 while for Barbey’s model a Delegationsnorm is lacking. In addition, recourse to Barents’ triad is difficult because it presupposes a higher authority C which delegates powers from body A to B,562 which would amount to accepting a superiority of the Council and the Parliament over the Commission. At the same time the reason for not des cribing the conferral of powers to international organizations as delegation does not apply to agencies, because there can be no doubt that the EU institutions can fully control EU agencies.

If a more flexible position is taken on the lack of hierarchy (ie the lack of a hierarchical higher norm or the lack of a hierarchy between the EU insti-tutions), one could subsume agencification under these notions of delegations by assuming there is an implicit Delegationsnorm available to the legislator or by taking the position that the legislator (and not one of its components) is the power superior to the executive. Of course, the latter would be based on a separation- of- powers argument and on the principle of democracy. This in itself would be problematic because even if the separation of powers may be relevant for the EU from a theoretical perspective, the EU’s constitutional charter is based on the institutional balance, and because the EU legislator (dep-ending on its constellation) does not necessarily have (qualitatively) the same democratic underpinnings as the legislator in national states, undermining its superiority claim.

Alternatively one could argue that despite the possibility for the EU insti-tutions to control agencies, the notion of delegation is not suited to describe their empowerment because a ‘true’ delegation is not at issue. This argument is akin to the Europeanization argument (cf. section IV 5.4.1.2) and was made by the CLS in its opinion on the amended 1984 proposal for OHIM (cf. sec-tion IV 5.4.3.2). The CLS noted that ‘because [these powers] have not at the moment been vested in any Community institution […] the decisions of the Court in the Meroni Case do not seem to apply in this context’.563 The Legal Service then referred to the Köster case, noting that Article 155 EEC stressed the Council’s freedom to decide (or not) to empower the Commission,564 seem-ingly ignoring the Romano judgment. In any case it has been noted (see section IV 5.4.3.2) that following the Lisbon Treaty, the decision whether or not to empower the Commission has been objectivized. As a result, the reasoning of

561 More precisely, the legislator could only confer but not delegate these powers. Alternatively, this question could also be solved by looking at the Commission as an integral part of the legisla-ture. Through its proposal the Commission would then delegate its powers. Bergkamp and Park indeed remark: ‘Conceivably, the legislative proposal made by the Commission could be viewed as an act of permanent delegation.’ See Lucas Bergkamp and DaeYoung Park, ‘The Organizational and Administrative Structures’, in Bergkamp (ed), The European Union REACH Regulation for Chemicals— Law and Practice, Oxford, OUP, 2013, p 24 at footnote 7. However, while the Commission indeed forms part of the legislator, this view is not followed here, since solely the Parliament and Council adopt the legislative act.

562 See n 495.563 Legal Service of the Council of the European Communities, n 388, p 6.564 Ibid, p 8.


the Legal Service, already contestable in 1985, no longer seems to hold under current EU primary law.

The ruling in Short- selling could have brought further clarity to this issue. However, whereas AG Jääskinen qualified the empowerment of ESMA as a con-ferral, the Court ignored the issue.565 The Court therefore also missed an op-portunity to clarify how the empowerment of EU agencies relates to the (other) delegations in the EU legal order, and primarily whether Meroni and Romano could be linked with the Köster case law.566 After all, in the German and US legal systems, no genuine distinction is made between delegations to the core executive (President and Bundesregierung) or the agencies (cf. section V 3). An analysis of the problem in Short- selling in relation to the power to adopt implementing and delegated acts under Articles 290 and 291 TFEU is dealt with later (cf. section IV 7.2.4.1).

5.7.3 ConclusionEven if the notion of delegation has been introduced in primary law through the Lisbon Treaty, there is no clear delegation model in the EU. Instead, there are dif-ferent legal regimes which have been defined by the Court, or in legal doctrine, as delegations. Some of these conferrals may indeed be defined as traditional dele-gations, while for others defining them as delegations seems out of place. This is foremost the case for the empowerment of the Commission under Article 291 TFEU, but also for the empowerment of international organizations, where the Court’s emphasis is not so much on control by the EU but rather on its autonomy. If one were to hold on to the concept of delegation to govern the EU’s relationship with international organizations, the consequence would be that the EU would be barred from developing its international relations, which in turn would be in conflict with the Treaty objectives.

Another regime may be noted in Article 290 TFEU, which itself speaks of dele-gation. For this reason it is difficult to reject that label for these conferrals, but it should still be noted that these delegations are different from the more traditional Meroni delegations because of the existence of an explicit Delegationsnorm. The situation is then most problematic for the EU agencies where the existence of a delegation cannot be ruled out, since these bodies may easily be controlled, but where none of the existing delegation models seem to apply, because of a lack of hierarchical relation, a lack of a Delegationsnorm, or the inapplicability of the nemo plus iuris rule.

565 Clément- Wilz remarks that the use of Article 114 TFEU as a legal basis even suggests that a delegation from the Member States (rather than from one of the EU institutions) is at issue when EU agencies are empowered. See Clément- Wilz, n 94, pp 347– 8.

566 In the Bankers’ bonus cap case, AG Jääskinen may have further blurred the difference between Köster and Meroni. See Opinion of AG Jääskinen in Case C- 507/ 13, UK v. Parliament and Council, ECLI:EU:C:2014:2394, para 62. For earlier examples or suggestions of a merging of Köster and Meroni, see n 376 and 459.


5.8 EU agencies and an EU Meroni doctrine

In this last section dealing with Meroni, a doctrine for the EU, staying close to the original doctrine, will be presented. This may then be juxtaposed with the solu-tions of the AG and the Court in the Short- selling case.

What could the Meroni doctrine mean for EU agencies when it is transposed as set out previously (see section IV 5.6) to the EU context? The four reinterpreted rules would mean when the legislator establishes an agency (i) that it cannot dele-gate powers which it does not possess itself; (ii) that the agency should be granted powers explicitly (even if the enabling clause may be worded generally); (iii) that the legislator should state the reasons why an agency is necessary and an appropriate means to achieve the objective(s) pursued; and lastly (iv) that the agency, following the elaboration of AG Van Gerven’s scheme, might be granted powers of practical execution, powers of management and administration, or powers to implement or concretize the decisions of principle adopted by the legislature and the Commission.

Granted, this re- interpretation of the ECSC Meroni doctrine— resulting in a Meroni- light— is debatable, on at least two accounts. Firstly, the dichotomist dis-tinction between executive and discretionary powers has been recast. In doing so the category of powers which cannot be delegated has shrivelled to the decisions of principle, other powers in principle being delegatable to internal bodies,567 even if they would fall under ‘discretionary’ powers in the original dichotomist categorization. Secondly, the scope left by Meroni for agencification would seem limited, since it cannot be said that the legislature, when empowering an agency, is delegating its own powers. From the section on the notion of delegation, it re-sults that this may be overcome by accepting a (hierarchical) primacy of the legis-lature (vis- à- vis the Commission) or by qualifying the empowerment of agencies as a conferral for which only part of Meroni would apply.

5.8.1 Meroni applied to ESMA’s powers under Regulation 236/ 2012: Short- selling

As noted, the Court of Justice itself had never actually applied its Meroni doctrine to the empowerment of an EU agency but, following the adoption of Regulation 236/ 2012,568 it was invited to do so by the UK, which challenged Article 28 of the regulation. It was not the first time that the UK had tried to reconfirm a strict reading of Meroni in the area of financial regulation,569 but now the UK

567 Whether a specific delegation is acceptable will depend on the control provided for. It is clear that even a delegation of a power of practical execution is impermissible if the delegate authority is beyond any measure of control.

568 Regulation (EU) 236/ 2012, OJ 2012 L 86/ 1.569 Two days before the European Council Summit of 9 December 2011 UK Prime Minister

Cameron declared the following in the House of Commons: ‘I think we should celebrate the fact that [the financial services sector] is a world- class industry, not just for Britain but for Europe— but it is absolutely vital for us to safeguard it. We are currently seeing it under continued regulatory attack from Brussels. I think that there will be an opportunity, particularly if there is a treaty at 27,


brought the issue to the Court, inter alia arguing that ESMA’s empowerment under Article 28 of the regulation violated Meroni.570

This plea raised the fundamental questions which have figured in the preceding sec-tions: (i) does Meroni apply to the EU legal order post- Lisbon, (ii) does Meroni apply to the empowerment of EU agencies, and (iii) what is a discretionary power? The options open to the Court in addressing this plea ranged from a rejection of Meroni’s relevance, over a re- interpretation of Meroni, to a strict application of the doctrine.571

5.8.1.2 The Opinion of AG JääskinenIn the event that the Court would not follow him in solving the case based on the UK’s fourth plea (see section IV 1.1.3.2), AG Jääskinen also analysed the other pleas. As regards Meroni he noted that the Court’s preoccupations in that case

were twofold; first, the Court was concerned about the absence of any judicial review of acts of the Brussels agencies; secondly, the Court sought to prevent the High Authority from delegating powers that were wider than its own, and which were so broadly defined as to be arbitrary. In my opinion, the Court was therefore upholding the constitutional imperatives of effective judicial control and institutional balance.572

As to the question whether Meroni could be applied to the EU agencies today, the AG observed:

[T] he Meroni case law remains pertinent in the context of delegation of implementing powers to an agency […] in that (i) powers cannot be delegated to an agency that are dif-ferent from the implementing powers the EU legislature has conferred on the delegating authority, be it the Commission or the Council, and (ii) the powers delegated must be sufficiently well defined so as to preclude arbitrary exercise of power […] The delegating authority ‘must take an express decision transferring them and the delegation can relate only to clearly defined executive powers.573

Dealing with the problem that the legislature is not actually delegating its own powers, the AG further noted that:

Indeed, the EU legislature is not acting as a ‘delegating authority’ in the sense of the Meroni judgment when it confers implementing powers on institutions, agents, or other bodies of the Union, but a constitutional actor exercising its own legislative competence as

to ensure that there are some safeguards— not just for the industry, but to give us greater power and control in terms of regulation here in the House of Commons. I think that that is in the interests of the entire country, and it is something that I will be fighting for on Friday.’ See Hansard of 7 December 2011, column 294. In return for agreeing on a Treaty amendment to tackle the euro crisis, the UK delegation demanded safeguards alluded to in the previous citation. The UK thus demanded that the unanimity rule would be reintroduced for financial sector legislation and that the ESAs’ executive powers would be clearly set out, the exercise of discretion being reserved to the Member States’ authorities. Should the latter demand have been granted it would have meant a co-dification of Meroni in primary law, at least for the agencies in the financial sector. In the end, the Fiscal Compact was simply concluded as a separate treaty outside the EU framework.

570 See Case C- 270/ 12: Action brought on 1 June, OJ 2012 C 273/ 3.571 For a discussion of these options see Chamon, n 24, pp 157– 8.572 Opinion of AG Jääskinen in Case C- 270/ 12, n 74, para 64. 573 Ibid, para 88.


conferred on it by the higher constitutional charter, i.e. the Lisbon Treaty. The executive and judicial powers that the EU legislature can confer on institutions or bodies are quali-tatively different from its own powers.…However, […] the prohibition on inordinately broad and/ or arbitrary implementing

powers remains relevant to the assessment of the legality under EU constitutional law of Article 28 of Regulation No 236/ 2012. The European Union legislature may not vest ESMA with an authority to pass implementing measures that would breach this principle because an implementing power will be validly conferred only if it is sufficiently specific, that is to say it must clearly specify the bounds of the power conferred […] In addition, ESMA cannot be empowered to take policy decisions.574

The AG therefore proposed that the legislator could confer executive powers which it does not hold itself, but that these could not be different from those ordinarily transferred to the Commission and Council.575 In assessing whether no discretionary powers were conferred on the ESMA, the AG stressed that the ESMA was (only) empowered to adopt certain pre- defined instruments, directed at pre- defined categories of addressees, under pre- defined conditions, following a pre- defined procedure. In addition to the framework set out in the legislative regulation itself, the AG noted that the Commission had further elaborated this in a delegated regulation.

While the AG’s observations were correct, this still did not mean that the ESMA would not exercise discretionary powers. The AG further addressed this point by noting that the ESMA had no discretion on whether to act or not, at the same time seemingly accepting that the ESMA would exercise discretionary powers,576 once it is determined that it should act:

[W] hile the application of the criteria and terms of Article 28 of Regulation No 236/ 2012 require complex assessments of facts, ESMA was established for precisely this purpose […], the need to make such technically difficult assessments on an objective and non- political basis is an argument in support of conferral of the powers concerned on an expert agency rather than on the Commission.577

The solution adopted by the AG resulted in a partially applicable reinterpreted Meroni, since powers are conferred rather than delegated to agencies. Obviously the policy decisions which are in any case reserved to the legislature cannot be taken, according to the AG, by agencies.578 Further, in order not to grant

574 Ibid, paras 91– 3. On the policy choices, the AG noted that the ‘essential value judgements’ had been made by the EU legislator. See ibid, para 97.

575 It may be noted that AG Jääskinen, just like the CLS in its opinion on the OHIM, cf. section IV 5.4.3.2, spoke of a conferral rather than a delegation. Because of this, the AG and CLS con-cluded that part of Meroni does not apply, but the AG, unlike the CLS, still stressed the prohibition to delegate discretionary powers.

576 This difficult element in the AG’s analysis was also noted by Skowron, who found that the AG did not make clear how the sanctioned ‘complex assessments of fact’ by the ESMA differed from the prohibited ‘discretionary powers’ under Meroni. See Magdalena Skowron, ‘Die Zukunft eu-ropäischer Agenturen auf dem Prüfstand’, (2014) 49 Europarecht 2, pp 258– 9.

577 Opinion of AG Jääskinen in Case C- 270/ 12, n 74, para 99 (emphasis added).578 Ibid, para 93.


discretionary powers the implementing power conferred must be sufficiently spe-cific, that is, the bounds of the power conferred must be clearly specified.

5.8.1.3 The ruling of the Court and its implications for the debate on agencification

Where the AG’s opinion shows an awareness of the legal issues identified in pre-vious sections, the Court’s pragmatism resulted in a ruling that appears much more like a simplification exercise.579 The Court’s solution was a re- interpretation of Meroni pushed to such an extreme that the Court appears to have overruled its jurisprudence.

While the Court confirmed Meroni’s applicability to EU agencies, it revisited its jurisprudence on two accounts, resulting in a Meroni- light.

Firstly, the Court reduced the Meroni doctrine to one rule, since in Meroni it ‘stated, in essence, that the consequences resulting from a delegation of powers are very different depending on whether it involves clearly defined executive powers […], or whether it involves a ‘discretionary power implying a wide margin of dis-cretion […]’580— this even if the Court in its previous jurisprudence had recalled a more complete Meroni doctrine,581 and the AG had presented a Meroni doctrine which was more faithful to the original.582

The Court continued by ascertaining whether the powers in casu were discre-tionary or not. The Court emphasized the context in which the ESMA would exercise that power, recalling at least five elements in the regulatory framework which circumscribed the margin left to the ESMA. Firstly, the ESMA can only exercise the power concerned in case of a systemic risk and when no national authority has acted adequately. Still, the Court did not take issue with the fact that it is up to the ESMA itself to decide whether these conditions are met.583 Secondly, ESMA has to take into account the criteria of paragraph three of Article 28. Thirdly, before taking any action, the ESMA is required to consult the European Systemic Risk Board (ESRB).584 Fourthly, it found that the decisions adopted by the ESMA were only of a temporary nature, somehow suggesting that a measure’s limited validity in time may affect the nature of the power pursuant to which

579 Since the proper implementation of EU law often depends on EU agencies, Short- selling should be welcomed from a practical point of view. For this reason Bernard, although critical of the Court’s reasoning, welcomed the Court’s pragmatic approach. See Elsa Bernard, ‘Quels pouvoirs pour les agences de l’Union européenne?’, (2014) 50 CDE 3, pp 658– 9.

580 Case C- 270/ 12, n 92, para 41 (emphasis added).581 Case C- 301/ 02 P, n 452, para 43.582 See Opinion of AG Jääskinen in Case C- 270/ 12, n 74, para 62.583 This was also remarked upon by Ohler: see Christoph Ohler, ‘Rechtsetzungsbefugnisse

der Europäischen Wertpapier- und Marktaufsichtsbehörde (ESMA)’, (2014) 69 JZ 5, p 250. See also Skowron, n 584; Rostane Mehdi, ‘Le pouvoir de décision à l’épreuve de “l’agenciarisation” de l’Union— Quelques questions constitutionnelles’, in Bertrand, Picod and Roland (eds), L’ identité du droit de l’Union européenne— Mélanges en l' honneur de Claude Blumann, Bruxelles, Bruylant, 2015, p 702.

584 Skowron has rightly questioned whether the consultation duty at issue really restricts the ESMA’s powers: see Magdalena Skowron, ‘Kapitalmarktrecht: Rechtmäßigkeit der Eingriffsbefugnisse der ESMA nach Art. 28 Leerverkaufsverordnung’, (2014) 25 EuZW 9, p 351.


the measure is adopted. Lastly, the Court referred to the Commission’s power to further elaborate the regulatory framework, which the Commission had done through Delegated Regulation 918/ 2012.585 Here the Court did not seem to take issue with the fact that agencies such as the ESMA have been granted the power to draft the delegated regulations of the Commission (cf. section I 1.2.1.3.1.f) and that in casu the ESMA had indeed drafted the delegated regulation.586

Did the cumulative effect of these elements constraining the ESMA result in a non- discretionary power being granted to it? This is rather doubtful, and is being debated,587 but the Court concluded that the ESMA’s powers ‘are precisely deline-ated and amenable to judicial review in the light of the objectives established by the delegating authority. Accordingly, those powers comply with the requirements laid down in Meroni’.588 Whether the judicial review referred would be mean-ingful is unclear, given the developing jurisprudence of the General Court in relation to its standard of review when it scrutinizes agency decisions (cf. section IV 5.4.2.2). Interestingly, the Court did not conclude that the ESMA’s powers are executive in nature, as one would expect from Meroni, but instead that they are ‘precisely delineated’. Here the Court ignored that, under the fourth rule as pre-viously identified (see section IV 5.6.2.2), such powers still need to be politically controlled. In Short- selling, however, the Court contents itself with affirming a continued judicial control.589

In a way, the Court had announced its conclusion in the beginning of its ana-lysis when it noted: ‘[U] nlike […] in Meroni v. High Authority, the exercise of the powers under Article 28 of Regulation No 236/ 2012 is circumscribed by various conditions and criteria which limit ESMA’s discretion.’590 However, this prelim-inary observation is debatable. After all, the Brussels Agencies formally func-tioned under the responsibility of the High Authority, which had also laid down guiding principles. As noted previously (see section IV 5.1.4.2.3) in Meroni, the Court found that these criteria still left too much leeway to the Office and Caisse, as their decisions could not ‘be the result of mere accountancy procedures based on

585 See Delegated Regulation 918/ 2012 of the European Commission, OJ 2012 L 274/ 1.586 See ESMA, Technical Advice on possible Delegated Acts concerning the regulation on short

selling and certain aspects of credit defaults swaps, ESMA/ 2012/ 263, pp 65– 7.587 Kohtamäki rightly concluded that the Court’s argumentation was hardly convin-

cing: see Natalia Kohtamäki, ‘Die ESMA darf Leerverkäufe regeln’, (2014) 49 Europarecht 3, p 328. Similarly, see Clément- Wilz, n 94, pp 341– 2. See also Mehdi, n 583, p 701 et seq. Martucci on the other hand remarked that ‘on ne peut qu’approuver la Cour de justice lorsqu’elle conclut: aucun “ large pouvoir discrétionnaire” n’est délégué à l’AEMF.’ See Francesco Martucci, ‘Les pouvoirs de l’Autorité européenne des marchés financiers à l’épreuve du droit constitutionnel de l’Union’, (2014) RAE 1, p 198. For other authors that have welcomed the Court’s ruling rather positively, see Jean- Claude Bonichot, ‘À propos de l’attribution du pouvoir réglementaire à l’Autorité européenne des marchés financiers’, (2014) RFDA 2, p 330; Berend Jan Drijber, ‘Verbod op short selling— grondslag bevoegdheden ESMA’, (2014) Ondernemingsrecht 7, pp 352– 4.

588 Case C- 270/ 12, n 92, para 53.589 Adamski also notes that the Court in Short- selling did not address, at all, the question of pol-

itical accountability: see Adamski, n 94, p 832.590 Case C- 270/ 12, n 92, para 45.


objective criteria laid down by the High Authority’.591 However, the criteria laid down by the legislator and the Commission are similarly vague and generally worded. Whether ESMA’s future decisions based on Article 28 of Regulation 236/ 2012 would then be the result of a ‘mere accountancy procedure’ based on ‘objective criteria’ laid down by the legislator and the Commission is questionable.

Instead, both in Short- selling and in Meroni, certain general criteria had been laid down, but it appears that the Court was less demanding in Short- selling.592 Lastly, and as was noted previously (see section IV 5.1.4.2.3), in Meroni the Court further remarked that the High Authority had had to intervene twice because the Office and Caisse could not find an agreement. The existence of this disagreement was an indication for the Court that discretionary powers were involved. Here it suffices to note that ESMA may exercise its contested power precisely when there is a disagreement among national authorities.

Because Meroni was narrowed down to a single requirement and because, sub-sequently, that requirement was hollowed out, Meroni as a limit to agencification has become almost ethereal. Here a parallel may be drawn with the development of the comitology system (cf. section III 4.4), where the Court also did not inter-fere when the institutions engaged in institutional experimentation.

How does the Short- selling case link with the analysis in the previous sections? Obviously the Court did not take over the EU Meroni doctrine, as suggested pre-viously (see section IV 5.6.2), but Short- selling is far from the end of the debate on Meroni. The Court for instance recalled that the prohibition of delegating dis-cretionary powers was the essence of Meroni, without necessarily discarding other elements of the doctrine. Because the Court did not pay attention to the nemo plus iuris requirement, the question of whether agencies are empowered through delegations or conferrals was not touched upon either, even though the Inet Hellas case (cf. section IV 5.5.1.3) could have been used as a stepping stone. As noted, the Court did not conclude that the power under Article 28 was ‘merely execu-tive’ as required by the original Meroni ruling. Instead it held that that power was ‘precisely delineated’. The Court left a lot of leeway to the legislator by empha-sizing the context in which a power is exercised. Generally speaking, the Court stressed three elements: firstly, an agency should only be competent in exceptional cases; secondly, the agency should follow a procedure involving other actors; and thirdly, the agency should act on pre- defined criteria.593 Here again, it could prove useful in the future to put limits on the legislator’s very large discretion by empha-sizing the third rule of the EU Meroni doctrine as identified previously (see section IV 5.6.1). Alternatively, should a sound legal basis be inserted in the Treaties (cf. section VI 2.2), the Court could also do away with Meroni and instead generalize its Köster jurisprudence.

591 Case 9/ 56, n 223, p 153 (emphasis added).592 This is also noted by Clément- Wilz, n 94, p 341.593 Note however, that the General Court in the Heli- Flight case remarked that the EASA’s com-

petence was not circumscribed by a pre- defined procedure or pre- defined criteria, without taking issue with this. See Case T- 102/ 13, n 379, para 90.


5.8.2 Concluding remarksThe Court’s ruling resulted in a double simplification, narrowing Meroni down to a single requirement and applying that requirement very flexibly.

Short- selling does not allow a straightforward integration in the EU Meroni doc-trine identified in the previous sections. This is firstly because the Court confirmed that Meroni applies to agencification and secondly because the Court largely ig-nored the issue of control. An integration of Short- selling is still possible, however, since the Court only referred to the ‘essence’ of Meroni. Further, the Court in the future could still emphasize the issue of control by relying on the institutional bal-ance, which was completely ignored here (cf. section IV 7.2.4.1.3). Alternatively, the Court could implicitly come back to its first basic choice and substitute its Meroni test in empowering agencies for a Köster- test, without a distinction being made between legislative empowerments of the primary or subsidiary executives, just as in the US (cf. section V 1.3).

5.9 Conclusion

This extensive analysis of the Meroni ruling in the present chapter was deemed necessary to provide a proper understanding of the ruling and of its relevance for agencification.

The lack of a thorough debate in legal doctrine appears clear from the fact that there is no real consensus on which doctrine the Court laid down, which is curious since everyone seems to agree that there is indeed ‘a’ Meroni doctrine. Meroni is further plagued by the fact that some authors present the ruling as a monolith, sometimes reducing it to the institutional balance or to the simple prohibition on delegating discretionary powers. The discussion of this academic dissensus has shown the need to come to a more thorough understanding and a greater consensus on Meroni. Therefore an ECSC Meroni doctrine has been pro-posed (see section IV 5.6.1).

Because Meroni is much more than a monolith, the arguments refuting its relevance for the EU agencies today were not accepted (see section IV 5.4.1), but these arguments (and those for a Meroni- light) did form the background for an EU Meroni doctrine (cf. section IV 5.6.2). The EU Meroni doctrine, unlike the ECSC Meroni doctrine, still allows for a lot of debate. The added value of the EU Meroni doctrine as presented is that (i) it stays true to the ori-ginal doctrine, since some rules governing delegation (in its broad sense) have a universal character, and (ii) it allows more flexibility than the ECSC Meroni doctrine.

The EU Meroni doctrine as presented is in line with most of the Court’s subse-quent references to Meroni, although the Court’s reasoning in some of these rul-ings was criticized. These rulings were also studied more in depth (cf. section IV 5.5). As a result, the incorrect representation of cases such as Tralli and Alliance for Natural Health as confirmation that ‘Meroni applies to EU agencies’ was also

Prelude to the Institutional Balance: Romano 249

rejected. Since none of the institutions seems to have a clear view on Meroni, the EU Meroni doctrine presently identified did not draw on the institutions’ views (cf. section IV 5.4.3).

To substantiate the analysis, a study was first devoted to identifying the dele-gation notion in the EU legal order (see section IV 5.7), and it was concluded that it is not straightforward to treat the empowerment of an EU agency as a Meroni- style delegation. At the same time, it had already been found (in section IV 5.4.1.1) that part of Meroni enshrines a set of rules that apply generally. Where the AG in Short- selling referred to a more complete Meroni doctrine, the Court reduced it to a single requirement. The AG was then confronted with the question of whether the empowerment of the ESMA was the result of a delegation or of a conferral, while the Court ignored this issue. In any event, both the AG and the Court settled on a restrictive reading of discretionary powers, resulting in a Meroni- light. By focusing on the nature of the powers, the Court ignored the real issue in Meroni, as identified (in section IV 5.6), which revolved around the possibility to exercise political and judicial control.

While Short- selling is a landmark case on agencification, it has evidently not settled the issues underlying this institutional development in a definitive fashion. If agencification respects the four rules identified (cf. section IV 5.6.2) (or three rules in cases of conferrals) this will enhance the legitimacy of the process. Following Short- selling agencification undoubtedly suffered in this regard, given the way in which the Court addressed the pertinent grievances put forward by the UK. Further, the Court completely neglected the institutional balance. This will be taken up in the following chapters.

6 A Prelude to the Institutional Balance: Romano

In the previous sections, arguments have been proposed against conflating the balance of powers in Meroni with the notion of institutional balance. The starting point for the sections devoted to that balance is the Romano ruling of the Court of Justice.594 As will be shown, the facts and legal context of the Romano judgment much better resemble those of agencification today, further stressing how remark-able it is that Meroni is taken as the guiding jurisprudence for agencies instead of Romano. As a result, it may be surprising for the reader to find that the present section is much shorter than the one on Meroni. The reason for this is that Romano is a much less complicated and debated case.

The reason why Romano is discussed as a prelude to the institutional balance is because the Court implicitly, yet undeniably, relied on the institutional balance in this case. The ruling therefore provides the link to an analysis of agencification through the lens of the institutional balance.

594 Case 98/ 80, n 528.


6.1 The Romano case

As was remarked elsewhere,595 Meroni only tells part of the Court’s tale on dele-gation, the 1981 Romano ruling being another interesting case. The repercussions of Romano for agencification appear even more severe than those of Meroni, which might also explain why Romano has not been proven to be popular jurisprudence in order to ‘frame’ agencification.

At least three fundamental aspects differentiate Romano from Meroni, bringing it closer to the factual and legal situation in which the EU agencies operate today. For one, the Romano case was ruled on under the EEC Treaty and not under the ECSC Treaty. Secondly, Romano concerned a conferral by the legislator, not a delegation by the Commission or High Authority. Lastly, the delegate authority in the Romano case was a body established under secondary law and not a body established by private law, albeit that the Administrative Commission in Romano did not have legal personality.

A last preliminary remark is on the Administrative Commission itself. This body was and is596 an institutional curiosity. Maas was the first (and probably only) author to dedicate a study to this body.597 The EU legislation providing for the establishment of the Administrative Commission was originally intended to be an international convention drafted under the auspices of the High Authority and the International Labour Office. When the final text of the convention was prepared, the entry into force of the Rome Treaties was imminent. Eventually the European Commission proposed to replace the Convention with an EEC regulation. The Administrative Commission then was a remnant of the regula-tion’s origins in international law in that it was made competent to interpret the Regulation and to decide in disputes on the interpretation and application of the Regulation. Maas remarked in 1966: ‘It seems that in changing the convention into an E.E.C. Regulation not enough account was taken of the fact that the entire Regulation was given another complexion in accordance with the structure of the Community, which was perhaps not quite so clear in 1958, but which has since become much more sharply outlined.’598 Maas analysed the powers of the Administrative Commission in the light of Meroni and concluded that the dele-gation was not defensible,599 proposing that the Administrative Commission be changed into a simple advisory body.600

595 See Chamon, n 391, pp 1060– 6. This article has formed the basis of the present section.596 For a recent discussion, see Henrik Wenander, ‘A Network of Social Security Bodies—

European Administrative Cooperation under Regulation (EC) No 883/ 2004’, (2013) 6 REALaw 1, pp 47– 50.

597 Maas, n 255 (CMLRev); Maas, n 255 (CDE).598 Maas, n 255 (CMLRev), p 55.599 Similarly, see Kapteyn, n 261, p 49.600 Von Bogdandy et al repeated this call forty years later: see von Bogdandy, Arndt, and

Bast, n 119, p 132. However, the new regulation replacing Regulation 1408/ 71 has neither abol-ished the Administrative Commission nor changed it into a comitology committee. See Articles 70 and 71 of Regulation (EC) 883/ 2004 of the European Parliament and of the Council, OJ 2004 L 166/ 1.


6.1.1 The actual Romano caseThe main issue in Romano was a straightforward problem, albeit of a rather tech-nical nature. Mr Romano was an Italian worker living in Belgium who was entitled to both an Italian and a Belgian pension. Under EU rules, the Belgian pension was not to be paid in full, but had to be reduced by the amount of the Italian pension. One of the contentious issues then related to the exchange rate to be used to con-vert the Italian pension into Belgian Francs, given the fluctuating exchange rate.

The Belgian social security institute justified the exchange rate it had used by referring to Decision No 101 of the Administrative Commission on Social Security of Migrant Workers,601 which the Council had granted the power to fix the exchange rates applicable under Article 107 of Regulation 574/ 72.602 The Labour Court in Brussels thus asked whether the decision of the Administrative Commission was lawful.

In his preliminary remarks AG Warner made the following pertinent observa-tions on the Brussels’ Court’s question:

It gives rise in limine to a point of a constitutional nature, viz. whether it was compat-ible with the Treaty for the Council to confer a legislative power on the Administrative Commission.

[…]The Administrative Commission is not of course a creature of the Treaty. It is the crea-

ture of Article 80 of Regulation No 1408/ 71, which provides that it ‘shall be attached to the Commission’; that it shall be ‘made up of a government representative of each of the Member States, assisted, where necessary, by expert advisers’; that ‘a representative of the Commission’ shall attend its meetings ‘in an advisory capacity’.603

While the Administrative Commission’s specific power at issue was provided for by the Council in its executive capacity (in Regulation 574/ 72), the latter had merely implemented the (formal) legislative provisions of Article 81 of Regulation 1408/ 71.604

The AG remarked that the Council was enabled under the Treaties and specif-ically Article 155 EEC to confer (material) legislative powers on the Commission, but that the Treaties nowhere suggested that such a conferral was possible to the Administrative Commission. The AG continued:

Moreover Article 173 of the Treaty gives this Court jurisdiction to review the legality of acts of the Council and of the Commission, whilst Article 177 gives it jurisdiction to rule on the validity and interpretation of acts of the institutions of the Community. The Administrative Commission is not an institution of the Community as defined by the Treaty.605

601 Decision No 101 of the Administrative Commission, OJ 1976 C 44/ 3.602 Regulation (EEC) 574/ 72 of the Council, OJ 1972 L 74/ 1.603 Opinion of AG Warner in Case 98/ 80, Romano, [1981] ECR 1259, pp 1263– 4.604 One could then argue that the Council in its executive capacity had truly delegated powers

(rather than having conferred powers in its legislative capacity) but this would have raised the same problem as in Inet Hellas (cf. section IV 5.5.1.3).

605 Opinion of AG Warner in Case 98/ 80, Romano, [1981] ECR 1259, pp 1264– 5.


The AG concluded therefrom that the Court did not have a direct jurisdiction to answer the referring Court’s question, since the Administrative Commission could not be brought under Article 177 EEC. Because the AG did not consider a solution similar to the one adopted in Les Verts (cf. section V 6.1.1.1), the logical result of this position was the immunity of the Administrative Commission’s acts from judicial review. The AG commented:

The idea that there may be set up for the Community an administrative body empowered to make binding decisions, but whose decisions are, in themselves, incapable of review by this Court seems to me incompatible with the scheme of the Treaty.606

Based on this argument alone, the AG felt that the Council’s delegation of powers to the Administrative Commission was unlawful. For the sake of completeness, the AG in addition referred to the established case law by the Court to support his conclusion. Curiously, the Meroni ruling did not figure in the cases cited. The AG did refer to the Köster case, noting that the compatibility of the comitology man-agement procedure with Article 155 EEC and the provisions on judicial protection precisely rested on the mere consultative function of the management committee. The AG concluded that Article 107(4) of Regulation 574/ 72 was invalid and, as a consequence, a decision of the Administrative Commission adopted pursuant to that provision could not have any legal effect.

The Court noted:

[I] t follows both from Article 155 of the Treaty and the judicial system created by the Treaty, and in particular by Articles 173 and 177 thereof, that a body such as the Administrative Commission may not be empowered by the Council to adopt acts having the force of Law.607

Just like the AG, the Court concluded, without referring to the Meroni ruling, that the delegation at issue was incompatible with the Treaties, and this on two grounds: firstly because Article 155 EEC (later Article 211 TEC, now repealed by the Treaty of Lisbon and in substance replaced by Articles 290 and 291 TFEU) only enabled the Council to confer powers on the Commission, and secondly be-cause Articles 173 and 177 EEC (Articles 263 and 267 TFEU) did not provide ad-equate remedies. As was noted previously, the requirement of a waterproof system of judicial protection was the very same concern which occupied AG Roemer’s mind in Meroni (cf. section IV 5.1.3). The AG and the Court therefore relied on the same grounds, but only the ground related to the legal remedies was elabo-rated, and this only by the AG.

6.1.2 Assessment of the caseBefore looking at Romano’s reception in legal doctrine and how it is perceived to relate to Meroni, it is interesting to look at a similar ruling prior to Romano which

606 Ibid, p 1265. 607 Case 98/ 80, n 528, para 20.


also related to a decision by the Administrative Commission. In van der Vecht the Court had been asked by the Dutch Centrale Raad van Beroep about the extent to which a judicial authority such as the Raad was bound by a decision of the Administrative Commission.608 The applicable regulation at that time was still Regulation 3,609 which was later replaced by Regulation 1408/ 71. However, just like the latter, Regulation 3 provided for the creation and empowerment of the Administrative Commission. Advocate General Gand observed that

if the regulation purported to empower the Administrative Commission to give authentic interpretations binding on other judicial authorities it would infringe Article 177 [EEC] […] Like the jurisdiction of this Court, the jurisdiction of such national courts or tribu-nals would be reduced to nothing if it were limited by a decision of the Administrative Commission.610

AG Gand concluded that the decisions of the Administrative Commission could at the most be afforded a ‘particularly authoritative’ quality, while remaining non- binding. The Court agreed with the analysis of AG Gand to the effect that the decisions could only have the status of opinion, adding that ‘[n] o other interpret-ation of Article 43 would be in accordance with the Treaty, in particular Article 177 thereof, which establishes a procedure to ensure the uniform judicial inter-pretation of the rules of Community law’.611

Although the same issue was raised in both cases and the Court (partially) relied on the same reasoning,612 only AG Warner referred to van der Vecht in Romano. Still the Court also added a second element in Romano, that is, the ref-erence to Article 155 EEC. This is of course highly relevant following the entry into force of the Lisbon Treaty. Although this point will be returned to in greater detail later, the amendments of the Lisbon Treaty now ensure legal protection vis- à- vis EU agencies or bodies like the Administrative Commission. Had the Court in Romano merely referred to its ruling in van der Vecht, the constitutional objections to a delegation of powers to the Administrative Commission would have been lifted by the Treaty of Lisbon. The ‘Romano doctrine’ now depends on the objection resulting from Article 155 EEC.

6.1.2.1 The relation between Romano and MeroniPerhaps because of this conciseness there is no consensus (pre- Short selling) among those authors that also refer to Romano on the precise meaning of this ruling and how it relates to the principles laid down in the Meroni judgment.613 Von

608 See Case 19/ 67, Van der Vecht, [1967] ECR 345.609 Réglement No 3 du Conseil, OJ 1958 30 561.610 Opinion of AG Gand in Case 19/ 67, Van der Vecht, [1967] ECR 345, p 361.611 Case 19/ 67, n 608, p 355.612 To be precise, the AG and Court seemed equally concerned about the jurisdiction of both the

Court and the national courts under the preliminary ruling procedure. This procedure is only a part of the ‘judicial system established by the Treaties’ on which the AG and Court relied in Romano.

613 Some authors mention Romano together with Meroni but do not attempt to clarify the rel-ationship between the two: see inter alia Schammo, n 273, pp 30– 6; Kapteyn, VerLoren Van


Bogdandy et al claim that the Court ‘applied the Meroni doctrine’ in Romano.614 According to Comte, Romano simply confirms Meroni.615 Geradin concurs with this, but notes the two should still be distinguished as the Court did not expli-citly rely on the principle of institutional balance in Romano.616 This should be juxtaposed with Griller and Orator, who claim that the institutional balance al-ready cited in Meroni was central to the prohibition in Romano.617 Indeed, Griller and Orator’s view carries more weight, since the Court referred to Article 155 EEC. Delegating the power at issue to a body other than the Commission could then be seen as encroaching on its powers and prerogatives, one of the three sub- principles constituting the principle of institutional balance according to Lenaerts and Verhoeven.618 The Court therefore took issue not so much with the Council delegating but with the fact that the Council did not delegate powers to the Commission.619

Although the institutional balance, through Article 155 EEC, was part of the Court’s reasoning in Romano, it cannot be said with certainty to have made up the core of the Court’s objection to the delegation as Griller and Orator seem to suggest. It seems impossible to make such a definite conclusion on this crucial element in the Romano judgment given the scant reasoning of the Court. The more thorough analysis by the AG would even seem to point in the opposite dir-ection, as the AG insisted not so much on Article 155 EEC but on the judicial system created by the Treaties. Remmert indeed suggests that it was because of this lacuna in judicial protection that the Administrative Commission could not wield external competences.620 On the other hand, Romano precisely adds the element of 155 EEC to the van der Vecht jurisprudence.

While Koenig and others assert that Meroni and Romano show a restrictive approach of the Court, they only conclude from those two judgments that com-prehensive law- making powers may not be delegated,621 which unfortunately does

Themaat, and Gormley, n 359, pp 244– 6; De Schutter, n 278, p 41; Florian Ermacora, ‘The Agency for the Cooperation of Energy Regulators (ACER)’, in Jones (ed), EU Energy Law, Leuven, Claeys & Casteels, 2010, Vol. 1, p 260.

614 von Bogdandy, Arndt, and Bast, n 119, p 132. Similarly Bernard refers to Meroni when she observes: ‘La Cour a certes confirmé par la suite, que cette jurisprudence [Meroni] n’était pas uniquement valable pour les délégations intervenant dans le cadre de la CECA.’ See Elsa Bernard, ‘Accord sur les agences européennes: la montagne accouche d’une souris’, (2012) RDUE 3, p 437.

615 Françoise Comte, ‘Agences européennes: relance d’une réflexion interinstitutionelle euro-péenne?’, (2008) RDUE 3, p 495. Similarly see Clément- Wilz, n 279, p 130; Correia, n 362, p 598; Stephanou, n 276, p 620; Van Cleynenbreugel, n 279, p 394; Christian Stoffaës, ‘Foreword to the Volume’, in Della Cananea (ed), European Regulatory Agencies, Paris, Éditions Rive Droite, 2005, p 17. According to Grimm, the Court upheld Meroni in Romano, see Nathalie Grimm, The Shift of Energy Regulatory Powers under the Framework of Directive 2009/ 72/ EC, Universität Wien, 2011, PhD Thesis, pp 188– 9.

616 Geradin, n 278, p 10.617 Griller and Orator, n 273, p 18.618 Lenaerts and Verhoeven, n 284, pp 44– 5.619 This was also observed by Gautier, see Gautier, n 492, p 399.620 Barbara Remmert, ‘Die Gründung von Einrichtungen der mittelbaren Gemeinschafts-

verwaltung’, (2003) 37 Europarecht 1, p 140.621 Koenig, Loetz, and Fechtner, n 280, p 231.


not really help in clarifying both judgments. To Schütze, Romano is a drastic expression of an expansive reading of Meroni.622 Lastly, according to Türk, the Romano ruling did not just confirm but even went beyond the Meroni judgment and restricted the possibilities for delegation even further,623 the practical con-sequence being that delegating decision- making powers to outside bodies is un-lawful under the Treaties.624 As a result, although some authors mention Romano together with Meroni, none of them have undertaken to clarify the relation be-tween these two judgments. Neither have they tried to explain how the Romano judgment can still leave room for EU agencies such as OHIM, CPVO, EASA, the ESAs, etc, which have been attributed the power to take binding decisions.

Whether Türk is correct in his assertion that decision- making powers may not be delegated to outside bodies ultimately depends on how ‘force of Law’ is con-strued. Türk himself seems to interpret this as the legal force making a decision or an act binding, an interpretation that may be supported based on the English text of the judgment. However, in the French, Italian, Dutch, and German- language versions of the judgment the Court refers to ‘actes revêtant un caractère normatif, atti di carattere normative, normatieve besluiten and Rechtsakte mit normativem Charakter’. Acts having the force of Law should then be read as binding acts of general application, leaving open the possibility to adopt binding acts in indi-vidual cases. The importance of this question in the light of the current agencifica-tion should immediately be clear. True regulatory agencies, which are granted the power to adopt normative measures, would be ruled out by Romano, but it could be that Romano does not prohibit the delegation of the power to take individually binding decisions— a power already granted to a number of agencies.

Looking at the subsequent case law of the Court and the opinions of the Advocates General, one would indeed be inclined to say the Court’s prohibition in Romano only related to the delegation of ‘legislative powers’.625 However, the choice of words should be read in the light of the facts of the Romano case, in which a normative measure of the Administrative Commission was indeed at issue. Yet a literal interpretation of the Court’s ruling in Romano, resulting in the

622 Schütze, n 560, p 674 at footnote 88. Elsewhere Schütze argues that Romano is an example of simply an ‘expansive reading’ of Meroni. See Schütze, n 147, p 235.

623 Similarly Kelemen, n 282, p 185; Georg Hermes, ‘Legitimationsprobleme unabhäng-iger Behörden’, in Bauer, Huber, and Sommerman (eds), Demokratie in Europa, Tübingen, Mohr Siebeck, 2005, pp 468– 9. According to Delzangles, in Romano ‘la Cour de justice complète sa construction jurisprudentielle en interdisant, de manière plus large [que dans Meroni], toute déléga-tion d’un pouvoir normatif ’. See Hubert Delzangles, ‘L’encadrement des agences américaines de régulation, un modèle pour l’Europe?’, in Peraldi Leneuf and Normand (eds), La légistique dans le système de l’Union européenne, Bruxelles, Bruylant, 2012, p 171.

624 Alexander Türk, ‘Case Law in the Area of the Implementation of EC Law’, in Pedler and Schaefer (eds), Shaping European Law and Policy: The Role of Committees and Comitology in the Political Process, Maastricht, EIPA, 1996, p 186.

625 See Opinion of AG Lenz in Case 236/ 87, Bergemann, [1988] ECR 5125, para 41; Case C- 102/ 91, Knoch, [1992] ECR I- 4341, para 52. In other cases the Court again referred to ‘acts having the force of law’. See Case 21/ 87, Borowitz, [1988] ECR 3715, para 19; Case C- 201/ 91, Grisvard and Kreitz, [1992] ECR I- 5009, para 25. The Court merely referred to the non- binding nature of the Administrative Commission’s decisions in Fitzwilliam: see Case C- 202/ 97, Fitzwilliam, [2000] ECR I- 883, para 32.


prohibition only relating to normative measures, would run counter to the very premises in both the Court’s and the AG’s reasoning. Whether a decision is of a general or individual scope does not appear to be relevant in the light of the Treaty Articles (Articles 155, 173, and 177 EEC) invoked by the Court and the AG. If one stays true to this underlying reasoning, the scope of the prohibition contained in that judgment should indeed be broadened to any binding decision, whether of general or individual application, as Türk suggested.

It is unfortunate that since 1981, with just one exception, Romano has only been referred to in the specific context of the work of the Administrative Commission,626 its potential for the general question on delegation remaining untapped. As a result, there is not much guidance when trying to identify the meaning of Romano. Things have been further complicated because the Treaty of Lisbon seems to have nullified the Court’s objection related to the available legal remedies, only leaving its objection related to Article 155 EEC, which has since been affected by numerous Treaty changes as well. How Romano should then be understood today will be taken up in the following section, illustrated by refer-ence to the Short- selling case.

6.1.3 Romano today: The ESMA’s powers under Regulation 236/ 2012 in Short- selling

Previously (see section IV 5.8.1) the circumstances of the case and the UK’s first plea relating to Meroni have been presented. In its second plea,627 the UK invoked Romano and argued that ‘Article 28 [of Regulation 236/ 2012] purports to em-power ESMA to impose measures of general application which have the force of law, contrary to the Court’s decision in [Romano]’.628

Just as in the section on the UK’s first plea, its second plea raises the same fun-damental questions which have been dealt with here: (i) does Romano apply to the EU legal order post- Lisbon, (ii) does Romano apply to the empowerment of EU agencies, and (iii) what is the power to adopt an act having the force of law? In addition, the two first pleas of the UK combined raised the question of whether Meroni and Romano both govern a single type of delegation, that is, do they con-stitute a single legal regime? Again the same basic options were open to the Court, ranging from finding Romano irrelevant, over reinterpreting Romano, to applying the ruling without any adaptation.

626 The notable exception is the Eggers case in which the Commission and AG Rozès did refer to Romano, arguing that the (comitology) Committee on Community Transit established by Article 55 of Regulation 222/ 77 could not adopt agreements which would be binding on the national customs offices (or courts) when applying EU customs law: see Opinion of AG Rozès in Case 302/ 81, Eggers, [1982] ECR 3443, p 3456. In the end the Court did not have to deal with this question because it was only asked in the event that it would answer the referring court’s first question in the negative, which was not the case.

627 The present discussion is largely based on an earlier contribution, see Chamon, n 70, pp 380– 403.

628 See Case C- 270/ 12: Action brought on 1 June 2012, OJ 2012 C 273/ 3.


6.1.3.1 The Opinion of AG JääskinenCompared with his analysis of Meroni, AG Jääskinen was much more concise when dealing with Romano. Still, two important elements in his Opinion should be noted. The first relates to the problem of how ‘acts having the force of Law’ should be construed. Referring to the other language versions, the AG concludes that it should be understood as ‘normative acts’ rather than binding acts. Here the AG clearly opted for a Romano- light.

Even more important, the AG further noted that ‘the evolution in the EU constitutional law that occurred under the Lisbon Treaty has indeed accom-modated the pivotal concerns with which the Court had to deal in Meroni and Romano’.629As far as judicial control goes, one could agree with the AG.630 However, whether the same may be said for the Court’s reference to Article 155 EEC in Romano is unclear. Already at the time of Romano the Court suggested that the Council could not freely decide on the body responsible for implement-ing EU law, and, as was noted previously (see section IV 5.4.3.2), the subsequent Treaty revisions only seem to have made this concern more acute by strengthen-ing the Commission’s position. Since Romano can be qualified as a prelude to the institutional balance, the section on the institutional balance will further deal with this question.

6.1.3.2 The ruling of the Court and its implications for the debate on agencification

As noted previously (see section IV 5.8.1.3), the Court’s ruling in Short- selling may be read as a simplification exercise, and one of its greatest victims was Romano. Although the Court confirmed Romano’s relevance for EU agencies, it equally noted that ‘it cannot be inferred from Romano that the delegation of powers to a body such as ESMA is governed by conditions other than those set out in Meroni’.631 From the prior analysis of the Romano ruling it should be clear that this statement of the Court is highly doubtful. Article 155 EEC, to which the Court referred in Romano, did not play a part in Meroni and lives on today in Articles 290 and 291 TFEU. The Court’s rejection of the UK’s plea on Romano then obviously had repercussions for the UK’s third plea. A critique on the Court’s ruling will therefore be further developed in a following section.

It suffices to note here that Romano cannot be completely subsumed under Meroni as the Court suggests. Furthermore, if Romano truly did not add anything to Meroni, it would be even more remarkable that neither the parties, nor AG Warner, nor the Court in Romano referred to Meroni.

629 Opinion of AG Jääskinen in Case C- 270/ 12, n 74, para 72.630 However, one could also argue that the Lisbon Treaty has not accommodated the concerns

of the Court in Meroni, since Article 263 TFEU still does not allow the acts of private bodies to be challenged.

631 Case C- 270/ 12, n 92, para 66.


6.2 Conclusion

While the present section has shown that Romano is a much more straightforward case than Meroni, the precise scope and effect of the ruling are unclear since they depend on how the notion of ‘acts having the force of law’ is understood and what the effect is of the metamorphosis of Article 155 EEC into Articles 290 and 291 TFEU.

The Court was given an opportunity to clarify the relevance of the Romano ruling for EU agencies in the post- Lisbon EU legal order in Short- selling and indeed seized it. The Court did so in a rather unconvincing way, as it seems doubtful that Romano may be completely subsumed under Meroni. Similarly, while the AG stressed the importance of the institutional balance (which was implicitly at issue in Romano), he disregarded Romano and the institutional balance in his further analysis of Articles 290 and 291 TFEU, opting instead for a democratic legitimacy test (cf. IV 7.2.4.1.2). However, the Court’s ruling does not necessarily mean Romano is dead and buried. Since the UK’s second and third pleas were linked, Romano could find a resurrection should the Court in the future refine its jurisprudence on Articles 290 and 291 TFEU.

7 Limits Flowing from the Institutional Balance

The institutional balance is often invoked by legal scholars when the possibility to delegate powers to bodies outside the Treaties is discussed. Unfortunately, the concept itself is quite elusive: it is not mentioned, let alone defined, in the Treaties and has instead been introduced by the Court in its case law.

In the present section, therefore, an attempt will be undertaken at clarifying the substance and function of the institutional balance in the EU legal order.

One important observation is the interconnectedness of the principle of insti-tutional balance with that of the separation of powers, even if it is unclear how both principles precisely relate to each other. Jacqué notes that the principle of institutional balance is no less to the Court than a substitute for the principle of separation of powers,632 a view shared by Conway633 and Huber.634 According to Barnett, the arrangements between the institutions in the EU are not governed by the separation of powers, but by the institutional balance.635 Prechal notes that the principle of institutional balance prevents the concentration of powers at European level and may therefore ‘be considered to be an equivalent to the doctrine of separation of powers, or rather “checks and balances” as they exist in

632 Jacqué, n 486, p 348.633 Gerard Conway, ‘Recovering a Separation of Powers in the European Union’, (2011) 17

ELJ 3, p 319.634 Peter Huber, ‘Das institutionelle gleichgewicht zwischen Rat und Europäischem Parlament

in der künftigen Verfassung für Europa’, (2003) 38 Europarecht 4, p 576.635 Hilaire Barnett, Constitutional & Administrative Law, New York, Routledge, 2011, p 210.

Limits Flowing from the Institutional Balance 259

national systems’.636 According to de Búrca, the institutional balance takes the place of the notion of separation of powers in the EU.637 Guillermin on the other hand claims both principles cannot simply be assimilated, notwithstanding the analogies they both share.638

Apart from this dissenting view, there seems to be a general consensus that both principles are strongly related. On the other hand and more recently, AG Trstenjak observed the following in the Audiolux case:

The institutional balance within the Community is not based on the principle of the separation of powers in the constitutional- law sense, but on a principle of the separation of functions, whereby the Community’s functions are intended to be exercised by the organs which are best placed to perform them under the Treaties. Unlike the principle of the separation of powers, which seeks partly to ensure that the individual is protected by moderating state power, the principle of the separation of functions is intended to ensure that the Community’s aims are effectively achieved.639

Still, how the institutional balance and the separation of powers precisely relate and whether the institutional balance actually serves its purpose as equivalent/ substitute to the separation of powers is seldom argued. Before tackling these questions, the venerable principle of separation of powers itself and whether it operates in the EU polity will be assessed.

7.1 Separation of powers

It should be clarified from the outset that the principle of separation of powers will only briefly be dealt with, so far as is necessary to gain a better understanding of the principle of institutional balance in the EU. As Conway points out, the former principle dates back as far as the writings of Aristotle, although the modern con-ception of the principle should be traced to Locke and Montesquieu and the more practically oriented writings of Madison.640 Conway also links the principle to other fundamental concepts such as democracy and the rule of law, claiming the normative attraction of the principle of a separation of powers essentially rests on its compatibility with the latter two.641 The principle is indeed intrinsically linked to the rule of law, as Gwyn showed in his study on the meaning of the separation of powers.642 According to Gwyn, although a number of distinct arguments may

636 Sacha Prechal, ‘Institutional Balance: A Fragile Principle with Uncertain Contents’, in Heukels, Blokker, and Brus (eds), The European Union after Amsterdam, The Hague, Kluwer Law International, 1998, p 280. Similarly Schmitter notes the functions of both principles may be com-pared. See Catherine Schmitter, ‘L’équilibre institutionnel’, in Philip, Barav, and Boutayeb (eds), Dictionnaire juridique des Communautés européennes, Paris, PUF, 1993, pp 473– 4.

637 de Búrca, n 273, p 58.638 Guy Guillermin, ‘Le principe de l’équilibre institutionnel dans la jurisprudence de la Cour

de justice des Communautés européennes’, (1992) 119 Journal du droit international 2, pp 319– 20.639 Opinion of AG Trstenjak in Case C- 101/ 08, Audiolux e.a., [2009] ECR I- 9823, para 104.640 Conway, n 633, pp 306– 7.641 Ibid, p 308.642 William Gwyn, The Meaning of the Separation of Powers: An Analysis of the Doctrine from Its

Origins to the Adoption of the United States Constitution, New Orleans, Tulane University, 1965, p 128.


be raised on behalf of the separation of powers, only the rule- of- law version of the principle must be based on a proper distinction between the different branches of government.643 Whether this means that the principle of separation of powers flows from the rule of law, as Fenwick and Phillipson suggest,644 or that the prin-ciple is an integral feature of the rule of law, as Allan claims,645 is another matter. Mapping the relations between the concepts of democracy, rule of law, and separ-ation of powers falls outside the scope of this excursion,646 but it will be assumed that these three concepts all form part of modern civilized polities so as to ensure a single goal, namely avoiding tyranny and arbitrary rule. Of course those three concepts are also equally elusive. Whether a given polity is a democracy or char-acterized by the rule of law or a separation of powers is always debatable.

As concerns the rule of law, Raz noted it is a political ideal which a legal system may lack or possess to a greater or lesser degree,647 and that it is essentially a nega-tive value, since the law inevitably creates a great danger of arbitrary power.648 The same observation applies to the separation of powers and democracy as well. Writing on a European Dyet, William Penn argued that chaos and strife could only be ended if men subordinated their liberty to the rules of their own making, incorporating themselves into society.649 But, arguing analogously to Raz, this in-corporation into society, by which the individual surrenders its absolute freedom, also entails new risks to his physical integrity. If people organize themselves into a society, those who hold power over that society and command its resources present a greater threat than they would have had that group remained individu-als outside of society. Power therefore needs to be split up and separated among different persons and branches, and those holding power should be (in)directly elected by the people and present themselves to the people at regular intervals.

Again, just like the rule of law, separation of powers and democracy are no more than abstract political ideals which polities may lack or possess to a greater or lesser degree. As regards the separation of powers, the issue is not so much whether the different powers can or cannot overlap— since the pure theory of

643 Other versions of the separation of powers are the efficiency version, the common interest version, the accountability version, and the balancing version. According to Gwyn, the rule of law is meaningless if there is no distinction between law- making and law- executing. See n 642.

644 Helen Fenwick and Gavin Phillipson, Text, Cases and Materials on Public Law and Human Rights, London, Taylor and Francis, 2011, p 131.

645 Trevor Allan, ‘Constitutional Justice and the Concept of Law’, in Huscroft (ed), Expounding the Constitution: Essays in Constitutional Theory, Cambridge NY, CUP, 2008, p 221.

646 According to Conway, ‘the rule of law can exist without democracy, but democracy needs the rule of law, for otherwise democratically established laws may be eviscerated at the stage of application by not being followed’. See Gerard Conway, The Limits of Legal Reasoning and the European Court of Justice, Cambridge, CUP, 2012, p 114. However, in this observation a separation of powers is also presupposed because the process of making the law is detached from the process of applying the law.

647 Joseph Raz, ‘The Rule of Law and Its Virtue’, (1977) 93 The Law Quarterly Review 2, pp 195– 8.

648 Ibid, p 206.649 William Penn, An Essay Towards the Present and Future Peace of Europe by the Establishment

of an European Dyet, Parliament, or Estates, London, Whitlock, 1696, pp 7– 10.


separation is quite untenable650— but when such overlap becomes too much so that one branch of government dominates another; or not enough so that the several branches of government operate on isolated islands of competence. A con-siderable number of authors also criticize the account of the ‘founding father’ of the modern principle, Montesquieu, himself. Some reject the idea of separation of powers as a constitutional principle.651 Allegedly Montesquieu also misunder-stood and had only undertaken a very superficial study of the British constitution when describing it in his De l’Esprit des Lois.652 Other scholars claim this critique is based on a misreading by the former authors,653 or that Montesquieu attempted to describe only an ideal type of a ‘constitution of liberty’.654

Montesquieu is also the subject of critique because of his understanding of the relation between the three powers. As regards the judicial power, Montesquieu claimed: ‘Des trois puissances dont nous avons parlé, celle de juger est en quelque façon nulle.’655 And when Montesquieu goes on to state that ‘Ces trois puissances devraient former un repos ou une inaction. Mais comme, par le mouvement nécessaire des choses, elles sont contraintes d’aller, elles seront forcées d’aller de concert’,656 he is actually talking about the two powers of the legislative branch (representing the commons and the nobles) and the executive, leaving out any important role for the judiciary. To the modern ear, this does not sound right. In fact, if there is any consensus on the principle today, it is precisely on the import-ance of the judicial branch and the need for its independence.657

Such a critique should not however be seen as too problematic, since Montesquieu did not himself lay down a doctrine of separation of powers but only contributed to its further development.658 Because the principle of separ-ation of powers is the starting point of an analysis of the institutional balance it was necessary to point to this debate but, fortunately, unnecessary to engage in it. As Vile concludes, Montesquieu’s writing is relevant because it upgraded the principle of separation of powers from an English theory to a universal criterion of constitutional government.659

To summarize, the ideal type of the principle could be described as fol-lows: public authority may be dissected into three different functions. These are

650 Eric Barendt, ‘Separation of Powers and Constitutional Government’, (1995) Public Law, Winter, p 608.

651 William Jennings according to Iain Stewart, ‘Men of Class: Aristotle, Montesquieu and Dicey on “Separation of Powers” and “Rule of Law”’, (2004) 4 Macquarie Law Journal, p 189.

652 Eg Albert Dicey, Introduction to the Study of the Law of the Constitution, London, MacMillan, 1939, p 338; William Holdsworth, A History of English Law, London, Methuen, 1972, Vol. X, pp 718– 24; Laurence Claus, ‘Montesquieu’s Mistakes and the True Meaning of Separation’, (2005) 25 Oxford Journal of Legal Studies 3.

653 Stewart, n 651, pp 187– 223.654 Mirkine- Guetzévitch according to Maurice Vile, Constitutionalism and the Separation of

Powers, Indianapolis, Liberty Fund, 1998, p 93.655 Charles- Louis De Montesquieu, De l’esprit des lois, Paris, Gallimard, 1995, Vol 1, p 333.656 Ibid, p 339.657 Fenwick and Phillipson, n 644, p 128.658 Vile, n 654, p 93.659 Ibid, p 106.


the legislative power, enacting norms of general application; the executive power, enforcing these norms; and the judicial power, applying the law in conflicts. This functional separation of powers should then be linked to an institutional (or org-anic) and personal separation of powers, meaning the different powers should be vested in different and distinct institutions staffed by different people.660 Because the principle is a means to an end and not the end in itself, a pure separation of powers should be rejected in favour of a partial one which results in a system char-acterized by checks and balances.661 What form these checks and balances should take is a question of practice rather than theory.

7.1.1 The separation of powers in the EUDriessen notes that there are different views on the appropriateness of analysing the EU in terms of separation of powers and that this is partially explained by the lack of academic consensus on the meaning of the principle.662 In his ana-lysis, however, he does not make a clear distinction between two distinct ques-tions: whether the principle is relevant to the EU and whether the EU pays re-spect to the principle. Obviously, the latter should only be answered if the first is answered affirmatively.

7.1.1.1 The principle’s relevance for the EUThe question of the principle’s relevance for the EU may be easily resolved, but was raised nonetheless and seems to flow from the debate on the nature of the EU itself. Because the principle of separation of powers has initially been developed for and within national polities, Achterberg questioned whether it was at all sens-ible to try to apply it to the EU.663 Ipsen claims the principle is no longer even adequate to structure the exercise of public power in modern states and should for this reason alone be rejected as a guiding principle for the EU.664

Of course, looking at the ratio of the principle, the irrelevance of such objec-tions becomes clear. Arguing that the principle is only relevant to national states would be tantamount to saying arbitrary rule is only a risk when powers are exer-cised by the national state. Furthermore, the ideal type cited above simply refers

660 Aalt Heringa and Philipp Kliver, Constitutions Compared: An Introduction to Comparative Constitutional Law, Antwerpen, Intersentia, 2007, p 7.

661 As Everson notes, protection against abuse of power (or arbitrary rule) may be understood in two ways. One can divide the different powers so as to ensure abuse of powers become impossible but one can also aim for a separate and ‘ideal’ dispensation of powers. Here the checks and balances come into play. See Michelle Everson, ‘European Agencies: Barely Legal?’, in Everson, Monda, and Vos (eds), European Agencies in between Institutions and Member States, Alphen aan de Rijn, Kluwer Law International, 2014, p 51.

662 Bart Driessen, Interinstitutional Convention as Checks and Balances in EU Law, Leuven, KUL, 2006, PhD Thesis, p 37.

663 Norbert Achterberg, ‘Rezension: Die Gewaltenteilung in den Europäischen Gemeinschaften’, (1968) 3 Europarecht 2, pp 240– 5.

664 Ipsen, n 120, p 318. Similarly, see Renaud Dehousse, ‘Comparing National and EC Law: The Problem of the Level of Analysis’, (1994) 42 American Journal of Comparative Law 4, pp 776– 7.


to public authority, regardless of at which level it is exercised. Ipsen’s observations on the novelty of European integration are not disputed here, but his deductions are. No matter how revolutionary the EU polity may be, it remains a hierarchical polity in which the powers to direct it are (partially) centralized, creating a risk of power abuse and arbitrary rule.

Lenaerts and Desomer claim that the Court of Justice has rejected the prin-ciple as unknown in the EU legal system in the joined cases 188 to 190/ 80,665 something which Nicolaysen also hinted at. According to the latter, the Court’s ruling ‘again confirms the originality of the system of Community functions since it does not allow itself to be fitted in the classic scheme of separation of powers yet it follows its own structural principles’.666 However, in that case the Court merely rejected the UK’s claim that all original law- making power is vested in the Council, clarifying that both the Commission and Council can lay down measures of general application, depending on the specific Treaty provisions in question.667 This case does show however that powers in the EU have not been functionally and institutionally separated along the same lines. The Court thus ruled that there is no single institution which holds the legis-lative function. Otherwise there would be some tension in the Court’s juris-prudence, since it ruled in DEB that ‘EU law does not preclude a Member State from simultaneously exercising legislative, administrative and judicial functions, provided that those functions are exercised in compliance with the principle of the separation of powers which characterizes the operation of the rule of law’.668 Paradoxically, EU law would require national polities to respect the separation of powers in order for them to qualify as a Rechtsstaat, but it would not require the EU to adopt a separation of powers in order to qualify as a Rechtsgemeinschaft.669

The confusion over the question of whether the separation of powers is relevant to the EU may probably be traced back to two key assumptions which a number of authors hold. Pescatore’s conclusion on the relevance of the principle for the EU is illustrative in this regard: ‘la doctrine “tripartite” de la séparation des pouvoirs n’est pas un principe d’explication valable pour un ensemble transnational tel que les Communautés européennes.’670 Thus, Pescatore assumes that the principle of separation of powers dictates an organic tripartite and he sees the principle as an explanatory principle. Neither assumption seems fully convincing. Firstly, the principle merely dictates that the three different functional powers be separated,

665 Koen Lenaerts and Marlies Desomer, ‘Towards a Hierarchy of Legal Acts in the European Union? Simplification of Legal Instruments and Procedures’, (2005) 11 ELJ 6, p 764; Joined cases 188 to 190/ 80, France, Italy and UK v. Commission, [1982] ECR 2545.

666 Gert Nicolaysen, ‘Rechtsprechung: Öffentliche Unternehmen, Transparenz der finanziel-len Beziehungen zum Staat’, (1983) 18 Europarecht 1, p 61. Translated from German by the author.

667 Joined cases 188 to 190/ 80, n 665, para 6.668 Case C- 279/ 09, DEB Deutsche Energiehandels- und Beratungsgesellschaft GmbH, [2010] ECR

I- 13849, para 58.669 See Case 294/ 83, Parti écologiste ‘Les Verts’ v. Parliament, [1986] ECR 1339, para 23.670 Pierre Pescatore, ‘L’Executif Communautaire: Justification du Quadripartisme institue

par les Traités de Paris et de Rome’, (1978) 14 CDE 4, p 388.


not that they be separated between three different institutions. Secondly, the prin-ciple is not explanatory but prescriptive in nature. Whether or not one can make sense of the EU, looking at it from a separation- of- powers perspective has no bearing on the relevance it has for the EU.

7.1.1.2 Identifying a separation of powers in the EUThe principle is therefore relevant to the EU, notwithstanding that it is a new type of polity. However, whether the principle is observed is a delicate question which may only be answered by evaluating the whole picture of power- exercise as it re-sults from the entire system of checks and balances. Such an exercise would fall outside the scope of this study and will therefore not be attempted.

Lenaerts notes that tracing the principle in the EU is an even more subtle undertaking than it is in national polities, since in the EU an organic under-standing of the principle is not practicable.671 He notes that it is a more fruitful exercise to look for a functional separation of powers in the EU and concludes that ‘[i] t can hardly be doubted that the principle of separation of powers, as defined in this functional perspective, is a structural feature underlying the [Union] legal order’.672 The identification of different functions (ie legislative, executive, judica-tive) indeed underlies the principle’s descriptive part, but a theoretical separation of functions in itself is quite meaningless. The principle’s normative dimension consists of the prescription that those different functions should then also be vested in different organs and persons, and only by linking the descriptive and normative dimension does the principle become meaningful.

Still it is correct to claim that the system of the Treaties does not allow for the easy identification of the three powers.673 The problem is furthermore exacerbated by the Treaty principles of conferred powers and subsidiarity. The legislative, ex-ecutive, and judicial functions in the integrated EU order are therefore partially exercised at EU level and partially exercised at national level. Focusing on the EU itself, Driessen notes that the judicial function is unproblematic and can be said to rest wholly with the Union’s courts.674 This is not the case for the legislative and executive functions. According to Driessen, the problems may be summarized as follows: (i) lack of hierarchy of norms, (ii) a problematic attribution of functions, (iii) fragmented execution, (iv) comitology and agencies.

Driessen notes that a hierarchy between legislative and executive norms, which the pre- Lisbon EU Treaties lacked, is necessary. Yet how a hierarchy of norms con-tributes to the principle of separation of powers is not immediately clear. Referring back to the triad of rule of law, democracy, and separation of powers, the concept of hierarchy of norms may be situated closer to the former two. A hierarchy of

671 Koen Lenaerts, ‘Some Reflections on the Separation of Powers in the European Community’, (1991) 28 CMLRev 1, pp 12– 13.

672 Ibid, p 14.673 Driessen, n 662, p 40.674 In this regard Dehousse rightly points out, referring to Lecourt, that the national judges may

also be seen as forming part of the EU judiciary: see Dehousse, n 664, p 776.


norms flows from the rule of law, ensuring law- makers themselves are bound by higher laws. Similarly, a hierarchy of norms which ranks parliamentary acts near the top strengthens democratic government. This is what Driessen himself seems to refer to when he posits that in the logic of a gesetzmäßigen Verwaltung a func-tional separation of powers necessitates the subordination of executive acts to le-gislative acts. A hierarchy of norms further also seems to depend more on the separation of powers than vice versa, as it would be hard to effectively enforce a hierarchy of norms without an independent judicial branch.

Whatever the case, the Treaty of Lisbon seems to have introduced such a hier-archy between the legislative, delegated, and implementing acts in Articles 289(3), 290, and 291 TFEU. This creates a suggestion of hierarchy along the order in which the types of acts are cited.675 But this would be hard to justify based on a pure separation- of- powers perspective. Instead the reason should be found in the greater democratic legitimacy of a delegated act, given the scrutiny by the Parliament (and the Council), as compared with that of an implementing act.676

Secondly, Driessen argued that the attribution of functions under the Treaties is wholly different from the classical attribution to three different branches. However, this does not mean that the EU does not respect the principle of separ-ation of powers. In the EU the legislative power is more fragmented than it is in national polities, where the fit between functional and organic separation is more elegant. The plethora of different legislative procedures cited by Driessen should not be exaggerated either. In reality, the ordinary legislative procedure is the dom-inant procedure. The overall resulting picture then is a legislator composed of three institutions, two of which form a bicameral Parliament, whereas the man-datory involvement of the third is indeed a genuine curiosity from a comparative perspective. Looking at the decision- making process in the EU from a broader perspective, this curiosity may be seen as a balancing element itself.677

As regards the fragmented executive at EU level, it is again unclear why this would pose a problem in the light of the separation of powers. Driessen notes that the Commission might appear to be the executive of the Union, but that this can be challenged from three sides. Firstly, the Parliament does not wish to restrain itself to a purely legislative or controlling role; secondly, the Member States play an important role in the execution of EU law; lastly, the Council is also an im-portant part of the EU executive.678 Concerning the first objection, this cannot

675 Lenaerts and Desomer, n 665, p 763; Allan Rosas and Lorna Armati, EU Constitutional Law, Oxford, Hart Publishing, 2010, p 51.

676 Precisely because of their lower democratic legitimacy as compared to the Commission, AG Jääskinen found it impossible in Short- selling to grant EU agencies the power to adopt delegated acts: see section IV 7.2.4.1.2.

677 Merijn Chamon, ‘Upholding the ‘Community Method’: Limits to the Commission’s Power to Withdraw Legislative Proposals— Case C- 409/ 13’, (2015) 40 ELRev 5, p 900.

678 What may be problematic is the Council, in its legislative capacity, creating special executive powers for itself following secondary legal bases. This has already come before the Court: see Case C- 133/ 06, Parliament v. Council, [2008] ECR I- 3189, and the pending case C- 363/ 14, Parliament v. Council, OJ 2014 C 329/ 10.


in itself be problematic. After all, the principle of separation of powers does not presuppose each different branch will control its lust for power itself; quite the contrary, the principle prescribes that through a separation of powers the other two branches will keep a check on that first one.

Concerning the objection that the proper executive of the EU is a triad com-posed of the Commission, Council, and Member States, it seems this critique assumes that a separation of powers is only possible when the executive function is exercised by a ‘unitary executive’. From a comparative perspective, this claim is already difficult. While it does for instance appear that the US Constitution establishes a unitary executive, this issue is still hotly debated in the US,679 even if Printz v. United States may again have strengthened the proponents of a unitary ex-ecutive theory.680 A unitary executive then is simply a practical choice in putting a system of checks and balances into practice, but not a conditio sine qua non in res-pecting the principle of separation of powers. In the EU Treaties a choice has also been made for an executive federalism inspired by Germany’s Vollzugsföderalismus. Not only is a unitary executive not imperative, but also, in German doctrine, the choice for an (at first sight) antipode— Vollzugsföderalismus— is even regarded as complementary with and completing the principle of separation of powers.681

The last two objections formulated by Driessen, comitology and agencifica-tion, can actually be subsumed under the fragmentation of the executive. The im-portant difference then between comitology and agencification is that the former (since the SEA) finds its origins and legitimacy in the Treaties, whereas the latter does not.

To conclude on the question of whether there is a separation of powers in the EU, Lenaerts’ original conclusion should be endorsed, but not simply because there is a functional separation of powers. After all, that is always the case, whether a polity wishes to acknowledge it organically or not. In the EU, the legislative and executive— not so much the judicial— powers are more fragmented than under

679 See Evan Caminker, ‘Unitary Executive and State Administration of Federal Law’, (1997) 45 University of Kansas Law Review 4, pp 1075– 112.

680 Printz v. United States, 521 U.S. 898. In Printz Justice Scalia, writing for the majority and referring to Madison’s double security argument (discussed later), held that the states could not be used as instruments of federal governance, further arguing that the contested arrangement ‘would also have an effect upon […] the separation and equilibration of powers between the three branches of the Federal Government itself. The Constitution does not leave to speculation who is to admin-ister the laws enacted by Congress; the President, it says, “shall take Care that the Laws be faithfully executed.” […] The Brady Act effectively transfers this responsibility to thousands of [Officers] in the 50 States, who are left to implement the program without meaningful Presidential control (if indeed meaningful Presidential control is possible without the power to appoint and remove)’. Bybee remarked how this federal case promoted the notion of a unitary executive, also affecting the position of the federal independent agencies. See Jay Bybee, ‘Printz, the Unitary Executive, and the Fire in the Trash Can: Has Justice Scalia Picked the Court’s Pocket’, (2001) 77 Notre Dame Law Review 1, pp 269– 88.

681 Hans- Jochen Vogel, ‘Die bundesstaatliche Ordnung des Grundgesetzes’, in Benda and Hesse (eds), Handbuch des Verfassungsrechts der Bundesrepublik Deutschland, Berlin, de Gruyter, 1994, p 1051. Without referring to Madison, Vogels repeats the latter’s double security argu-ment: see The Federalist No 51.


a tripartite separation of powers. The EU is characterized by an intricate system with numerous checks and balances which introduce the separation of powers in the EU, even if the EU system is not based as such on the principle. Again this is because this principle is not so much explanatory but rather prescriptive. Even if one cannot explain the institutional features of a polity by referring to the separ-ation of powers, one can still scrutinize it in the light of that principle.

This conclusion can also be criticized. After all, finding a separation of powers in the EU has been possible by hollowing out our more conventional notion of the principle. By focusing on the ratio of the principle it was possible to dismiss certain elements of the traditional conception as not strictly necessary. What is left is the idea that in exercising public authority a polity is exercising different functions and that these functions may not all be conferred on the same institu-tions and persons, but should be fragmented over at least as many actors as there are functions, where those actors should be relatively independent of each other.

7.1.1.3 Institutional balance as separation of powers?The conclusion of the previous section should again draw attention to the prin-ciple of institutional balance and how it precisely relates to the principle of separ-ation of powers. Since the latter does not depend on the former to be of relevance for EU constitutional law, it cannot be said that the institutional balance is to the EU polity what the separation of powers is to national polities.682 Above it was noted that only Guillermin is keen on keeping both concepts apart even if both principles do serve a similar function, that is, that of guaranteeing certain rights and competences to all of the actors involved in the polity concerned.683 Still, Guillermin is not entirely clear why the two principles cannot be assimilated.

To answer this question, it should be noted that the principle of separation of powers is a political– philosophical principle and is seen in liberal democracies as a fundamental principle to safeguard (political) freedom. It is an abstract theoret-ical notion against which different polities can be gauged. This is of course totally different for a principle such as the institutional balance. There is no abstract institutional balance against which the institutional balance in the EU can be gauged. The principle was conjured by the Court in Luxemburg and ‘is laid down in the Treaties’. Because the Treaties themselves are the standard, it is impossible to measure them up to any abstract notion of ‘institutional balance’. Both prin-ciples thus operate at different levels: whereas the separation of powers is relevant to evaluate the legal architecture and political functioning of the EU, the insti-tutional balance is only useful in evaluating the political functioning of the EU.

In this regard it is interesting to note that the argument proposed by a number of authors to the effect that the institutional balance is not an obstacle to further agencification, but that the latter actually strengthens the institutional balance

682 It seems that at the most it could be said that the institutional balance is the expression (rather than the equivalent or substitute) of the separation of powers in the EU.

683 Guillermin, n 638, p 344.


(cf. section IV 5.4.2.1.1), seems to be based on the efficiency variant684 of the sep-aration of powers: agencies allow the ‘proper’ institutions to work more efficiently and effectively.685 Because this is undeniably so, agencification could indeed be justifiable and even laudable under a separation- of- powers perspective (efficiency variant), but this does not necessarily mean agencification also strengthens the institutional balance. Another example in the executive sphere may be found in Blumann’s observation on the Parliament’s increasing say in comitology: ‘Cette évolution soulignait d’ailleurs assez bien le hiatus, pour ne pas dire la contradic-tion, existant entre le principe d’équilibre institutionnel et celui de séparation des pouvoirs.’686 This observation can only be supported since it indeed made perfect sense for the Parliament to demand equality with the Council in comitology mat-ters under the institutional balance, whereas such an intrusion— which included a parliamentary veto power over the Commission’s executive decision- making— by the (co- )legislator is more questionable from the perspective of the separation of powers.

Now that it has been determined that both principles are relevant to the EU, independently from each other, it is necessary to look more closely into the notion of institutional balance.

7.2 The EU’s institutional balance

The ‘notion’ of institutional balance refers to ‘a system for distributing powers among the different Community institutions, assigning to each institution its own role in the institutional structure of the Community and the accomplish-ment of the tasks entrusted to the Community’.687 The notion, often dubbed a principle of EU law, has not been spared academic critique. Indeed it has even been raised that in fact it is no principle at all. According to Tridimas, ‘a principle must be judged on the basis of two parameters: the intrinsic value of the right that it embodies, and how well it structures the judicial inquiry’.688 These two param-eters signify that if the institutional balance is to be a genuine principle of EU law, it should have a sufficiently clear meaning which guides the Court and helps it in solving legal issues brought before it.

According to von Bogdandy, the institutional balance guarantees the ‘strings of accountability’ and the observance of procedural rules without it being a genuinely clear principle (yet), since the provisions on the competences and the cooperation between the different institutions are too complex and divergent.689 Chalmers does not claim as forcefully that the institutional balance is not a principle but he

684 See n 643.685 Although these arguments would also seem to fit well with AG Trstenjak’s view on the insti-

tutional balance as promoting effective action at EU level (cf. n 639).686 Blumann, n 543, p 37.687 Case C- 316/ 91, Parliament v. Council, [1994] ECR I- 625, para 11.688 Tridimas, n 169, p 2.689 Armin von Bogdandy, Europäische Verfassungsrecht, Berlin, Heidelberg, 2003, p 199.


is of the opinion that the Court of Justice should take on a more assertive stance in interinstitutional conflicts and be more receptive when the institutional balance is raised in preliminary procedures, if it is to be deserving of the title of legal prin-ciple rather than ‘a bat for EU institutions to swat each other with’.690 Chalmers’ critique stands in contrast with AG Trstenjak’s view in Audiolux, in which she observed that ‘the Court established […] the notion of “institutional balance” […] and accorded it the role of a normative, actionable formal principle’.691 However, the analysis of the Court’s case law below does not provide much support for des-cribing the institutional balance as an actionable principle; instead, it confirms Chalmers’ observation.

According to de Witte, given the Treaties’ detail, there is no place for a (judge- made) principle governing the horizontal relations between the institutions.692 Prechal largely agrees with this but still sees an important role for the institutional balance as a gap- filling principle, since the rules governing the position of the sep-arate institutions and the relations between them are not always clear.693 These conflicting views may be reconciled by noting that the Treaties are indeed detailed in relation to the legislative process, but not so much in the executive sphere. In addition, given the fast pace of successive Treaty revisions, there does seem to be some potential in the gap- filling function as proposed by Prechal.

Finally, Lenaerts and Verhoeven do make the case for the institutional balance as a principle of EU governance, although they see more potential in the institu-tional balance as a political rather than a legal principle. That political principle ‘requires the makers of the European constitution to shape institutions and the interactions between them in such a manner that each interest and constituency present in the Union is duly represented and co- operates with others in the frame of an institutionalized debate geared towards the formulation of the common good’.694 This again shows the difficult relation between the notion of institu-tional balance and the principle of separation of powers, since these authors pro-pose an abstract balance, which the Treaty drafters themselves should take into account when drafting new Treaties.695

While this political ‘institutional balance’ has a certain appeal, Lenaerts and Verhoeven do not clarify how it would work in practice: how can a self- standing political principle of institutional balance inform the Treaty authors? Instead, the question as to which societal interest should be attributed what weight in the decision- making process seems to relate more to questions of democratic

690 Damian Chalmers, ‘Justifying Institutional Accommodation’, (2008) 33 ELRev. 4, p 456. Similarly, see Roland Bieber, ‘The Settlement of Institutional Conflicts on the Basis of Article 4 of the EEC Treaty’, (1984) 21 CMLRev 3, p 509.

691 Opinion of AG Trstenjak in Case C- 101/ 08, n 639, para 105 (emphasis added).692 Bruno de Witte, ‘Institutional Principles: A Special Category of General Principles of EC

Law’, in Bernitz and Nergelius (eds), General Principles of European Community Law, The Hague, Kluwer Law International, 2000, pp 150– 1. However, de Witte does make an exception for the Court’s Chernobyl ruling.

693 Prechal, n 636, pp 277– 8.694 Lenaerts and Verhoeven, n 284, p 47.695 Ibid, pp 47– 8. Similarly, see Michel, n 3, pp 67– 9.


legitimacy and interest- representation.696 In the following parts, any reference to the institutional balance then is a reference to the legal notion or principle of in-stitutional balance.

7.2.1 A static or dynamic institutional balance?This last issue touches upon another vital question regarding the institutional bal-ance: whether it is a static or a dynamic concept (cf. section IV 5.4.2.1.2). Nord observed that the institutional balance continually modifies itself and that the division of powers between the institutions has not always corresponded to what is laid down in the Treaties,697 but this should be understood as confirming that institutional practice may be very dynamic rather than confirming the institu-tional balance itself is dynamic.

Is there a possible second institutional balance, then? The argument by Lenaerts and Verhoeven on a political institutional balance implies the existence of a dy-namic institutional balance: the existence of an abstract political institutional balance allows the legal institutional balance to be interpreted dynamically. It is this very same view which underlies the assumption of a number of authors that the institutional balance can be rebalanced to ‘absorb’ the shock of agencification (cf. section IV 5.4.2.1.2). Frank observes: ‘In general, one can safely argue that it is a dynamic notion. The practice of today is to apply the institutional balance strictly, notwithstanding the fact that it seems controversial to hang on to a con-cept which is static. Moreover, this rigid interpretation blocks the evolution of the Community.’698 However, the latter is only partially true, since the Union may still evolve as long as it is within the limits set out by the Treaties, and even then it can still evolve if the Treaties are amended. Frank’s argument raises the unad-dressed question of why the EU should be permitted to evolve without recourse to the proper procedure for such evolution, that is, a Treaty change.

Authors such as Geradin and Petit and Jacqué note that the Court in Chernobyl used the institutional balance not as a static principle but dynamically.699 Here, two distinct issues seem to be confused: whether the notion or principle is static/

696 Following Chernobyl, Hilf noted that ‘Der Kern des Urteils zielt also auf den gerichtlichen Schutz des “institutionellen Gleichgewichts”, das sich in der Tat seit der Abfassung der Verträge nachhaltig zugunsten des Europäischen Parlaments verschoben hat. Zwar ist nach der Meßlatte des Demokratieprinzips im Hinblick auf die Verteilung der Gesetzgebungsbefugnisse bei weitem noch kein “institutionelles Gleichgewicht” erreicht’. See Meinhard Hilf, ‘Das Klagerecht des Europäischen Parlaments im Organstreit’, (1990) 25 Europarecht 3, p 278. Driessen correctly noted that such an institutional balance would be a political rather than a legal notion. See Bart Driessen, Interinstitutional Conventions in EU Law, London, Cameron May, 2007, p 45.

697 Hans- Robert Nord, ‘Quelques réflexions sur l’évolution institutionnelle des Communautés européennes’, in Teitgen (ed), Mélanges Fernand Dehousse: La Construction Européenne, Paris, Nathan, 1979, p 142.

698 Frank, n 273, p 128.699 Geradin and Petit, n 113, p 149; Jacqué, n 486, p 386. See also Savino and Zinzani: Mario

Savino, ‘The Constitutional Legitimacy of the EU Committees’, (2005) Les Cahiers européens de Sciences Po 2005/ 3, p 8; Zinzani, n 280, p 54.


dynamic, and what its effects are when applied. A static notion or principle could well have dynamic effects. If a static institutional balance could only have static effects, it would indeed be a useless principle. This distinction is also apparent in AG Van Gerven’s Opinion and the Court’s ruling in Chernobyl (cf. section IV 7.2.3.1): the AG did not want to ‘adjust’ the institutional balance as requested by the Parliament, whereas the Court in its ruling found that it had to ‘enforce’ the institutional balance.

Presenting the institutional balance as dynamic relates in a difficult way with the Court’s observation that the institutional balance is laid down in the Treaties. Tournepiche seems to suggest that the balance is established by the Treaties but may subsequently evolve on its own, within the limits ultimately set by the Court, even if the balance may not be modified: ‘la pratique institutionnelle […] peut […] faire évoluer cet équilibre qui ne peut en principe être modifié que par une révision des traités.’700 For the sake of clarity, however, it would seem preferable here to speak of institutional practice which may evolve, whereas the institutional balance may not be modified, unless by a revision of the Treaties.701

Further, few authors actually reason why the institutional balance should be a dynamic concept, although Schuppert has probably expressed the underlying reasoning when he observed:

While the elements of stability and dynamics, rigidity and flexibility find themselves in a balanced relationship in the national constitution, the European Community’s con-stitution is necessarily dynamic and geared towards evolution, since the objective of the Community is the progressing European integration, which further implies a process of evolution in the constitution which guides this integration.702

Integration is indeed an ongoing process, and therefore needs a sufficiently dy-namic constitutional charter, but this argument does not solve the question of whether this makes the institutional balance dynamic. In any case, what seems clear is that if the principle of institutional balance is a dynamic principle, it risks becoming an unworkable legal principle for the EU judge. Its ‘dynamism’ would make it an ‘unguided missile’, and the same goes for its use in the present study. A ‘rigid’ institutional balance as a legal principle on the other hand does

700 Anne- Marie Tournepiche, Les accords interinstitutionnels dans l’Union européenne, Bruxelles, Bruylant, 2011, p 479.

701 The same remark may be made in relation to Smith and Kelemen who note that the devel-opment of the institutional balance results from formal (Treaty) and informal (practice) changes. See Mitchell Smith and Daniel Kelemen, ‘The Institutional Balance: Formal and Informal Change’, July 1997, Centre for European Policy Studies Working Document 11, pp 1– 2.

702 Gunnar Folke Schuppert, ‘Zur Staatswerdung Europas: Überlegungen zu Bedingungsfaktoren und Perspektiven der europäischen Verfassungsentwicklung’, (1990) 1 Staatswissenschaften und Staatspraxis: rechts- , wirtschafts- und sozialwissenschaftliche Beiträge zum staatlichen Handeln, p 37. Translated from German by the author. Echoing this view, see Everson et al, n 290, p 28. Similarly, Constantinesco notes that ‘[l] e caractère dynamique de l’ intégration euro-péenne soumet l’ équilibre institutionnel établi par les traités à évolution.’ See Vlad Constantinesco, ‘Les institutions communautaires: Présentation génerale, Fasc. 200’, in Loy (ed), Traité de droit européen, Paris, Ed. Techniques, 1989, p 18.


lend itself as a tool to scrutinize institutional developments and practices such as agencification.

7.2.2 The institutional balance as a legal principleLenaerts and Verhoeven derive three rules from the legal principle of institutional balance: (i) each institution should enjoy a sufficient independence in order to ex-ercise its powers; (ii) institutions should not unconditionally assign their powers to other institutions; and (iii) institutions may not in the exercise of their own powers encroach on the powers and prerogatives of other institutions.703

So far, Lenaerts and Verhoeven have been the only ones to try and map out the content of the institutional balance in detail, making it unclear whether the three rules they deduced are generally accepted. Some, however, have implicitly argued for a fourth rule to be added to those of Lenaerts and Verhoeven, arguing that there is also a vertical dimension to the institutional balance.704

According to Vos, the institutional balance enshrines a strict division of powers between the institutions which can be seen as a reflection of the Member States’ concern that the integrity of their powers be maintained. The notion of institu-tional balance should then be defined widely, encompassing the vertical relation between the Union and the Member States. According to Vos, this would explain why Member States need a say not only over the legislative process at EU level, but also over the implementation of EU law.705 However, since the institutional bal-ance would be a judge- made principle, such a function is doubtful. Even if the ref-erence to the institutional balance in the Protocol on the application of the prin-ciples of subsidiarity and proportionality added to the Amsterdam Treaty could have been interpreted as supporting such a vertical dimension of the notion,706 it should be noted that this reference has been deleted in the Protocol added to the Lisbon Treaty.

To be sure, Vos is not alone in her claim. Fischer- Appelt for instance notes that ‘already early on the ECJ worked out the principle of institutional bal-ance to scrutinise both the legality or illegality of transfers of power between the Community organs (horizontally) or to scrutinise the infringements in the attribution of competences between the Communities and the Member States (vertically)’.707 However, the vertical dimension of the division of powers within the EU legal order is governed by the principle of attribution of powers, making it unclear why recourse to the institutional balance to solve these ver-tical power- delimitation problems would be necessary. It could of course be argued that the notion of institutional balance is the horizontal expression of

703 Lenaerts and Verhoeven, n 284, pp 44– 5. 704 See also Reuter, n 263.705 Vos, n 542, p 223.706 Paragraph 2 of the Protocol provided: ‘The application of the principles of subsidiarity and

proportionality shall respect the general provisions and the objectives of the Treaty, particularly as regards the maintaining in full of the acquis communautaire and the institutional balance; […].’

707 Fischer- Appelt, n 8, p 169. Translated from German by the author.


the principle of attribution of powers,708 but this would not sit well with the Court’s apparent emphasis on that notion being a principle in its own right (cf. section IV 7.2.3.1).709

In the following section the case law of the Court of Justice in which the in-stitutional balance figured will be analysed to gain a better understanding of the critique on the status of the institutional balance as a principle of EU law, and to verify whether a fourth rule, on the alleged vertical dimension of the principle, should be added to the three rules of Lenaerts and Verhoeven.

7.2.3 The institutional balance in the case law of the Court of JusticeIn the following sections this case law will be analysed and briefly commented upon. To bring some structure to this section, the three rules of Lenaerts and Verhoeven and the fourth rule of Vos will be used as subheadings.

7.2.3.1 The three rules of Lenaerts and VerhoevenLenaerts and Verhoeven do not make clear in which category they list Chernobyl, although that case made the institutional balance famous. Because the first rule emphasizes the autonomy of the institutions, the Chernobyl case should be listed under this rule. The precursor to Chernobyl was Comitology. In that case, the Parliament tried to rely on the institutional balance to acquire locus standi as plaintiff under the then Article 173 EEC (now Article 263 TFEU), which the Court rejected. In reply to the Parliament’s claim that it would not be able to defend its prerogatives without the power to bring action for annulments, the Court declared that other means for review were still available.710 In Chernobyl, two years later, the Court reconsidered its earlier Comitology judgment and fol-lowed the Parliament’s original reasoning. The Court then also further clarified the notion of institutional balance, stating:

Th[e] prerogatives [of the European Parliament] are one of the elements of the institu-tional balance created by the Treaties. The Treaties set up a system for distributing powers among the different Community institutions, assigning to each institution its own role in the institutional structure of the Community and the accomplishment of the tasks entrusted to the Community.

Observance of the institutional balance means that each of the institutions must exer-cise its powers with due regard for the powers of the other institutions. It also requires that it should be possible to penalize any breach of that rule which may occur.

The Court […] must therefore be able to maintain the institutional balance and, con-sequently, review the observance of the Parliament’s prerogatives when called upon to do

708 For such an argument, see Priebe, n 27, pp 75– 8.709 According to Constantinesco, the institutional balance results from the principle of attrib-

uted competences (to the EU) and powers (to the institutions) and is a principle in itself. See Vlad Constantinesco, ‘L’équilibre institutionnel dans la Constitution de l’Union européenne’, in Apogée (ed), Le droit de l’Union européenne en principes: liber amicorum en l’ honneur de Jean Raux, Rennes, Apogée, 2006, p 485.

710 Case 302/ 87, Parliament v. Council, [1988] ECR 5615, para 27.


so by the Parliament, by means of a legal remedy which is suited to the purpose which the Parliament seeks to achieve.711

Interestingly enough, the Court justified this contra legem (or praeter legem)712 in-terpretation of the Treaties by pointing to the need to maintain the institutional balance. This should be juxtaposed with the opinion of AG Van Gerven in the same case. In his opinion, AG Van Gerven paid special attention to working out a solution while remaining as much as possible within the lines set by the Court in Comitology. To this end AG Van Gerven made a distinction between adjusting or re- establishing the institutional balance and upholding an adequate and coherent system of legal protection. According to the Advocate General, Comitology consti-tuted a ‘refusal by the Court to accede to the Parliament’s request to alter the insti-tutional balance in its favour’ because such a move did not belong to the province of the courts. On the other hand, ensuring the Parliament enjoys effective legal protection was a task for the Court.713 AG Van Gerven thereby cleverly succeeded in presenting a workable solution without having to get into the repercussions of his solution for the institutional balance.

The Court reached more or less the same result as the AG but did not rely so much on the narrowed- down argument of effective legal protection; rather, it em-phasized the fundamental interest in maintaining and observing the institutional balance. Thus, granting the Parliament (limited) locus standi against the wording of the Treaties did not upset the institutional balance— a question the AG was keen on avoiding— but was actually necessary to uphold the institutional balance.

In the Lamberts case on appeal before the Court of Justice, the European Ombudsman and the European Parliament claimed that the CFI, in declaring an action for damages in relation to the conduct of the Ombudsman admissible, had jeopardized the institutional equilibrium established around the Ombudsman and called into question the Ombudsman’s independence under Article 228(3) TFEU.714 The careless wording of this argument, in that there is an institutional balance established around the Ombudsman, appears to be another indication that the institutions do not hesitate to gratuitously use the institutional balance to further their interests in Court. The Court, however, ruled that the Ombudsman’s independence was not at stake, since an action for damages concerns not the personal liability of the Ombudsman but the liability of the Union.715

Similarly, in the FIAMM case on appeal before the Court of Justice, the Council argued that the non- contractual liability for lawful acts should be rejected since it ‘would limit both the freedom of choice inherent in the Commission’s right of initiative and the legislature’s discretion, calling into question the separation of

711 Case 70/ 88, n 356, paras 21– 3.712 See AG Mengozzi’s juxtaposition of reading Chernobyl as praeter legem versus a contra legem

interpretation of the Treaties. Opinion of AG Mengozzi in Case C- 354/ 04 P, Gestoras Pro Amnistía e.a. v. Council, [2007] ECR I- 1579, paras 168– 73.

713 Opinion of AG Van Gerven in Case C- 70/ 88, n 358.714 Case C- 134/ 02 P, Ombudsman v. Lamberts, [2004] ECR I- 2803, para 42.715 Ibid, para 48.


powers and institutional balance intended by the Treaty’.716 The Court indeed rejected such a non- contractual liability in order to safeguard the legislator’s freedom when defining policies,717 without commenting however on the institu-tional balance argument.

Under the second subrule, Lenaerts and Verhoeven mention Romano.718 However, because of the Court’s reference to Article 155 EEC it seems more appropriate to classify it under the third rule prohibiting the institutions from encroaching on the prerogatives of other institutions.719 After all, it was found that the Court in Romano took issue with the fact that the Council violated the Commission’s prerogatives by empowering the Administrative Commission, not with the Council delegating its own executive powers.

The most important part of the Court’s institutional balance jurisprudence can then be identified in relation to the third subrule, even if the Court will often ignore the institutions’ arguments.720 Rather, if the institutional balance plays a role, it will often be invoked by the Court itself. In Köster, the Court validated the comitology system in the CAP finding that these committees did not distort ‘the Community structure and institutional balance as regards both the relation-ship between the institutions and the exercise of their powers’.721 Here already the Court seemed to emphasize that the institutional balance dealt with the relation between the institutions and the way they exercise their powers, rather than con-stituting a constitutional protection for private parties.722 In Isoglucose, the Court enforced the institutional balance to strengthen Parliament’s position,723 and in Wybot the Court confirmed that it would verify out of its own motion whether the institutional balance is respected by the institutions.724

In Commission v. Ireland the Court confirmed for the first time that it was also bound by the institutional balance, since it is the sole prerogative of the Commission to decide when to bring proceedings against a Member State under Article 258 TFEU.725 Of course, the institutional balance also works the other

716 Joined cases C- 120/ 06 P and C- 121/ 06 P, FIAMM and Fedon v. Council and Commission, [2008] ECR I- 6513, para 142.

717 Ibid, para 174.718 Le Bot also mentions Romano as part of the jurisprudence where the institutional balance

acts as a limit to the independence of the institutions. See Fabien Le Bot, Le principe de l’ équilibre institutionnel en droit de l’union européenne, Paris, Université Panthéon- Assas, 2012, PhD Thesis, pp 316– 17.

719 Chamon, n 391, p 1063.720 See eg Case 302/ 87, n 710, para 19; Case C- 533/ 03, Commission v. Council, [2006] ECR

I- 1025, para 28; Case C- 77/ 11, Council v. Parliament, ECLI:EU:C:2013:559, para 28; Case C- 270/ 12, n 92, para 72.

721 Case 25/ 70, n 314.722 This protection for private parties was still apparent in Ehlermann’s commentary on

Köster: see Ehlermann, n 265, p 253.723 Case 138/ 79, Roquette Frères v. Council, [1980] ECR 3333, para 33.724 See Case 149/ 85, Wybot, [1986] ECR 2391, para 23.725 Case 415/ 85, Commission v. Ireland, [1988] ECR 3097, para 9. Similarly, see Joined cases

T- 68/ 89, T- 77/ 89 & T- 78/ 89, Società Italiana Vetro SpA, e.a. v. Commission, [1992] ECR II- 1403, para 319.


way and protects the prerogatives of the Court vis- à- vis the other institutions, as confirmed by the General Court in Pergan Hilfstoffe v. Commission.726

A less straightforward step was taken in Elio Fiorucci v. OHIM, where the CFI noted that it could not take a decision in lieu of the OHIM since otherwise it would ‘upset the institutional balance on which the division of jurisdiction be-tween OHIM and the Court is based’.727 The Court thereby extended the in-stitutional balance, which, although ‘laid down in the Treaties’, apparently also includes bodies such as EU agencies which are not even foreseen in the Treaties.

In France v. Commission, the competence of the Commission under Article 218 TFEU to conclude guidelines on regulatory cooperation and transparency with a third country was at issue. Although the Court concluded that the guidelines did not fall within the scope of Article 300 EC given they were non- binding, it immediately added:

Nevertheless, this judgment cannot be construed as upholding the Commission’s argu-ment that the fact that a measure such as the Guidelines is not binding is sufficient to confer on that institution the competence to adopt it. Determining the conditions under which such a measure may be adopted requires that the division of powers and the insti-tutional balance established by the Treaty in the field of the common commercial policy be duly taken into account.728

Through this ‘mystical’ part of its judgment,729 the Court made clear that it did not wish to extend the Commission’s external powers730 and to prevent a ‘grey zone’ from emerging as regards to the possibility to adopt ‘soft law’. Pitschas notes that the Court accorded legal significance to the Guidelines even though they were non- binding, without clarifying the nature of this legal significance.731 The Court probably feared that the institutions could construct a whole system of acts which were non- binding given the lack of formal competence for the adoption of such acts, but which, as a result of their non- binding nature, would also be immune from scrutiny. An elaborated body of such non- binding acts could also upset the institutional balance because it could de facto, but not de iure, affect the discretion which other institutions enjoy in subsequent rulemaking.

According to Baroncini, the Court indeed wanted to secure its own jurisdic-tion over these instruments.732 From an institutional balance perspective, it could

726 See Case T- 474/ 04, Pergan Hilfsstoffe v. Commission, [2007] ECR II- 4225, para 77.727 Case T- 165/ 06, Fiorucci v. OHIM, [2009] ECR II- 1375, para 67. See also Case T- 135/ 08,

Schniga GmbH v. CPVO, [2010] ECR II- 5089, para 85.728 Case C- 233/ 02, France v. Commission, [2004] ECR I- 2759, para 40.729 Nikolaos Lavranos and Ronald van Ooik, ‘Zaak C- 233/ 02 Franse Republiek

t. Commissie van de Europese Gemeenschappen (“EG- VS Richtsnoeren”)’, (2004) 52 SEW 12, p 447.

730 Kamiel Mortelmans, ‘De handelspolitieke bevoegdheden van de Europese Commissie’, (2004) Ars Aequi, p 455.

731 Christian Pitschas, ‘Keine Verletzung des Monopols zur Gesetzesinitiative der Kommission durch nicht bindende Leitlinien’, (2004) 15 EuZW 14, p 435.

732 Elisa Baroncini, ‘La Cour de Justice et le treaty making power de la Commission europée-nne’, (2006) RDUE 2, p 417.


thus be said that the Court acted to safeguard its own prerogatives.733 The Court’s insistence on the importance of the institutional balance in this case should be juxtaposed with its stance in Reynold Tobacco e.a. (discussed later).

In that case the Court confirmed the General Court’s reasoning that any act of an institution carries in it an incidental implication that the institution in ques-tion has adopted a position as to its competence to adopt that act, but that this position itself has no independent significance. Thus, whereas the Commission could not upset the institutional balance by deciding to commence proceedings against a private party before a foreign court, this could be the case when the Commission enters into agreements for practical cooperation with third coun-tries, although both acts were of a non- binding nature.

In the Minimum common lists case, the Parliament challenged a number of provisions of Council Directive 2005/ 85 through which the Council had created a secondary legal basis for itself to adopt a common list of safe third countries. Whereas the Treaty provisions prescribed either a procedure whereby the Council decides unanimously after having consulted the Parliament or the codecision pro-cedure, the Council’s secondary legal base foresaw a third option: the Council could adopt the list with QMV after consulting the Parliament. The Court refused to validate the possibility for Treaty institutions to create secondary legal bases, stating:

To acknowledge that an institution can establish secondary legal bases […] is tantamount to according that institution a legislative power which exceeds that provided for by the Treaty.

It would also enable the institution concerned to undermine the principle of institu-tional balance which requires that each of the institutions must exercise its powers with due regard for the powers of the other institution.734

The Court here also defined the institutional balance as a principle for the first time.

7.2.3.2 The possible vertical dimension of the institutional balanceIn Commission v. Portugal, the defendant relied on the institutional balance to argue that each Member State is itself responsible for defining the measures which it considers necessary for the protection of the essential interests of its security pursuant to Article 346 TFEU.735 If such a direct vertical dimension of the in-stitutional balance had been accepted, this would have come close to confirming Vos’ argument,736 but the Court dismissed this argument— without commenting on the institutional balance— by stating that Article 346 TFEU cannot be read as allowing a Member State to derogate from the Treaties by relying on no more than its essential security interests.737

733 Looking at the problem from the perspective of the EU’s autonomy, one could also argue that the Court stepped in to prevent conflicts from arising between the international legal order and the EU legal order, similarly to the Court’s function in the procedure under Article 218(11) TFEU.

734 Case C- 133/ 06, n 678, paras 56– 7.735 Case C- 38/ 06, Commission v. Portugal, [2010] ECR I- 1569, para 49.736 See n 705. 737 Case C- 38/ 06, n 735, para 64.


What is more, as regards a vertical function of the institutional balance, it was also rejected by AG Poiares Maduro in his opinion to the Waste Shipments case. In that case the correct legal basis of Regulation 1013/ 2006 was at issue. The Commission argued that the Regulation should have been adopted on the basis of Articles 175 and 133 EC (current Articles 192 and 207 TFEU) instead of solely on the former Article. The AG remarked that the disagreement on the legal basis did not affect the institutional balance because either way, the procedure to adopt the Regulation would have remained the same.738 On the other hand, he remarked that the true repercussions of the problem lay in the distribution of competences between the Union and the Member States, which would be greatly altered if the measure should also have been adopted on Article 133 EC.739 By neatly distinguishing the legal consequences hidden behind the controversy, the AG in effect excluded a possibly direct vertical dimension to the institutional balance. Furthermore, his strict rea-soning also argues against the view that the institutional balance is (merely) a hori-zontal expression of the principle of attribution of powers.740

Still, there is one line of cases in which the Court of Justice has alluded to a possible vertical dimension of the institutional balance. These cases concern the attempts made by sub- national entities in the EU to acquire the standing of privileged parties in EU procedural law. In 1997 the Court decided by reasoned order that entities such as Wallonia and Tuscany could not claim standing as if they were EU Member States since otherwise this ‘would undermine the institu-tional balance provided for by the Treaties, which, inter alia, govern the condi-tions under which the Member States […] participate in the functioning of the Community institutions’.741 Still, as Le Bot remarks, the reasoning here is rather unclear;742 Le Bot further observes that the concern or interest at issue, which militates against granting privileged standing to sub- national entities, is not so much that of the institutional balance but rather that of the unity of the national state.743 All in all, because the reasoning based on the institutional balance in this line of cases seems rather unconvincing, it does not seem strong proof of a vertical dimension to the institutional balance either.

7.2.3.3 Missed opportunities for the institutional balance as an ‘actionable’ principle

So far, the above- cited case law shows that the institutional balance applies be-tween the institutions of the EU, imposing obligations both on these institu-tions as well as on other actors. It equally seems that Court does not accept the same parallelism as regards the protection offered by the institutional balance, re-serving the latter to the institutions. A number of non- privileged parties have also

738 Opinion of AG Poiares Maduro in Case C- 411/ 06, Commission v. Parliament and Council, [2009] ECR I- 7585, para 6.

739 Ibid, para 7. 740 See text following n 706.741 See Case C- 95/ 97, Région wallonne v. Commission, [1997] ECR I- 1787; Case C- 180/ 97,

Regione Toscana v. Commission, [1997] ECR I- 5245. This was further confirmed in Case C- 417/ 04 P, Regione Siciliana v. Commission, [2006] ECR I- 3881, para 21.

742 Le Bot, n 718, p 125. 743 Ibid, p 129.


relied on the institutional balance, especially drawing analogies with Chernobyl, to secure their private rights, albeit unsuccessfully.

7.2.3.3.1 Private parties and the institutional balanceFollowing Meroni, one would expect the institutional balance to be of special rele-vance to private parties given the Court’s qualification of the balance of powers as a ‘fundamental guarantee’.744

This issue re- emerged thirty years later in Vreugdenhil II, in which the Court ruled on the conditions of the non- contractual liability of the Union. In Vreugdenhil I, the Court had previously invalidated Article 13a of Commission Regulation 1687/ 76 implementing Council Regulation 754/ 76, because the Commission had acted ultra vires.745 To regain its losses, Vreugdenhil applied for damages and the Court was asked to rule on whether there had been any suffi-ciently flagrant violation of a superior rule of law for the protection of the indi-vidual. The Court, following AG Darmon,746 ruled: ‘the aim of the system of the division of powers between the various Community institutions is to ensure that the balance between the institutions provided for in the Treaty is maintained, and not to protect individuals.’747

Vreugdenhil II is of paramount importance because the Court completely aban-doned the protective function (at least to the benefit of private parties) of the institutional balance which stood central in Meroni. Constantinesco noted that Vreugdenhil II ‘décevra ceux qui ont assisté à la montée du principe de l’équilibre institutionnel et qui l’ont considéré comme une utile création jurisprudentielle permettant de garantir les droits des institutions mais aussi ceux des opérateurs économiques’,748 further concluding that ‘[l] e champ d’application du principe de l’équilibre institutionnel semblerait alors devoir se limiter au seul contentieux inter institutionnel’.749 Where Constantinesco still used the conditional tense, the case law cited presently seems to take away all doubt that the function of the institutional balance has changed. Other authors were more critical than Constantinesco, noting that the ruling puts in doubt whether the Community (at that time) really was a community based on the rule of law.750 Recently, this criti-cism has been partly repeated by AG Bot in Artegodan v. Commission,751 although the Court in that case again confirmed its ruling in Vreugdenhil II.752

744 Case 9/ 56, n 223, p 152. 745 Case 22/ 88, Vreugdenhil BV, [1989] ECR 2049.746 Opinion of AG Darmon in Case C- 282/ 90, Vreugdenhil BV v. Commission, [1992] ECR

I- 1937, para 38.747 Case C- 282/ 90, Vreugdenhil BV v. Commission, [1992] ECR I- 1937. See also inter alia Joined

Cases T- 64/ 01 & T- 65/ 01, n 462, para 116.748 Vlad Constantinesco, ‘Jurisprudence— Recours en indemnité’, (1993) 120 Journal du

droit international 2, p 405.749 Ibid, p 406.750 See X, ‘CJCE, 13 mars 1992, Industrie en Handelsonderneming Vreugdenhil BV, Aff.

C- 282/ 90’, (1992) 2 Revue Europe Mai (5); Francette Fines, ‘Responsabilité extracontractuelle’, (1993) La Semaine Juridique II 22093.

751 Opinion of AG Bot in Case C- 221/ 10 P, Artegodan GmbH v. Commission, ECLI:EU:C: 2011:744, paras 40– 4.

752 Case C- 221/ 10 P, Artegodan GmbH v. Commission, ECLI:EU:C:2012:216, paras 81– 2.


An important and controversial shift in purpose of the institutional balance, between Meroni and Vreugdenhil II, may thus be noted. Subsequently, British Steel v. Commission left no doubt that the institutional balance lost not only its protective purpose, but also any protective function for private parties. In British Steel v. Commission,753 the applicant requested the annulment of a Commission decision since the latter had allegedly failed to adhere to the terms of the Council’s assent required under Article 95 ECSC. While the final decision was different from the text on which the Council gave its assent, the Court observed that the breach of a procedural rule will only constitute sufficient ground for a contested act’s annulment if in the absence of that irregularity the contested act could have been substantively different. It then cited the Council which had declared that ‘the decisions adopted by the Commission were indeed consistent with what had been agreed by it’.754

This part of the Court’s reasoning in British Steel sits uncomfortably with its earlier judgment in Isoglucose in which it had ruled that a breach of a fundamental institutional rule is in itself a sufficient ground for the annulment of an act con-taining such an irregularity,755 as further confirmed in Parliament v. Council.756 Could the institutions then rely on established institutional practice, within the bounds of the relevant Treaty provisions, as suggested by AG Poiares Maduro in the Minimum common lists case?757

The Court indeed suggested as such in Greece v. Council, where it noted that this margin is ‘limited by the separation of powers, as laid down in the Treaty, between the institutions. The Court must therefore make sure that in the context of interinstitutional cooperation the institutions do not ignore the rules of law and do not exercise their discretionary power in a manifestly wrong or arbitrary way.’758 Yet it could be argued that this is exactly what happened in British Steel. The assent of the Council under Article 95 ECSC fulfilled a specific function. This function would be undermined if an institutional practice was developed according to which the Commission’s final decisions could still be substantively different from the text consented to by the Council.

In Tralli v. ECB, the applicant’s argument on the institutional balance was less clear,759 and it was in any event rejected by the Court, noting that ‘it is suf-ficient to recall that that principle is intended to apply only to relations between Community institutions and bodies’.760

753 Case T- 243/ 94, n 403. 754 Ibid, para 190.755 Koen Lenaerts, Dirk Arts, and Ignace Maselis, Procedural Law of the European Union,

London, Sweet and Maxwell, 2006, p 285.756 Case C- 65/ 93, Parliament v. Council, [1995] ECR I- 643, para 21.757 Opinion of AG Poiares Maduro in Case C- 133/ 06, Parliament v. Council, [2008] ECR

I- 3189, para 29.758 Case 204/ 86, Greece v. Council, [1988] ECR 5323, para 27.759 The link between the alleged violation of the institutional balance and the applicant’s rights

was also unclear in Volkswagen, a case on state aid. Again here the CFI simply dismissed the argu-ment. See Joined Cases T- 132/ 96 & T- 143/ 96, Freistaat Sachsen, Volkswagen AG and Volkswagen Sachsen GmbH v. Commission, [1999] ECR II- 3663, paras 220 and 241.

760 Case C- 301/ 02 P, n 452, para 46.


In FNAB and others v. Council the applicants tried to circumvent the strict ad-missibility requirements under Article 230 EC by arguing that Chernobyl should be applied analogously. The applicants also claimed that the Court should raise an infringement of the institutional balance of its own motion, without a need for them to meet the admissibility requirements for an action for annulment. Perhaps unsurprisingly, the Court dismissed the applicants’ pleas and stated:

Although, as noted by the appellants, the Court of Justice stated emphatically, in […] Meroni v. High Authority, that the balance of powers that characterises the institutional structure of the Community constitutes a fundamental guarantee granted by the Treaty in particular to the undertakings and associations of undertakings to which it applies, that statement cannot be interpreted as providing a remedy for any natural or legal person who considers that an act of a Community institution has been adopted in breach of the principle of institutional balance, regardless of whether the act in question is of direct and individual concern to that person.

Finally, the appellants cannot rely on [Chernobyl] to show that their application should be declared admissible […] That judgment is based on the need to ensure continuing insti-tutional balance and judicial supervision of observance of the Parliament’s prerogatives. It is therefore not relevant to the admissibility of an action brought by a natural or legal person.761

A few months later, the Netherlands Antilles tried to convince the Court with the same arguments, so as to be treated as a (semi) privileged party762 rather than as an ordinary legal person. The Court did not deal with the arguments of the Netherlands Antilles on the analogy with the Chernobyl case, but AG Léger did and rejected any useful analogy.763

In Philip Morris and others (discussed previously), the applicants challenged the decisions of the Commission to start proceedings against them before a United States district court. As regards the admissibility of the proceedings before the CFI, one of the arguments was related to the institutional balance. The appli-cants claimed that the Commission had unilaterally added to its powers granted under the Treaties and that acts liable to affect the institutional balance should not escape judicial review. However, the Court ruled that the contested decisions did not have any binding legal effects and that the ‘seriousness of the alleged in-fringement by the institution concerned or the extent of its adverse impact on the observance of fundamental rights cannot justify an exception to the absolute bars to proceedings laid down by the Treaty’.764 On appeal, the Court of Justice validated the reasoning of the CFI.765

7.2.3.3.2 Member States and the institutional balanceThe cases discussed in the previous section show the limited usefulness of the insti-tutional balance for non- privileged litigants. Of course, besides the institutions, the

761 Case C- 345/ 00 P, FNAB, Setrab & Est Distribution Biogam SARL v. Council, [2001] ECR I- 3811.762 Case C- 452/ 98, Nederlandse Antillen v. Council, [2001] ECR I- 8973.763 Opinion of AG Léger in Case C- 452/ 98, Nederlandse Antillen v. Council, [2001] ECR I- 8853.764 Joined Cases T- 377/ 00, T- 379/ 00, T- 380/ 00, T- 260/ 01 & T- 272/ 01, Philip Morris

International e.a. v. Commission, [2003] ECR II- 1, para 87.765 Case C- 131/ 03 P, Reynold Tobacco e.a. v. Commission, [2006] ECR I- 7795.


Member States are also privileged litigants, raising the question of how the Court treats their pleas related to the institutional balance, given that the latter ‘is in-tended to apply only to relations between Community institutions and bodies’.766 This would mean that the Member States as such do not enjoy a special position under the institutional balance, without ruling out that they could still have an interest in invoking institutional balance violations, eg to protect the prerogatives of the Council or (in some cases, for smaller states) even the Commission.

In the Tobacco case, the question was raised as to whether Directive 2001/ 37 could be based on both Articles 95 and 133 EC.767 The German Government argued that because the Directive was adopted following the codecision pro-cedure, while Article 133(4) EC provided the Council decides by itself, the insti-tutional balance was jeopardized. AG Geelhoed acknowledged the institutional balance’s importance but commented: ‘In this, however, the Court establishes a direct link with the prerogatives of the European Parliament and the democratic principles underlying them’. From this point onward it was easy for the AG to state that it was hard to see which interest is adversely affected by use of the codecision procedure. In the words of the AG: ‘Considered from the aspect of insti-tutional balance: if there is any procedure in Community law which is designed to achieve an optimum balance between different authorities, that would appear to me to be the codecision procedure.’768

Here the AG ignored one of the fundamental differences between the insti-tutional balance and the principle of separation of powers, since there simply is no clear ‘optimum balance’ against which the institutional balance can be meas-ured. And even if there were, the only relevant balance would be the one laid down by the treaty authors in primary law. If AG Geelhoed’s reasoning were to be followed, questions on legislative procedure would automatically lead to the codecision procedure, regardless what the Treaties provide.769 AG Geelhoed seems to have been aware of this as well when he continued, stating: ‘[e] ven if there is an adverse effect on the recognized interest of the Council in being able to take decisions alone, it was the choice of the Community legislature itself— and thus also that of the Council— to leave that interest out of account.’ However, given that the institutions cannot alter the legislative procedure prescribed by the Treaties, nor renounce their prerogatives as laid down in the Treaties,770 that ob-servation cannot convince. The Court found that since the Directive could have been adopted on the single legal basis of Article 95 EC, the argument invoking the institutional balance could not be pertinent.771

766 See n 760.767 Case C- 491/ 01, British American Tobacco (Investments) Ltd, [2002] ECR I- 11453.768 Opinion of AG Geelhoed in Case C- 491/ 01, British American Tobacco (Investments) Ltd,

[2002] ECR I- 11453, para 180.769 That upholding the institutional balance may have positive consequences from a demo-

cratic perspective but that it is not the institutional balance’s principal function is also noted by Le Bot: see n 784.

770 Case 68/ 86, UK v. Council, [1988] ECR 855, para 38.771 Case C- 491/ 01, n 767, para 110.


Member States have invoked the institutional balance on numerous other occasions before the Court, but the Court did not always give proper at-tention to these arguments.772 In Germany v. Commission, for instance, the applicant invoked the institutional balance to show it was not possible under the Treaties to delegate a power to impose penalties on economic operators to the Commission— not too farfetched an argument, since at that time the only Treaty provisions which provided for the possibility to impose sanc-tions granted this power to the Council.773 The Court declined to engage with the German government’s reasoning, referring back to its Köster case law in which it emphasized that the (only) relevant distinction is between essential and non- essential elements, further concluding that the sanctions in casu fell under the non- essential elements of the common organization of the market.774

In Brasserie du Pêcheur the German government again invoked the institutional balance when the Court had to rule on Member State liability for breaches of directly effective EU law attributable to the legislature of that Member State. The German government argued that such a general right to reparation for individuals could only be created by legislation and could not be created by judicial decision. Such a judicial encroachment upon the powers of the legislator would moreover be incompatible with the institutional balance. Again the Court brushed the German government’s arguments aside and ruled that such a non- contractual liability of the Member States did exist.775

7.2.3.3.3 ConclusionThe present section has illustrated Chalmers’ critique in that the Court is not receptive enough towards arguments based on the institutional balance. A number of private parties have, unsuccessfully, tried to have the Court draw an analogy with its Chernobyl case law. In British Steel, the applicant unsuc-cessfully asked the Court to draw an analogy with the Isoglucose case, so that upsetting the institutional balance would amount to the infringement of an essential procedural requirement, causing the contested measure to be void. In Vreugdenhil II the Court ruled that the institutional balance was not in-tended to protect individuals, thereby setting part of its Meroni ruling aside. This jurisprudence and the cases in which the Member States have invoked the institutional balance can also be marked as a missed opportunity for the Court to clarify the purpose, nature, and content of the institutional balance. The Member States invoked the institutional balance in a plethora of dif-ferent cases, but the Court largely ignored these arguments instead of properly addressing them.

772 Apart from the cases discussed, see Case C- 239/ 01, Germany v. Commission, [2003] ECR I- 10333; Case C- 58/ 94, Netherlands v. Council, [1996] ECR I- 2169.

773 See Articles 172 and 87(2)a EEC (currently 261 and 103(2)a TFEU).774 Case C- 240/ 90, n 460.775 Joined cases C- 46/ 93 & C- 48/ 93, Brasserie du Pêcheur, [1996] ECR I- 1029.


7.2.3.4 Concluding remarksAn updated analysis of the Court’s jurisprudence confirms the three rules pro-posed by Lenaerts and Verhoeven, whereby especially the third rule seems to be the most important. At the same time it should be noted that the Court has not clarified to a great extent which institutions come under the institutional balance. Although the Court clearly included itself in the institutional balance, contrary to Guillermin’s view,776 the question of how far the institutional balance should be stretched is more difficult to answer. Using the conceptual framework of Hilf,777 the Court seems to have extended the institutional balance even to the tertiary layer in Elio Fiorucci, but this does not sit well with the Court’s own assertion in Chernobyl that the institutional balance is laid down in the Treaties. In his study on the institutional balance, Le Bot also precludes the possibility that bodies under secondary law could be part of the institutional balance,778 and echoes Guillermin in concluding that (at least the core) of the institutional balance con-sists of the institutional triangle.

A number of the cases discussed have also shown the Court’s reluctant stance in addressing institutional balance arguments, regardless of whether they were pre-sented by the institutions, Member States, or private parties. These are indeed missed opportunities because the Court could also further clarify the nature, pur-pose, and function of the institutional balance by refuting such often frivolous arguments.779

The cases referred to previously also offer interesting insights in the institutional balance’s function. Although it has been dubbed a principle, it most of the time ‘merely’ functions as supporting instead of steering the reasoning of the Court. This would confirm de Witte’s view to the effect that there is actually no room for a judge- made principle of institutional balance. However, Chernobyl and France v. Commission are two interesting exceptions. In those cases the Court relied on the institutional balance as a principle to solve a legal problem brought before it. Chernobyl may be seen as the apogee of the institutional balance jurisprudence of the 1980s which left many commentators speaking of a ‘principle of institutional balance’.780 Apparently, this principle was of such fundamental importance that it legitimized a contra legem (or praeter legem)781 interpretation of the Treaties.

776 Guillermin, n 638, p 328.777 Hilf, n 10, p 4.778 Le Bot, n 718, pp 157– 60. Le Bot puts forward a material criterion to identify those organs

which are covered by the institutional balance. Whereas being an ‘institution’ may create a pre-sumption that the organ also forms part of the institutional balance, this is only really the case when the organ has been granted a decision- making power in the decision- making process of the EU. However, the ongoing agencification undermines this, since some agencies indeed meet the material criterion while not being institutions.

779 See also Merijn Chamon, ‘Zaak C- 77/ 11, Raad t. Parlement’, (2014) 62 SEW 2, p 90.780 See for instance Bruno de Witte, (1991) 39 SEW 11, p 756; Jean- Paul Jacqué, ‘La légiti-

mation active du Parlement européen ou il n’était pas nécessaire d’espérer pour entreprendre’, (1990) 26 RTDE 3, p 627; Kamiel Mortelmans, ‘Rechten en plichten van Europese instellingen: de actieve legitimatie van het Europees parlement en de openbaarheid van het Commissiebestuur’, (1990) Ars Aequi, p 985.

781 See n 712.


The case drew the attention not only of legal academics but also of practition-ers. However, the enthusiasm of these privileged and non- privileged parties in invoking the institutional balance was not rewarded by the Court. In fact there is not a single case following Chernobyl in which a party successfully relied on the institutional balance, that is, a case in which the notion was invoked by one of the parties and where the Court explicitly endorsed such an ‘institutional balance argument’. The only cases in which the notion thus played a decisive role are those in which the Court out of its own motion raised the institutional balance in its assessment of the parties’ arguments.782

The reluctance of the Court to accord any meaningful role to the institutional balance has resulted in Chalmers’ critique, although of course it could also have to do with the tendency of the parties to invoke the notion, given its inherent sub-jectivity, as a last- ditch effort to convince the Court of their arguments. Perhaps it came as a surprise then that the Court seemed to reaffirm the importance of the institutional balance by qualifying it for the first time as a principle in 2008.783

7.2.4 The institutional balance: A principle with sufficiently clear content and independent function?

From the sections above a rather complex image of the institutional balance arises. As regards its function, Le Bot concluded that ‘[s] on objectif est de protéger les éléments fondamentaux du système institutionnel de l’Union européenne’, result-ing in ‘une fonction de protection systémique’.784 And whereas its content has been adequately described by Lenaerts and Verhoeven, other problems persist; for instance, the scope of the institutional balance not being clear from the Court’s jurisprudence.

Further, if the institutional balance is a principle, it is rarely used as one, most of the time being a shorthand for Article 13(2) TEU to substantiate the Court’s rea-soning.785 If it is a principle it is also a strange one in the sense that parties cannot rely on it. Le Bot observes that the institutional balance is indeed a general prin-ciple of EU law, but that the Court of Justice does not seem to have drawn all the necessary conclusions from this fact.786 The Court’s jurisprudence in this regard

782 Notably in Köster, Isoglucose, Wybot, Commission v. Ireland, Vetro SpA, Pergan Hilfstoffe, France v. Commission, Minimum Common Lists, Elio Fiorucci, and Schniga.

783 Case C- 133/ 06, n 678, para 57.784 Le Bot, n 718, p 243. Le Bot further refutes that the principal function of the institutional

balance is the promotion of democracy: ‘si l’argument de l’équilibre institutionnel peut incidem-ment avoir des vertus de promotion démocratique, là ne réside pas sa fonction principale.’ Michel however sees three main functions to the principle: (i) interest representation, (ii) preventing the abuse of power, (iii) efficiency in decision- making, (iv) securing democratic legitimacy: see Michel, n 3, pp 86– 90.

785 In further cases with an interinstitutional dimension following Short- selling, this has not changed: see eg Merijn Chamon, ‘The Dividing Line between Delegated and Implementing Acts, Part Two: The Court of Justice Settles the Issue in Commission v. Parliament and Council (Visa reci-procity)’, (2015) 52 CMLRev 6, p 1623.

786 Le Bot, n 718, p 298.


can be neatly summarized by its statement in Tralli where it observed that the institutional balance ‘is intended to apply only to relations between Community institutions and bodies’. The Court’s reticence to accord any substantial mean-ingful role to the institutional balance leads to a final obstacle: the Court seems to claim a monopoly of some sort on the institutional balance, since it does not seem to allow a party before it (not even the institutions themselves) to rely on the notion in their pleas. Where the institutional balance plays a meaningful role in the Court’s ruling, it is on the Court’s own initiative.

Because of this, the notion of institutional balance under Article 13 TEU has not realized its potential to become a principle of institutional balance. Still, the institutional balance could be a general principle of EU law, since it applies to the whole legal system (although it would not be uniform and would vary depending on the subject matter). In addition it would be an indeterminate principle, since it would not favour an expansive reading of the scope of EU law (such as the prin-ciple of supremacy or effet utile) and neither would it favour a restrictive reading thereof (such as the principles of conferred powers or subsidiarity).787

The Court’s jurisprudence may then shed light on the function of a principle of institutional balance. In Chernobyl and in France v. Commission the Court used the institutional balance as a gap- filling principle. In Chernobyl, the Court was confronted with an inconsistency in primary law following the SEA. In France v. Commission, the Court was confronted with a grey zone for which the Treaties did not provide exact guidance. Although the institutional balance functioned as a principle, it cannot be said that it structured the judicial enquiry (cf. section IV 7.2). For this, again, the Court should develop its jurisprudence and a sort of institutional balance test should be developed.

Because the principle’s function would be gap- filling, as Prechal suggested,788 certain common problems will be solved without reference to the institutional balance, especially if the relevant Treaty provisions are sufficiently clear, as al-luded to by de Witte in relation to controversies on the legal basis of legislative measures.789 On the other hand, the institutional balance could be used as a def-ining principle in those cases in which the Treaty provisions do not give sufficient guidance.

This may be the case following a new Treaty (amendment) when no consensus can be agreed on how the new provisions should be understood. In deciding be-tween rival interpretations, the Court could then rely on the institutional bal-ance. The latest Treaty revision through the Treaty of Lisbon provides ample examples: the precise distinction between the realms of Articles 290 and 291

787 For a discussion of the general/ specific and the expansive/ indeterminate/ restrictive principles in EU law, see Beck, n 172, pp 194– 207.

788 See n 693.789 See n 692. This is not to say that applying the relevant rules (centre of gravity and absorption

doctrines) mechanically produces results: the Court will still have to engage in an interpretation ex-ercise. In the case on the legal base for ratifying the Rotterdam Convention for instance, AG Kokott and the Court came to different conclusions even if they started from the same premises. See Case C- 94/ 03, Commission v. Council, [2006] ECR I- 1.


TFEU,790 the new institutions in the external policy of the EU (the double- hatted High Representative, the European External Action Service, and the permanent President of the European Council), etc.

A second type of situation in which the institutional balance may play a role, then, is when a given institutional practice or development, on which there is an institutional consensus, does not find its origins in the Treaty but finds itself in a ‘grey zone’. In the past this has been the case for the comitology system, which the Court sanctioned in a number of cases, starting with Köster. In France v. Commission the Court used it as principle to prevent a ‘grey zone’ from arising. It should be noted here that if there is a consensus between the institutions on a certain practice, challenging it will often depend on private parties (or Member States), which again stresses the need to allow private parties (and Member States) to rely on the institutional balance as a principle before the Court. For instance, it was no coincidence that it was a Member State, the UK, which challenged the empowerment of an agency for the first time in Short- selling. Indeed, agencifica-tion is a significant grey phenomenon which could be scrutinized through the lens of the institutional balance.

7.2.4.1 An institutional balance test as a tool to scrutinize agencificationBecause a test is useful to transform the notion of institutional balance into an independent principle which may be weighed against other principles and require-ments (eg efficiency of decision- making) and guide the Court to come to solutions in specific cases, an embryonic test will be proposed here. This test should then also help in the present study to scrutinize the process of agencification under the principle of institutional balance.

When a party raises the alleged infringement of the institutional balance, such pleas will often come down to a request to sanction or censure the exercise of a given power by a given institution or body. In a first step it could then be det-ermined whether the principle of institutional balance may indeed be invoked. If the Treaty provisions themselves provide sufficient guidance, reference to the institutional balance is unnecessary and the Court should resolve the case by ap-plying and interpreting the specific Treaty provisions.

However, if the Court concludes that the Treaty provisions provide insufficient guidance, it could proceed. In a second step, the granting or withholding of the

790 These new Articles have indeed already given rise to a number of cases before the Court: see cases C- 427/ 12, Commission v. Parliament and Council, ECLI:EU:C:2014:170; C- 65/ 13, Parliament v. Commission, ECLI:EU:C:2014:2289; C- 88/ 14, Commission v. Parliament and Council, ECLI:EU:C:2015:499. In the last case, AG Mengozzi made the interesting suggestion of applying an institutional balance test: see Chamon, n 785, p 1623. While the institutions have adopted a Common Understanding on delegated acts (see Annex to Council Document 8640/ 11 of 4 April 2011— for a short discussion, see Fabien Le Bot, ‘La soft law et les procédures d’adoption des actes de l’Union eu-ropéenne’, (2014) RUE 576, pp 142– 3), the Common Understanding cannot prevent controversies from arising. See for instance the practice of the Council and Parliament, referred to in the introduc-tory chapter (cf. section I 1.2.1.3.1), whereby the Commission is instructed to cooperate with an ESA when working out delegated acts.


contested power should be put into the legal context as it results from the Treaties. Here the Court may employ its whole arsenal of interpretative techniques and especially those of systemic and purposive interpretation further taking into account ‘genetic’ and historical arguments,791 by relying on the genesis of Treaty provisions and the travaux préparatoires.

In a third step it could be verified what the effect would be if the power con-cerned would be sanctioned or rebuked.

Reapplying this test to the two cases in which the institutional balance may be said to have been employed as a principle largely confirms the outcome in those cases. In Chernobyl the Court was confronted with a Treaty inconsistency follow-ing the SEA, the Parliament not being able under Article 173 EEC to challenge an act of the Council that would deprive it of its newly acquired powers under Article 100a EEC. The Court had to put the power claimed by the Parliament, that is, active locus standi under Article 173 EEC, in its context. This is some-thing AG Darmon did extensively in his opinion in Comitology,792 presenting an overview of the Parliament’s position (at that time) in proceedings before the Court. In a third step the Court looked at the effect of, in this case, withholding such a power to the Parliament, and had to conclude that this would result in the inability of the European Parliament to safeguard the prerogatives it had acquired under Article 100a EEC.

The issue in France v. Commission had to do with a ‘grey zone’. The possi-bility that the Commission could autonomously conclude binding agreements with third countries had already been rejected and the question remained as to whether the Commission did have a power to conclude non- binding agreements with third countries. This power had to be put further into the context of the Commission’s autonomous international function. Undoubtedly the possibility of some autonomy for the Commission in this field was not completely exotic.793 Therefore the Court had to look further into the effects of granting such a power to the Commission. For the reasons set out previously, the Court could have felt such a power might upset the institutional balance, as the Commission could otherwise escape the jurisdiction of the Court by adopting soft law agreements and at the same time narrowing the de facto discretion of the other institutions (in casu the Council) to adopt hard law.

This embryonic test does not resolve all the outstanding issues related to the institutional balance, an important one being the possibility for a non- privileged party to rely on the principle. The Court itself seems very reluctant, to say the least, to allow private parties to rely on the institutional balance, but the test as

791 Building on the work of Bengoetxa, Beck elaborated the different techniques of interpret-ative argumentation and identifies linguistic, systemic, teleological, functional, consequentialist, and comparative criteria as well as the genetic or historic and special legal arguments. See Beck, n 172, pp 187– 233.

792 Opinion of AG Darmon in Case 302/ 87, Parliament v. Council, [1988] ECR 5615, para 3. The AG inter alia referred to Case 13/ 83, Parliament v. Council, [1985] ECR 1513; Case 20/ 85, Roviello, [1988] ECR 2805; Case 294/ 83, n 669.

793 See Baroncini, n 732, pp 371– 4.


formulated above would not a priori rule out this possibility. Another question is whether the Court could raise an infringement of the institutional balance out of its own motion. Because the Court’s duty is to ‘ensure that in the interpret-ation and application of the Treaties the law is observed’ (cf. Article 19 TEU) and because the institutional balance would ultimately refer to the constitutional foundations of the EU, this should be the case, even if the test set out above only provides for a residual function for a principle of institutional balance.

7.2.4.1.1 The test applied to ESMA’s powers contested in Short- sellingAs noted, the first two pleas in Short- selling invoked Meroni and Romano and were dealt with in the preceding. In its third plea the UK argued that ‘Article 28 [of Regulation 236/ 2012] purports to confer on ESMA a power to adopt non- legislative acts of general application, whereas in the light of Articles 290 and 291 TFEU, the Council has no authority under the Treaties to delegate such a power to a mere agency outside of these provisions’.794

The reason why the UK referred to both Articles may be that the delimitation between these two acts itself is not clear. As was noted previously, this is one of these grey areas in which the Court could bring some clarification by making use of a principle of institutional balance.795 If the ESMA were to impose a ban on shorting it would undoubtedly adopt an act of general application without it being formal legislation, but it does not seem that such an act would supplement or amend Regulation 236/ 2012. Instead it would seem that the ESMA would be adopting an act to implement Regulation 236/ 2012 under uniform conditions, hence meeting the precondition of Article 291(2) TFEU.

a. Articles 290 and 291 TFEU under the first step From the perspective of the institutional balance test, the UK argued that the grant of power by the legis-lator to ESMA violated the Commission’s prerogatives. The first step of the test is indeed passed, because the Treaties do not contain a general enabling clause to empower agencies and the two Articles invoked are not entirely clear. The rele-vant part of Article 290 TFEU provides: ‘A legislative act may delegate to the Commission the power to adopt non- legislative acts of general application to sup-plement or amend certain non- essential elements of the legislative act’ (emphasis added). This provision could then be interpreted as providing a possibility to the legislator to delegate powers to the Commission, the authority explicitly men-tioned, or to another yet unqualified authority. The textual provisions of Article 291 TFEU in turn provide: ‘Where uniform conditions for implementing legally binding Union acts are needed, those acts shall confer implementing powers on the Commission, or, […] on the Council ’ (emphasis added). Here the legislator has some choice, but prima facie only between two authorities, the Commission and the Council. Hence, it might appear that an institutional balance test cannot be applied but, given that Article 173 EEC also appeared rather clear on the lack of

794 See Case C- 270/ 12: Action brought on 1 June 2012, OJ 2012 C 273/ 3.795 See n 790. However, in Biocides, the Court did not make use of the institutional balance.


the Parliament’s standing in Chernobyl (discussed previously), it cannot be com-pletely ruled out.

b. Article 291 TFEU under the second step Suppose, despite the wording of Article 291 TFEU, that the reference to the Council is seen as only one of the possible exceptions to the rule that the Commission should be empowered to adopt implementing acts. Through this reference, the Treaty authors would then have acknowledged that ‘in duly justified specific cases’ bodies other than the Commission could be empowered, explicitly naming the Council as only one of them. Still, because the Treaties do not provide for the establishment and em-powerment of EU agencies in general, the Court would not have had many use-able points of departure to elaborate a reasoning allowing EU agencies to adopt implementing acts. The main point would be the reference to the agencies in the Treaties on the EU’s system of legal protection, which was indeed extensively referred to by AG Jääskinen (cf. section IV 7.2.4.1.2) and by the Court itself (cf. section IV 7.2.4.1.3).

The references to the agencies in these Articles would indeed be meaningless if they were not allowed to adopt any binding acts.796 However, this does not auto-matically mean that agencies should be allowed to adopt implementing acts given the Commission’s prerogatives under Article 291 TFEU. Can other principles or values be invoked to be weighed off against the institutional balance in this case? In line with the argument elaborated previously (see section IV 2), the subsidiarity principle may be noted here. In addition, one could refer to the nature and pur-pose of the implementing act which, depending on the policy field and the matter to be regulated, requires highly technical expertise. To ensure the efficiency and effectiveness of the decision- making, EU agencies’ involvement together with the Commission, or exceptionally independently, could be justified here. Again this is something which the AG also hinted at.797

c. Article 290 TFEU under the second step Under the second step two different alternative questions may actually be raised. Can an agency adopt a delegated act (or also an implementing act, cf. supra) and can the legislator create atypical acts (resembling the acts described in Articles 290 and 291 TFEU) and empower agencies to adopt these atypical acts?

A systemic interpretation of the Treaties would not result in any clues support-ing such demarches, since again the Treaties do not in the first place explicitly allow for the creation and empowerment of EU agencies. No serious genetic or historic arguments could then be identified either. As is noted later (see sec-tion VI 2.1), the topic of EU agencies has been brought up numerous times in intergovernmental conferences (IGCs) but the matter was never dealt with. One should therefore assume that the drafters of the Lisbon Treaty, when introducing

796 However, Orator notes that the reference to the agencies in Article 277 TFEU on the plea of illegality cannot be taken as a sanctioning of agencies exercising normative powers, since it would mean that any office, body, or agency may be empowered to adopt generally binding acts. Instead Article 277 TFEU is broadly worded to ensure a complete system of judicial protection. See Orator, n 84, p 855.

797 See Opinion of AG Jääskinen in Case C- 270/ 12, n 74, para 99.


the novel instrument of the delegated act, were aware of the existence of EU agencies in the EU’s administrative sphere, but consciously only mentioned the Commission as the delegate authority in Article 290 TFEU. Indeed, as regards the implementation of EU law at least two amendments were, unsuccessfully, tabled by several members of the Convention referring to agencies as authorities which could be empowered by the legislator.798

As regards the possibility for the legislator to create further ‘atypical acts’ by secondary legislation which meet the characteristics of a delegated act, it is also doubtful that the Treaty drafters had this in mind when elaborating the would- be Article 290 TFEU.799 After all, the working group dealing with this issue in the convention was called the ‘Working Group on Simplification’ and it, successfully, proposed to introduce the only binding atypical act (ie the Beschluß ) in the offi-cial catalogue of Article 288 TFEU.800 From a teleological perspective there is no immediate reason why Article 290 TFEU should be interpreted so as to allow the ESMA to adopt delegated acts (or to adopt atypical acts which resemble delegated acts) when the Treaties allow the same results to be achieved by granting such authority to the Commission. Lastly, the reasons cited to allow agencification in relation to Article 291 TFEU do not seem as forceful in relation to Article 290 TFEU.

d. Articles 290 and 291 TFEU under the third step The test for Article 290 TFEU would thus seem to fail on the second step, but for the sake of com-pleteness the basic elements of a third step will also be sketched. Suppose for instance that under the second step a teleological interpretation of Articles 114 iuncto 26 TFEU leads to the conclusion that for the proper functioning of the internal market in financial services, Article 290 (and/ or 291) TFEU should be interpreted as allowing the legislator to delegate powers to a body such as the ESMA.801 If so, the Court should immediately define new limits to this added flexibility for the legislator, setting out when exceptions to the rule of Articles 290 and 291 TFEU may apply. Otherwise the Court would give carte blanche to the legislator, allowing it to completely hollow out the Commission’s role under Articles 290 and 291 TFEU. These limits should be set as such to preserve the prerogatives of the Commission.

Such limits would then be defined in the light of the reason why the legis-lator chose to grant the agency a power to adopt delegated/ implementing acts. For

798 See Secretariat of the European Convention, Contribution submitted by Mr Erwin Teufel, CONV 495/ 03, CONTRIB 198; Secretariat of the European Convention, Suggestion for amend-ment of Article 28 by Mr Andrew Duff a.o., http:// european- convention.eu.int/ docs/ treaty/ pdf/ 28/ Art28Duff.pdf.

799 Lefèvre makes the general distinction between the atypical acts for which the EU assumes (political) responsibility and those for which it does not. Yet, in both cases those acts are always adopted by actors whose intervention is foreseen by the Treaties. See Lefèvre, n 121, pp 14 and 43.

800 To allow for flexibility, the working group did not want to touch on the atypical acts devel-oped by the institutions, but it did mention that these acts are ‘in principle devoid of any binding legal force’. See Final Report of Working Group IX on Simplification, CONV 424/ 02, p 6.

801 This in itself would be doubtful under Declaration 39 on Article 290 TFEU annexed to the Treaty of Lisbon, which reads: ‘The Conference takes note of the Commission’s intention to continue

http://european-convention.eu.int/docs/treaty/pdf/28/Art28Duff.pdf





instance, in so far as an argument on technical expertise could convince, the limits should be defined in light of the agencies’ greater expertise and the technicality of detailed delegated/ implementing decisions.

In order to safeguard the prerogatives of the Commission, the agency’s power to adopt implementing acts would further have to be subject to further rules laid down in Commission delegated acts, or even be provided for in those delegated acts, rather than in formal legislation. The fact that this is only possible for agency empowerments under Article 291 TFEU distinguishes the situation under Article 291 TFEU from that under Article 290 TFEU.

7.2.4.1.2 AG Jääskinen’s testDespite mentioning the institutional balance, AG Jääskinen used a different test in his Opinion. His differentiation of the delegated acts under Article 290 TFEU from the implementing acts under Article 291 TFEU was based on a democracy argument.

According to the AG, agencies cannot be empowered to adopt delegated acts because the latter may alter the normative content of legislative acts. This requires a democratic accountability which only the Commission has.802 As re-gards the implementing acts, however, the AG noted that the same restriction does not apply.803 AG Jääskinen correctly observed that the fact that the dis-tinction between implementing and delegated acts is not always clear cannot be invoked as a reason for rejecting such fundamental consequences of the distinction.804

The problem remains however that, just like Article 290 TFEU, Article 291 TFEU does not make any references to agencies, and even the comitology regu-lation adopted under the latter Article does not mention the agencies a single time.805 The AG’s willingness to grant implementing powers to the EU agencies then rested on two elements. Firstly, there are the different functions of delegated and implementing acts. Secondly, the AG referred to the system of judicial pro-tection to which the EU agencies have now also been subjected, which would be pointless if they could not adopt binding acts.

While the second argument may seem sensible, accepting the existence of an implicit Delegationsnorm in the sense of Barbey (cf. section IV 5.7.1), inferred from the system of judicial protection, is still quite a fundamental step to take. As regards the first argument, while it is evident that a sufficient degree of demo-cratic legitimacy is in order when (material) legislation is adopted, democratic concerns are not absent from the implementing acts either.806 It is not entirely

to consult experts appointed by the Member States in the preparation of draft delegated acts in the financial services area, in accordance with its established practice.’

802 See Opinion of AG Jääskinen in Case C- 270/ 12, n 74, para 85.803 Ibid, para 86. 804 Ibid, para 78 at note 103.805 Regulation (EU) 182/ 2011, OJ 2011 L 55/ 13.806 See the scrutiny mechanism in Article 11 of the Comitology Regulation.


clear therefore why the lack of democratic legitimacy of EU agencies is problem-atic only for delegated acts but not for implementing acts.

In addition, the distinction on democratic grounds is not of a fundamental nature. As will further be noted, one could take the Parliament’s democratic scru-tiny powers vis- à- vis the Commission and mirror them in the acts laying down the statutes of the EU agencies. The EU agencies would then find themselves in the same position as the Commission, except for the fact that the Parliament’s scrutiny powers over the agencies would only be laid down in secondary law. Still, if the Delegationsnorm may be implicitly presumed under the Treaties, the Parliament’s scrutiny powers over the EU executive (including the agencies) may surely be presumed to be all- encompassing and derived from primary law as well. As a result, if the right mechanisms were instituted to allow the Parliament to ex-ercise the same control over EU agencies as over the Commission, there would no longer be any reason to deny EU agencies the power to adopt delegated acts under the reasoning of the AG. What is more, if the Parliament’s scrutiny powers over the agencies surpassed those it has vis- à- vis the Commission, the AG’s argument would even have the legislator empower agencies rather than the Commission under Article 290 TFEU.

7.2.4.1.3 The ruling of the Court and its implications for the debate on agencification

Whereas AG Jääskinen still emphasized the importance of the institutional balance— without, however, including it in his analysis— the Court went a step further and completely ignored it. In a sense this had been foreseen by the Court’s rejection of any relevance for Romano beyond Meroni (cf. section IV 6.1.3.2).

As regards the UK’s first two pleas, it was expected that the Court would re-interpret its Meroni and Romano jurisprudence. However, some anticipated that it would not be so easy for the Court to interpret the clear wording of Articles 290 and 291 TFEU flexibly enough to allow the kind of agencification at issue in Short- selling,807 while others warned of a ‘foolish judicial disregard for the vital need to ensure continuing financial stability within Europe’ if the Court were to strike down Article 28, despite good legal reasons to do so.808

As noted previously (see section IV 7.2.4.1.1), it would appear difficult to in-terpret those two Articles in a way that does not accord special importance to the role to be played by the Commission in the EU’s executive sphere. Still, this is exactly what the Court did in Short- selling after it observed that the question to be addressed was

whether the authors of the FEU Treaty intended to establish, in Articles 290 TFEU and 291 TFEU, a single legal framework under which certain delegated and executive powers may be attributed solely to the Commission or whether other systems for the delegation

807 Chamon, n 24, pp 158– 9. 808 Everson, n 661, p 50.


of such powers to Union bodies, offices or agencies may be contemplated by the Union legislature.809

The issue, also noted by Triantafyllou (cf. section IV 5.4.1.2), of whether Articles 290 and 291 TFEU lay down a ‘norme de délégation revêtant un caractère exclu-sive’810 was then answered in the negative.

Remarkably, since the possibility to do so was refuted previously (see section IV 7.2.4.1.1), the Court here appeared to rely on a genetic argument811 by refer-ring to the intentions of the Treaty authors. However, in its subsequent reasoning it does not really show the intent of the primary legislator. After remarking that the Treaties do not contain an explicit Delegationsnorm, it derives the above- noted intent from Articles 263, 265, 267, and 277 TFEU.812 The Court concluded from this that agencies may be empowered to adopt (general) binding acts and further found that the ‘conferral of powers [in casu] does not correspond to any of the situations defined in Articles 290 TFEU and 291 TFEU’.813 Of course this state-ment cannot be doubted, but the question was whether that non- correspondence meant the conferral was ultra vires. The Court did not really expound on this and simply concluded that the empowerment of the ESMA under Article 28 of Regulation 236/ 2012 did not undermine the system laid down in Articles 290 and 291 TFEU.814 In light of the observations made in the section on an insti-tutional balance test (cf. section IV 7.2.4.1.1), it should be clear that the Court thereby disregarded the effects on the Commission’s prerogatives.

The Court simply postulated that the empowerment in casu did not corres-pond with the situations described in those Articles, without indicating the ele-ments differentiating ESMA’s powers under Article 28 from the Commission’s powers under Article 291 TFEU. The Court did further note that the ESMA’s power should be seen in its context and that the national authorities and the ESMA should

be in a position to impose temporary restrictions on the short selling of certain stocks, credit default swaps or other transactions in order to prevent an uncontrolled fall in the price of those instruments. Those bodies have a high degree of professional expertise and work closely together in the pursuit of the objective of financial stability within the Union.815

The reasoning of the Court on the third plea is then unconvincing. The Court sets out the secondary law context in which the ESMA’s power needs to be situated without showing how this could be relevant in interpreting Articles 290 and 291 TFEU. Rather, the Court seems to start from the premise that the (ESMA’s) inter-vention powers are necessary and that the authorities must have the competence

809 Case C- 270/ 12, n 92, para 78. 810 See n 320. 811 See n 791.812 Case C- 270/ 12, n 92, para 80. For a critique, see n 796. In addition it should be noted that

the Court in Grimaldi found that also non- binding acts can form the subject of preliminary ques-tions regarding validity or interpretation under Article 267 TFEU. See Case C- 322/ 88, Grimaldi, [1989] ECR 4407, paras 7– 9.

813 Ibid, para 83. 814 Ibid, para 86. 815 Ibid, para 58.


to address an uncontrolled fall in the price of certain financial instruments. If the premise is that the contested power needs to be upheld, it is only logical to find that the ESMA’s powers could not undermine the system of Articles 290 and 291 TFEU.816

In essence the Court has made clear that the Treaty authors did not intend to establish a single legal framework under Articles 290 and 291 TFEU.817 The Court thereby appears to give the legislator carte blanche to elaborate this frame-work, without explicitly setting a limit on the legislator’s freedom. The only limits which could be inferred from the Court’s ruling are its references to (i) the tem-porary nature of the measures which may be adopted by the ESMA, (ii) the extra-ordinary circumstances in which the contested power will be exercised, and (iii) the fact that the ESMA has a high degree of professional expertise. However, even if the legislator were required to show these conditions are met, this does not appear to sufficiently safeguard the Commission’s prerogatives under those Articles. In addition, Ohler rightly noted that the modest limits which the Court sets for empowering agencies are in contrast with the requirements for empower-ing the Commission under Article 290 TFEU,818 making it easier to empower an atypical body like an agency than to empower a Treaty institution. Lenaerts on the other hand is more optimistic and notes that the resulting flexibility ‘ena-bles the EU legislator to evaluate the institutional capacities of the EU institu-tions vis- à- vis those of EU agencies’.819 As a result, for ‘the adoption of complex rules requiring technical and professional expertise’, agencies may be empowered. Lenaerts does not find the legislator’s discretion in this problematic, since the legislator, following the Biocides case,820 also has a discretion when it chooses

816 On this tautological reasoning of the Court, see also Clément- Wilz, n 94, p 346.817 The Court seems to have ignored this itself when in Biocides it seemingly presented Articles

290 and 291 TFEU as a closed system, noting that the notion of implementing act cannot be under-stood without an assessment of the notion of the delegated act. See C- 427/ 12, n 790, para 35. The purport of the Court’s ruling in the cases on the unitary patent package is less clear. In those cases Spain had questioned the empowerment of a select committee of the EPO (cf. section IV 5.5.1.4.3) inter alia because it violated Article 291(2) TFEU. The Court noted that the powers in question came under 291(1) TFEU and that EU regulation was a ‘special agreement’ pursuant to the EPC, meaning the Member States had retained implementing powers and were simply invited by the EU legislator to exercise them in an EPO committee. Remarkably, the Court found that Article 291(2) TFEU was not applicable since allegedly administering the renewal fees for the unitary patent does not require uniform conditions. See Case C- 146/ 13, n 466, paras 75– 82.

818 Ohler, n 583, p 251. Similarly, Skowron notes the difficulty in the Court’s argument: Both the Commission and the Council are institutions set up by the Treaties and their exercise of execu-tive powers is strictly spelled out in Articles 290 and 291 TFEU. Agencies, however, are not pro-vided for in the Treaties and would not require any legal basis in primary law to exercise executive powers. See Skowron, n 584, p 353.

819 Koen Lenaerts, ‘EMU and the EU’s Constitutional Framework’, (2014) 39 ELRev 6, pp 762– 3. However, Kollmeyer correctly argues that the Court was wrong to find that the Commission, under Article 290 or 291 TFEU, is not called upon to exercise technical expertise. See Daniel Kollmeyer, Delegierte Rechtsetzung in der EU – Eine Analyse der Art. 290 und 291 AEUV, Baden- Baden, Nomos, 2015, p 355. The idea that agencies, rather than the Commission, should be em-powered when technical expertise is required could give rise to a new delimitation line (general vs. technical expertise) similar to the delimitation lines between essential and non- essential elements and delegated and implementing acts.

820 See C- 427/ 12, n 790.


between empowering the Commission either under Article 290 or 291 TFEU.821 Of course, this ignores the fact that Biocides, differently from Short- selling, deals with the choice between two Treaty- provided scenarios.

In all, it appears as if the Court could only reject the UK’s third plea by ig-noring the institutional balance. The result of this is that the idea of an actionable principle of institutional balance suffers another blow, since the Court continues its stepmotherly relation with that principle. Short- selling then is another argu-ment for those who find the institutional balance not to be a principle at all. More generally, the Court’s sanctioning of this institutional experimentalism, without setting clear limits for the legislator, is also a potential Pandora’s box when the coherence of EU law is already under pressure.822

Specifically for the EU agencies and agencification, it unlocks the door to truly (US- style) regulatory agencies, since the arsenal of instruments at the disposal of EU agencies has taken a qualitative step. One major hurdle here is that the intel-ligible standard in the US (cf. section V 1.3) still allows for greater agencification than the Court’s Meroni- light doctrine. In any event, Short- selling does not ad-dress the question of the fundamental characteristics of the EU administration. Where EU agencies fit in should first depend on a definition of the role of the Commission. This issue can only be addressed in primary law and will be ad-dressed in a final chapter.

7.3 Conclusion

The present part started by differentiating the institutional balance from the clas-sical principle of separation of powers, finding both relevant for the EU legal order.

For the moment the academic debate on the nature of the institutional balance cannot be concluded, even if the notion has the potential of being a genuine principle of EU law which could help structure the enquiry of interinstitutional controversies, such as the process of agencification in the present study.

The UK’s challenge in Short- selling was taken as a case to apply an institutional balance test to the process of agencification. The fact that ESMA’s empower-ment to adopt acts under Article 291 TFEU was doubtful under various stages of the test shows the need to provide for a proper legal basis for agencification in the Treaties. In addition, the juxtaposition of the analysis under an institu-tional balance perspective and the analysis following a ‘democracy test’ by AG Jääskinen shows how the latter test does not take into account the prerogatives of the Commission which the Court is held to safeguard. The Court’s decision in Short- selling has however seriously put into doubt the prospects for a genuine principle of institutional balance. The Court confirmed its ongoing practice whereby it seems to consistently refuse to enforce the institutional balance as an actionable principle.

821 Lenaerts, n 819, p 763. 822 Chamon, n 70, p 398.

Conclusion on the Legal Limits 297

8 Conclusion on the Legal Limits

The preceding sections dealt with the legal limits to agencification. In legal doc-trine, authors have mainly focused on the problem of delegation rather than the issues related to the competence of the EU to establish agencies and how EU agen-cies should be assessed under the principles of subsidiarity and proportionality. On the other hand, the (political) institutions seem to have focused not on any of the legal limits but more on their own (political) interests. This, combined with the pragmatism which has driven the agencification of the EU administration, has meant that theory on agencification has largely followed practice.

While almost any legal basis in the EU Treaties would seem to suffice to est-ablish an EU agency, it was found that a more serious scrutiny of the necessity of an EU agency as organizational form should be undertaken by the institu-tions. While recourse to the instrument of the legislative regulation to establish an agency is technically not necessary, it is the sole appropriate instrument in practice.

As regards the principles of subsidiarity and proportionality, the tension be-tween these two and within proportionality was noted. Again here it was found that the institutions should be more considerate to these principles or exercise more self- restraint when establishing and empowering EU agencies.

The discussion of Meroni showed the remarkable dissensus in legal doctrine in relation to that jurisprudence. It was found that Meroni is partially relevant beyond the realm of ‘true delegations’ and that it is also relevant in the post- Lisbon EU context, even if that context, and especially the agencification in that context, militate in favour of a renewed look at Meroni. Indeed, multiple types of delegation were identified in the EU legal order and the delegations in Meroni are not identical to the mechanism used to empower EU agencies. However, the Meroni- light doctrine laid down by the Court in Short- selling went beyond the EU Meroni doctrine proposed here (cf. section IV 5.6.2) and has resulted in a rather ethereal Meroni.

Because the Court engaged in a simplification exercise, Short- selling is only a first step in clarifying the limits to empowering agencies. From a theoretical per-spective, the Court only referred to the essence of the Meroni doctrine. Practically, it remains to be seen how the Court might test the new standard on discretionary powers. In the absence of a clear legal basis to empower agencies it would appear advisable to require the legislator to elaborately design the context in which an agency exercises its powers. Again, however, it may be expected that the Court will defer to the assessment of the legislator. Still, even the new Meroni- light standard is not comparable to the intelligible standard developed by the US Supreme Court (cf. section V 1.3).

The discussion on Romano showed how it is much more than a Meroni- bis but contains in it the problem faced by agencification in the light of the institutional balance, that is, the fact that by empowering an EU agency the legislature intrudes on the prerogatives of the Commission. Following the theoretical discussion on


the institutional balance, a test was devised and applied to the problem in rela-tion to Regulation 236/ 2012, the outcome of which confirmed that agencification in the current primary law context remains highly problematic either because the Treaties do not leave scope for agencification or because the Commission’s prerogatives are not safeguarded where such a scope does exist. In the light of this analysis, the ruling of the Court in Short- selling is remarkable since it ef-fectively relegates Romano to a Meroni- bis. Because of the indisputable link be-tween Romano and the institutional balance, the Court’s relegation of Romano also meant it could completely ignore the institutional balance in Short- selling. Of course, the fact that the Court found Romano to be no longer relevant does not mean that the institutional balance dimension to agencification has become irrelevant. Instead, this problem lingers on and will resurface. Ideally this issue should be settled in the Treaties, where an EU administration (in the broad sense, cf. section I 1.4) for the twenty- first century should be defined.

VControlling Agencification

From the chapters on the legal limits to agencification it follows that this pro-cess should be sufficiently controlled by the institutions, so as to uphold the EU’s values and the general principles of EU law and to safeguard the institutions’ own prerogatives under the Treaties.

To better understand the problem of controlling agencification and because agencification is not a purely EU phenomenon, the summarized findings of a comparative excursion shall first be presented in this chapter. Because the EU itself is a federal- type polity, experiences in two other federal polities will be looked at. The first is the United States of America, which harbours some of the most powerful independent agencies at federal level and where agencies are said to find their origin. The second is the Federal German Republic, which has acted much like a model to the EU in the sphere of implementation of federal legisla-tion, even if the Court of Justice has refuted that possible analogies could have legal consequences in the EU legal order.1 Drawing on the two examples will then help give some background when evaluating the control mechanisms established at EU level. In a final section to this chapter the control exercised by the Court will be looked at in more detail.

1 Agencification in the United States of America

The US being the birthplace of the independent regulatory agency,2 it is fitting to first look at the control over agencies in the US legal system. Just like the EU constitutional charter, the US Constitution does not explicitly address the issue of agencies and instead entrusts executive power to the President.3

Placing the responsibility for the executive branch with the President, the last sentence of Article I, section 8 of the Constitution contains the ‘Necessary and Proper clause’, providing that it is Congress’ responsibility to design the federal government.4

1 See Case C- 359/ 92, Germany v. Council, [1994] ECR I- 3681, para 38.2 See Chapter I, n 27.3 See Article II, section 1 and Article II, section 3 of the US Constitution.4 See Dominique Custos, ‘The Rulemaking Power of Independent Regulatory Agencies’,

(2006) 54 The American Journal of Comparative Law Fall Suplement, p 615.

Controlling Agencification300

1.1 Designing the executive sphere

That clause, according to Chief Justice Marshall in McCulloch v. Maryland, ‘is made in a constitution, intended to endure for ages to come, and consequently, to be adapted to the various crises of human affairs’.5 Given that flexibility, inde-pendent agencies could indeed appear necessary and proper to deliver policy in the current modern society, even if this could not have been foreseen by the draft-ers of the US Constitution in 1787.

While Congress’ power to design the federal government has not been con-tested so much, an outstanding issue related to agencies in the US federal system was the question of where precisely the agencies should be situated in the federal government. As noted in the introductory chapter, the answer was clear for the cabinet and executive agencies. However, the question of the independent agen-cies and whether it is constitutionally sound to grant executive powers to inde-pendent bodies which are not firmly situated under the President’s authority is a different matter. That authority has almost completely been equated with the power to remove officials from office. The Supreme Court held in Myers v. US that the President should hold this power,6 but it may be circumscribed by Congress, which has the power to design the federal government.

1.2 Designing independent agencies vs. the President’s control over the executive

As a result, in Humphrey’s Executor the Supreme Court accepted that Congress could establish ‘independent’ agencies, the principal officers of which are only re-movable ‘for cause’ by the President.7 The Court’s jurisprudence on this issue was then elaborated in cases such as Bowsher v. Synar and Morrison v. Olson,8 and one of the latest rulings in this line of cases is Free Enterprise Fund v. PCBOA. In this case the Court had to rule on a double ‘for cause’ protection, shielding PCBOA members from removal by the Securities and Exchange Commission, the mem-bers of which were in turn protected from removal by the President. The Court found this arrangement contrary to Article II of the US Constitution since the President could no longer ensure the faithful execution of the laws.9 The Court then further linked this with a democracy argument, noting that if the President was incapable of ensuring the law’s execution, the people could not properly pass judgement on the President’s functioning.10

5 US Supreme Court, McCulloch v. State of Maryland, [1819] 17 United States Reports 316, 415.6 US Supreme Court, Myers v. US, [1926] 272 United States Reports 52.7 US Supreme Court, Humphrey’s Executor v. US, [1935] 295 United States Reports 602.

However, there is no unified ‘for cause’ standard: see Kent Barnett, ‘Avoiding Independent Agency Armageddon’, (2012) 87 Notre Dame Law Review 4, p 1352.

8 See US Supreme Court, Bowsher v. Synar, [1986] 478 United States Reports 714; US Supreme Court, Morrison v. Olson, [1988] 487 United States Reports 654.

9 US Supreme Court, Free Enterprise Fund v. Public Company Accounting Oversight Board, [2010] 130 Supreme Court Reporter 3138, 3147.

10 Ibid, 3154– 5.

Agencification in the United States of America 301

1.3 Control over which powers?

The apparent fixation in the US on control over agencies would have the ob-server believe that the question as to which powers may be delegated by Congress to agencies— so fundamental a question in the EU legal order— is uncontested. Of course, this is not true, but it does appear that this matter has largely been settled by legislative practice and before the US Supreme Court. An important preliminary observation here is that the US Supreme Court does not distinguish between Congressional delegations to the President and delegations to agencies.

For instance, that agencies may perform adjudicative functions through the administrative law judges which they house was confirmed in Atlas Roofing,11 and that agencies may combine such quasi- judicial powers with quasi- legislative powers was confirmed in Humphrey’s Executor. Inter alia in Union Bridge Co. v. US the Court had already ruled that Congress could delegate some legislative power, subject to it having fixed a primary standard, in which case the delegate authority held ‘the mere executive duty to effectuate the legislative policy declared in the statute’.12 From an EU perspective, the analogy in words between Union Bridge and the CJEU’s ruling in Meroni is evident, as the emphasis in both cases is on the ‘mere executive’ duties or powers of the delegate authority.

The Court later relaxed its standard and held the delegation of legislative power to be permissible, as long as Congress had laid down an ‘intelligible principle’13 which can be worded very vaguely or openly.14 Only on two occasions,15 in 1935, did the Court strike down a Congressional delegation as being insuffi-ciently precise.16 Justice White in Chadha concluded that despite the early con-servative cases,17 ‘restrictions on the scope of the power that could be delegated diminished and all but disappeared’.18 Following the two 1935 cases, the Court

11 See US Supreme Court, Atlas Roofing Co., Inc. v. Occupational Safety Commission, [1977] 430 United States Reports 442, 455.

12 US Supreme Court, Union Bridge Co. v. US, [1907] 204 United States Reports 364, 385 (em-phasis added). The Court applied this to an independent agency in US v. Chicago: see US Supreme Court, United States v. Chicago, [1931] 282 United States Reports 311, 324.

13 US Supreme Court, Hampton & Co. v. US, [1928] 276 United States Reports 394, 409.14 US Supreme Court, New York Central S. Corp. v. United States, [1932] 287 United States

Reports 12, 25; US Supreme Court, Federal Radio Commission v. Nelson Bros. B. & M. CO., [1933] 289 United States Reports 266, 285.

15 In Clinton v. City of New York, decided in 1998, the US Supreme Court struck down the President line- item veto under the Line Item Veto Act of 1996. The Act gave a power to the President to change the text of duly enacted statutes by cancelling certain budget appropriations, which amounted to the creation of a new procedure to pass legislation, not provided for under the Constitution. However, Calabresi has argued that this case was not about legislative procedure but about a delegation of powers from Congress to the President. See Steven Calabresi, ‘Separation of Powers and the Rehnquist Court: The Centrality of Clinton v. City of New York’, (2004) 99 Northwestern University Law Review 1, pp 85– 6. Justice Scalia, writing for the minority, indeed also suggested as such. See US Supreme Court, Clinton v. City of New York, [1998] 524 United States Reports 417, 465. Similarly, see Justice Scalia’s dissent in Mistretta, US Supreme Court, Mistretta v. United States, [1989] 488 United States Reports 361, 416– 22.

16 US Supreme Court, Panama Refining Co. v. Ryan, [1935] 293 United States Reports 388; US Supreme Court, A.L.A. Schechter Poultry Corporation v. United States, [1935] 295.

17 See n 12.18 US Supreme Court, INS v. Chadha, [1983] 462 United States Reports 919, 985.


soon reaffirmed its limited scrutiny of delegations as long as Congress provided some standard guiding the delegate authority.

In addition, judicial review now intervenes at the stage where a delegated power is exercised, scrutinizing the delegated act in the light of the Constitution and the delegating statute19 rather than scrutinizing the delegating statute in the light of the Constitution,20 further in line with the general rule of constitutional avoidance.21

1.4 Interpretation of agencies’ statutory mandates

Since the focus has shifted to the legality of agencies’ actions under their own mandate (and ultimately the Constitution), the question of interpretation of those mandates has gained importance. Although it is ‘the province and duty of the judi-cial department to say what the law is’,22 the Court soon recognized that agencies may also be competent to interpret the law. In Skidmore, the Court accepted that it ought to defer to an agency interpretation in so far as that interpretation does not relate to a pure question of law and in so far as it has the power to persuade.23

In Chevron, however, the Court introduced a new test, and in doing so abolished the importance of the distinction between pure questions of law and the applica-tion of law.24 The underlying reasoning of the Court’s ruling has been identified as resting on (the traditional idea of) agency expertise, the greater accountability of (executive) agencies compared to politically unaccountable judges,25 and the idea

19 See François van der Mensbrugghe, ‘The Danger of Excessive Delegations to Independent Administrative Agencies: the Example of the U.S.A.’, in Della Cananea (ed), European Regulatory Agencies, Paris, ISUPE, 2005, p 118.

20 See for example Brown & Williamson, where the Food and Drug Administration’s competence to regulate tobacco products was at issue. The FDA, against its previous position, argued that it had competence to regulate. The US Supreme Court held that the FDA was not competent. First, the FDA had considered tobacco to be unsafe, but did not ban it going against the regulatory scheme of the FDCA. Secondly, since Congress had adopted tobacco- specific legislation it had confirmed the FDA lacked competence to regulate tobacco products. Lastly, the Court held that given the cul-tural and economic importance of tobacco, it could not be assumed that Congress had implicitly granted competence to the FDA. See US Supreme Court, FDA v. Brown & Williamson Tobacco Corp., [2000] 529 United States Reports 120, 143, 156, 160.

21 See US Supreme Court, United States v. Delaware & Hudson Co., [1909] 213 United States Reports 366, 408– 9. The Court held: ‘It is elementary when the constitutionality of a statute is assailed, if the statute be reasonably susceptible of two interpretations, by one of which it would be unconstitutional and by the other valid, it is our plain duty to adopt that construction which will save the statute from constitutional infirmity.’ For an application, see US Supreme Court, Industrial Union Department, AFL- CIO v. American Petroleum Institute, [1980] 448 United States Reports 607, 646.

22 US Supreme Court, Marbury v. Madison, [1803] 5 United States Reports 137, 177.23 US Supreme Court, Skidmore v. Swift & Co., [1944] 323 United States Reports 134, 140. See

also Matthew Krueger, ‘In Search of the Modern Skidmore Standard’, (2007) 107 CLRev 6, p 1239.

24 See Linda Jellum, ‘The Impact of the Rise and Fall of Chevron on the Executive’s Power to Make and Interpret Law’, (2012) 44 Loyola University Chicago Law Journal 1, p 165.

25 Farina has described this as a general ‘separation of powers’ rationale which also in-cluded the idea that the judiciary should not aggrandize its own powers when Congress


that Congress chooses to delegate power by drafting ambiguous statutes.26 The Court therefore ought to defer to an agency (i) if the statute is unclear and (ii) if the agency’s interpretation is reasonable.

Yet in its subsequent jurisprudence27 the straightforward Chevron test was fur-ther complicated, much to the dismay of Justice Scalia,28 whereby a Chevron step zero29 was introduced in which a court should first check whether an agency in-terpretation meets the conditions for Chevron deference. After all Chevron itself dealt with (i) an interpretation having the force of law,30 (ii) adopted following a notice- and- comment procedure under the Administrative Procedure Act, (iii) by an executive agency.31 In cases such as MCI and Brown & Williamson,32 the Court further seems to have introduced a non- delegation dimension to its Chevron step zero, whereby Congress cannot be presumed to have had the intent to delegate ‘major questions’ to agencies, since major questions should be reserved to the legislator.33

Chevron and its ‘domain’34 remain a contentious topic in academic doctrine in the US and in the jurisprudence of US federal courts. However, the fact that the US Supreme Court mitigated part of its Chevron ruling in its later cases should not obfuscate the remarkably strong position which agencies hold in the US fed-eral system and the rather large gap which exists between how the ‘administrative

intentionally leaves ambiguous provisions in its statutes for agencies to fill in. See Cynthia Farina, ‘Statutory Interpretation and the Balance of Power in the Administrative State’, (1989) 89 CLRev 3, p 466.

26 David Gosset, ‘Chevron, Take Two: Deference to Revised Agency Interpretations of Statutes’, (1997) 64 University of Chicago Law Review 2, pp 688– 90.

27 See US Supreme Court, Christensen v. Harris County, [2000] 529 United States Reports 576.28 See inter alia US Supreme Court, US v. Mead Corp., [2001] 533 United States Reports 218, 245;

US Supreme Court, Barnhart v. Walton, [2002] 535 United States Reports 212, 226– 7.29 Thomas Merril and Kristin Hickman, ‘Chevron’s Domain’, (2001) 89 Georgetown Law

Journal 4, p 873.30 However, what should be understood by the force of law is not clear and in Christensen but

especially in Mead the Court further complicated the issue.31 The deference in Chevron was not only based on the agency’s expertise but also on its pol-

itical accountability, coming under the authority of the President. Evidently this second element applies much less to the independent agencies but the US Supreme Court has still afforded Chevron deference to independent agencies. For a discussion, see Randolph May, ‘Defining Deference Down: Independent Agencies and Chevron Deference’, (2006) 58 Administrative Law Review 2, pp 429– 53.

32 In MCI, the US Supreme Court ruled that where an agency has the authority to ‘modify’ a section of the basic statute, this cannot be understood as granting the power to fundamentally alter that section, since ‘modifying’ is commonly understood as (only) ‘changing slightly or incremen-tally’. See US Supreme Court, MCI Telecommunications Corp. v. American Telephone and Telegraph Co., [1994] 512 United States Reports 218, 225– 6. In Brown & Williamson (see n 20), the Court held that it could not be presumed that Congress implicitly delegated authority over tobacco products to the FDA, given the significance of the tobacco sector in US economy and culture.

33 However, Sunstein has proposed to include these questions in Chevron step One, since this would otherwise mean that courts become competent to solve major questions for which they have no superior claim than the executive branch. See Cass Sunstein, ‘Chevron Step Zero’, (2005) 92 Virginia Law Review 2, p 243.

34 See Merril and Hickman, n 29.


state’ in the US functions today and how one would expect it to function if one were to be guided solely by the provisions of the US Constitution.

1.5 US agencies’ constitutionality

Bringing together the insights on (i) Congress’ power to establish new agen-cies, (ii) the almost complete evaporation of the non- delegation doctrine in the Supreme Court’s jurisprudence, and (iii) the deference (Chevron or otherwise) granted to agencies’ statutory interpretations, it is possible to come to a general understanding of (independent) agencies’ constitutionality.

Steele and Bowman have indeed sought to argue the constitutional legitimacy of independent agencies in the light of the US Constitution and the Supreme Court’s jurisprudence cited previously. Both authors reject the idea of a rigid sep-aration of powers, often advanced by proponents of the unitary executive theory. Thus, under the separation of powers, it is not necessary for every public authority exercising enforcement powers to be wholly and exclusively brought under the executive branch and thereby under the President’s authority. Rather, Congress has the competence to design the federal government so that the President is ena-bled to ‘faithfully execute the laws’. If an independent agency is ‘necessary’ and ‘proper’ to ensure this, Congress may establish one, and does not act against the Constitution by limiting the President’s power of removal.35

To Steele and Bowman, agencies, despite being ‘independent’ and perhaps not fitting with the traditional notion of separation of powers, are still ‘under control’. After all, it is Congress which establishes the agency and at the same time lays down its powers. Through the budget procedure, the House and Senate have the opportunity to scrutinize agency action and to influence future policy by pro-viding for conditions in the budget. The commissioners in independent agencies being principal officers of the United States, their appointments are also con-firmed by the Senate. Finally, Congress always has the opportunity to use the nuclear option, that is, to abolish an agency. Less drastically, it may also review the mandate of an agency.36

The agencies are further controlled by the executive branch, since the President nominates the officers of the United States and has power to remove them, al-though in the case of the independent agencies this power is circumscribed. The Office of Management and Budget, which is part of the executive office of the President of the United States, is competent to review and revise agencies’ budget requests and legislative recommendations.37 Finally, most agencies are dependent

35 See Charles Steele and Jeffrey Bowman, ‘The Constitutionality of Independent Regulatory Agencies under the Necessary and Proper Clause: The Case of the Federal Election Commission’, (1987) 4 Yale Journal on Regulation 2, pp 377– 8. On the evolution of the President’s power of removal see ibid, pp 388– 91.

36 Ibid, pp 384– 5.37 The Reagan Administration expanded the OMB’s control over the agencies through Executive

Orders 12291 and 12498. See Damien Geradin, ‘The Development of European Regulatory Agencies: What the EU Should Learn from the American Experience’, (2004) 11 CJEUL 1,


upon the Justice Department for their litigation before courts.38 As regards the judicial branch, it is clear that agencies are not an exception to the rule of law. As a result, courts can review agency action, even if judges will often defer to an agency’s assessment, given its expertise on its subject matter.

Steele and Bowman thus conclude that independent agencies may be neces-sary and proper; that, despite their independence, they are sufficiently controlled; and that rather than confusing the separation of powers, they strengthen it by introducing checks on the executive.39 Yet, from the perspective of the present study, Steele and Bowman’s analysis feels incomplete, since it does not take into account the powers conferrable on agencies. Indeed, by arguing that Congress has the power to establish and empower agencies and by concluding a sufficient degree of ongoing and ex post control over agencies’ actions is retained, both authors seem to suggest that everything in between these two steps is of little importance.

How may this be explained? As noted above in the discussion of the US Supreme Court’s jurisprudence, the Court has laid down a very generous standard by which to gauge Congressional authorizations. If this standard is respected, Congress has not delegated its own powers (which is impermissible) but has only authorized an agency to enforce Congress’ legislation.40 Of course this difference appears rather semantic,41 and the US Supreme Court’s juris-prudence on this issue appears inspired by a large deal of pragmatism. It is this pragmatic view which explains the far- reaching powers delegated to agencies.42 To this pragmatic view is added what can only be described as a large degree of confidence (albeit not unconditional) by the Court in the two other branches of government and a genuinely balanced system of checks and balances. The US federal system and its institutions seem to have acquired the necessary maturity allowing for the creation and empowerment of independent agencies without at the same time risking the subversion of the balance between the three branches established in the Constitution. Indeed, potential shocks caused by the delega-tion of powers to agencies and the exercise of such powers are being absorbed, the traditional three branches having acquired the necessary (political) strength and authority.

p 44. These Executive Orders were replaced by Executive Order 12866, subsequently amended by Executive Order 13497. Shapiro notes that there is uncertainty as to whether the US independent agencies are legally required to submit their legislative recommendations to the OMB, even if they usually do: see Martin Shapiro, ‘A Comparison of US and European Independent Agencies’, in Rose- Ackerman and Lindseth (eds), Comparative Administrative Law, Cheltenham, Edward Elgar, 2010, p 296.

38 See U.S.C. Title 28, Part II, Chapter 31, § 516.39 On this last issue, see Steele and Bowman, n 35, p 386.40 See Jellum, n 24, p 157.41 See ibid; US Supreme Court, Whitman v. American Trucking Associations, Inc., [2001] 531

United States Reports 457, 472– 6.42 See for instance the Court’s observation in Mistretta: ‘[O] ur jurisprudence has been driven by

a practical understanding that in our increasingly complex society […] Congress simply cannot do its job absent an ability to delegate power.’ See Mistretta v. United States, n 15, 372.


Without anticipating the subsequent chapter, it may already be noted, in so far as is necessary, that the EU differs considerably from the US on this point. The EU is hardly a mature polity. The status of its constitutional charter is in no way comparable to that of the US Constitution. Rather, it is the subject of an almost permanent diplomatic conference, with five major revisions in the past thirty years. As a result, the positions of the EU’s main institutions are not settled.

A second important difference between the US and EU systems is the manner of the application and enforcement of federal legislation. This was already noted in the introductory chapter and it is fundamental in a comparison of the agenci-fication processes in the US and EU, even if this is ignored in some comparative analyses.43

In the US Supreme Court’s jurisprudence there are strands of cases related to the separation of powers and related to federalism. Both strands may at times come together in a single case, or sometimes it may be debated under which strand a single case should be situated.44 Generally, however, the ques-tion of delegation of powers to agencies firmly forms part of the separation- of- powers jurisprudence. Simply put, in the federal system of the US, a decision by Congress to grant powers to an independent agency is an affair exclusive to the federal level, and this decision generally does not have a vertical dimension linking the federal to the state level. As was explained, this is because in some federal systems, such as that of the US, the federal level has its own administra-tion. The EU however is characterized by a Vollzugsföderalismus and, as a result, the question of establishing and empowering a new federal administrative body is not a matter of exclusive interest to the federal level, but is intrinsically linked to the vertical order of competences and thus to the Member States’ compe-tences and interests.

2 Agencification in the Federal Republic of Germany

In light of the last remark in the preceding section, the case of the Federal Republic of Germany is interesting, since the EU shares with Germany the logic of Vollzugsföderalismus. The question of implementation and enforcement of le-gislation by the Bund then also carries in it a vertical dimension, raising issues of competence of Bund and Länder. At the same time, it should be noted that,

43 See Geradin, n 37; van der Mensbrugghe, n 19; Christian Stoffaës, ‘Foreword to the Volume’, in Della Cananea (ed), European Regulatory Agencies, Paris, Éditions Rive Droite, 2005, pp 137– 49; Miroslava Scholten, The Political Accountability of EU Agencies: Learning from the US Experience, Maastricht, Universitaire Pers Maastricht, 2014. Shapiro does seem to note this issue: ‘Member States administrations implement most EU legislation. In contrast to the US, few EU agencies directly administer regulatory or other programs.’ See Shapiro, n 37, p 300.

44 Examples of such cases already mentioned are Clinton v. New York and US v. Printz. See n 15 and Chapter I, n 267.

Agencification in Germany 307

unlike in the EU, hybrid forms of administration in between direct and in-direct administration are, as a rule, not tolerated by the German Basic Law: ‘The Basic Law also precludes, apart from limited exceptions, a so called mixed administration.’45

2.1 German agencies

The German Grundgesetz (GG),46 unlike the US or EU constitutional charters, con-tains a rather elaborate section on ‘The Execution of Federal Laws and the Federal Administration’.47 The Federal Authorities (selbständige Bundesoberbehörden) are the closest thing to agencies which may be found in the German system.48 Looking at the practice of the multilevel EU administration and the national agencies49 which engage in the Boards of EU agencies or the networks set up by them, Germany is indeed mostly represented by such selbständige Bundesoberbehörden, although the body representing Germany in the European Supervisory Authorities (ESAs) is an Anstalt.50 The identification of these bodies is rather straightforward precisely because of the elaborate section on execution of federal legislation in the Grundgesetz.

2.2 The relation between Bund and Länder

As Ziller points out, the Vollzugsföderalismus in the EU is elaborated only in a single Article (ie 291 TFEU), whereas in the Grundgesetz an entire section (made up of nine Articles) spells out the modalities of the implementation of federal legis-lation or Bundesverwaltung.51 Article 83 GG lays down the general rule similar to Article 291(1) TFEU by providing that ‘[t] he Länder shall execute federal laws in their own right insofar as this Basic Law does not otherwise provide or permit’.52 Article 84 GG provides the rules for the execution of federal legislation by the Länder ‘in their own right’, while Article 85 GG provides for the Länder execution

45 See Bundesverfassungsgericht, Hartz IV- Arbeitsgemeinschaften, [2007] 119 BVerfGE 331, 365. Translated from German by the author.

46 The German Parliament has put online an English translation of the Grundgesetz: see https:// www.btg- bestellservice.de/ pdf/ 80201000.pdf.

47 This is section VIII of the GG.48 Kristian Fischer, ‘Quangos— An Unknown Species in German Public Law? German

Report on the Rulemaking Power of Independent Administrative Agencies’, in Riedel and Wolfrum (eds), Recent Trends in German and European Constitutional Law, Berlin, Springer, 2006, pp 159– 60.

49 Often of course it will be the ministries and regular departments themselves who are involved in an EU agency.

50 In the ESMA, EBA, and EIPOA Germany is represented by a federal institution (Anstalt); in the EMA, ERA, ACER, BEREC, and ENISA it is represented by federal higher authorities (Bundesoberbehörde).

51 Jacques Ziller, ‘Multilevel Governance and Executive Federalism: Comparing Germany and the European Union’, in Birkinshaw and Varney (eds), The European Union Legal Order after Lisbon, Alphen Aan Den Rijn, Wolters Kluwer, 2010, p 265.

52 Article 30 GG generally provides: ‘Except as otherwise provided or permitted by this Basic Law, the exercise of state powers and the discharge of state functions is a matter for the Länder.’


of ‘federal commission’ which is only possible in those (material) fields expressly mentioned in the Basic Law.53 These Articles will not be further discussed be-cause the provisions which are immediately relevant here are those dealing with the Bundeseigene Verwaltung, that is, implementation of federal legislation by the Bund itself, provided for in Article 86 GG.54

The two bodies (selbständige Bundersoberbehörde and Anstalten)55 identified above are then mentioned in Article 87 GG, which is reproduced in abridged form here:

1) The foreign service, the federal financial administration, and […] the administration of federal waterways and shipping shall be conducted by federal administrative authorities […]

2) Social insurance institutions whose jurisdiction extends beyond the territory of a single Land shall be administered as federal corporations under public law […]

3) In addition, autonomous federal higher authorities as well as new federal cor-porations and institutions under public law may be established by a federal law for matters on which the Federation has legislative power. When the Federation is confronted with new responsibilities with respect to matters on which it has legislative power, federal authorities at intermediate and lower levels may be estab-lished, with the consent of the Bundesrat and of a majority of the Members of the Bundestag, in cases of urgent need.56

Although only Article 87(3) first sentence GG is directly relevant here, other parts of the Article were reproduced to make clear that the provisions of para-graph 3 are applicable ‘in addition’ (Außerdem) to the rules laid down in the two preceding paragraphs. Because of this, paragraph 3 deals with the so called fakultative Bundesverwaltung, either through the federal higher authorities (Bundesoberbehörden), making it fakultative unmittelbare Bundesverwaltung; or through the federal corporations and institutions under public law (Körperschaften und Anstalten des öffentliches Rechts), making it fakultative mit-telbare Bundesverwaltung.57

2.2.1 Article 87(3) first sentence GG: An unlimited enabling clause?Because Article 87(3) first sentence GG contains an optional course of action within the exceptional category of Bundeseigene Verwaltung, it should not come as a surprise that this provision has foremost been studied from a federalist perspec-tive. In the Kreditwesen case, the Constitutional Court found that this provision

53 For a more elaborate discussion of these two types see Ziller, n 51, pp 268– 74.54 The Bundeseigene Verwaltung therefore is one possible form of Bundesverwaltung, next to the

Bundesverwaltung foreseen in Articles 84 and 85 GG.55 When Helfritz worked out a framework for EU agencies (cf. section VI 2), he noted that in-

spiration could be drawn from Article 87(3) first sentence GG. See Vark Helfritz, Verselbständigte Verwaltungseinheiten der Europäischen Union, Berlin, Weissensee Verlag, 2000, p 182.

56 The first sentence of paragraph three has been put in bold for emphasis, since it is this provi-sion in Article 87 GG which provides for the bodies under scrutiny here.

57 Mittelbar reflects the separate legal personality of the bodies concerned.


gives the competence (to the Bund) not only to establish new bodies, but also to entrust them with tasks for which the Basic Law does not explicitly provide that they may come under Bundesverwaltung.58 It is difficult to overestimate the imp-ortance of this qualification. As Kalkbrenner noted at the time, the areas in which the Bund holds legislative competence (see Articles 73 and 74 GG) are vast and the Court’s ruling allowed the federal level to establish its own administration in these fields as it sees fit. Combined with a strong centripetal force in the German fed-eracy, Kalkbrenner predicted that the logic of Article 83 GG could be reversed,59 since the use of this provision by the Bund precludes the further competence of the Länder in the field affected,60 as confirmed by the Constitutional Court in Moratorium Gorleben.61

Because of the alleged lack of real constraints on the use of the possibility pro-vided for by Article 87(3) first sentence GG, Papier has described the provision as a Trojan horse to the general rule of Article 83 GG.62

Indeed, there are few limits to the use of the provision. Notwithstanding that only the legislator (and not the executive) may establish a Bundesoberbehörde or an Anstalt, the federal law by which these bodies are established does not re-quire the consent of the Bundesrat63 and may be adopted by the Bundestag alone (contrary to the bodies provided for in Article 87(3) second sentence GG).64 As a result, the Länder lack control not just when new federal authorities are estab-lished, but also when such authorities are granted powers which are traditionally Länder powers. For this reason, ten of the sixteen Länder proposed, albeit unsuc-cessfully, in the 1990s to amend the provision to the effect that the Bundesrat’s consent would be required to establish a Bundesoberbehörde or Anstalt.65 In par-allel with the exercise by the Bund of its concurrent competences, which is subject to the Erforderlichkeitsklausel, it has been argued that the establishment of a new federal body should be subject to the condition that it is necessary (erforderlich) because existing Bund- and Länderbehörden cannot adequately fulfil the tasks concerned.66 However, again in the Kreditwesen case, the Constitutional Court rejected such a reading of Article 87(3) first sentence GG, inter alia observing

58 Bundesverfassungsgericht, Kreditwesen, [1962] 14 BVerfGE 197, 210.59 Helmut Kalkbrenner, ‘Zur Errichtung von Bundesoberbehörden nach Art. 87 III GG’,

(1963) 18 JZ 7, p 211.60 Dieter Hömig, Grundgesetz für die Bundesrepublik Deutschland, Baden- Baden, Nomos,

2010, p 515.61 See Bundesverfassungsgericht, Moratorium Gorleben, [2001] 104 BVerfGE 238, 247.62 Hans- Jürgen Papier, ‘Die Regionalisierung der gesetzlichen Rentenversicherung aus verfas-

sungsrechtlicher Sicht’, (1995) 4 Neue Zeitschrift für Sozialrecht 6, p 242.63 Germany’s bicameral federal parliament is composed of a Bundestag, directly elected by the

German people, and the Bundesrat, which is composed of the governments of the Länder, much like the Council of Ministers in the EU.

64 See Martin Ibler, ‘Artikel 87 GG’, in Maunz and Dürig (eds), Grundgesetz: Kommentar, München, Beck, Ergänzungslieferung 64 2012, p 209.

65 Article 105(1) of the original draft of the Basic Law prescribed the consent of the Bundesrat for laws establishing Bundesoberbehörden but never made it to the final text of the Basic Law. On this and on the proposed amendment in the 1990s see ibid, pp 198– 201.

66 See the authors cited by Ibler, n 64, pp 210– 11.


(under an a contrario reasoning) that only Article 87(3) second sentence GG refers to ‘cases of urgent need’.67 Lastly, while Article 87(3) first sentence GG prescribes that the agencies be established through law (durch Bundesgesetz), their further empowerments do not require legislative action and may be done by an executive Rechtsverordnung (under Article 80(1) GG).68

Only one limit was further read into Article 87(3) GG by the Constitutional Court and prescribes that a Bundesoberbehörde or an Anstalt may only be est-ablished for the purpose of entrusting them with tasks which, out of their nature, may be practically executed centrally (at federal level) for the entire territory of the Bund without needing to rely on further subordinate author-ities (Mittel- and Unterbehörden) and without needing to rely on the Länder authorities.69

Reicherzer has identified a weak point in this limit, which greatly resembles the logic of the weakness in the application of the subsidiarity principle at EU level identified by Van Nuffel (cf. section IV 2.2). According to Reicherzer the Bund is free to devote as many resources to a Bundesoberbehörde as it chooses and it can therefore circumvent the limit, since any task may be effectively performed at cen-tral level if the resources devoted to that task are unlimited.70 Thus, just as the EU legislator can set as ambitious EU objectives as it wants (affecting the question of whether Member States may attain these objectives by themselves), the Bund may devote as many resources as it wants to a Bundesoberbehörde (affecting the ques-tion of which tasks a Behörde may practically undertake).

2.3 German agencies and the separation of powers

As noted, the question of establishing and empowering Bundesoberbehörden and Anstalten has foremost received attention under a federalism rather than a separation- of- powers perspective. Here again the detail of the German Basic Law should be noted. In its Article 80(1) it explicitly provides that the legislator may authorize the government or a minister to adopt ‘statutory instruments’ provided that the legislator also specifies the content, purpose, and scope of the

67 Kreditwesen, n 58, 213– 14.68 The Bundesverwaltungsgericht has noted that an Anstalt could be empowered either through

(i) the law originally establishing the Anstalt, (ii) a subsequent law, or (iii) a Rechtsverordnung. See Bundesverwaltungsgericht, [1996] 102 BVerwGE 119, 126. Appel and Eding conclude the same reasoning applies for Bundesoberbehörden. In their study, they also noted that the Rechtsverordnung in casu could only be adopted following the Bundesrat’s consent, finding that the prerogatives of the Länder were sufficiently ensured. Indeed this would appear to be the case, much more so than when a Bundesoberbehörde is empowered by a law under Article 87 (3) first sentence GG. See Markus Appel and Annegret Eding, ‘Verfassungsrechtliche Fragen der Verordnungsermächtigung des § 2 II NABEG’, (2012) Neue Zeitschrift für Verwaltungsrecht 6, p 344.

69 See inter alia Bundesverfassungsgericht, Zollkriminalamt, [2004] 110 BVerfGE 33, 49; Kreditwesen, n 58, 211.

70 Max Reicherzer, ‘Bundesoberbehörden: Trojanische Pferde für den Föderalismus? Zur Verfassungsmäßigkeit der Zuständigkeitsbestimmungen im TEHG’, (2005) Neue Zeitschrift für Verwaltungsrecht 8, p 877.


power conferred. This is indeed very similar to Article 290 TFEU but, unlike in the EU, the Grundgesetz further provides that if the legislator so provides, the government may subdelegate this power to another body. The lack of attention under a separation- of- powers perspective may also be explained by the fact that the Bundesoberbehörden, despite being qualified as selbständig, and the Anstalten, despite their separate legal personality, are still subject to the supervision of the highest federal authorities (ie the ministries) who may also still issue direc-tives.71 Because of this, any analogy with the American federal agencies discussed above other than their qualification as ‘regulatory authority’ has traditionally been rejected.72 In recent years, however, Ludwigs notes two trends which have strengthened the independence of Bundesoberbehörden. Specifically as concerns the Bundesnetzagentur, he notes pressures from EU legislation which requires the Member States to establish increasingly independent regulatory authorities (cf. section III 3.1)73 and, subsequently, a more progressive interpretation of this new legislation by the German administrative courts.74

Under German public law, the strengthened independence of regulatory auth-orities, imposed by EU law, relates with difficulty to (and ultimately confronts) the principle of democracy.75 This was also apparent in Commission v. Germany, which dealt with the EU requirement that national data supervisors act ‘with complete independence’. The German Länder had implemented the EU Directive in such a way that a supervisor was still subject to scrutiny by its Land govern-ment, which the Commission deemed incompatible with the Directive. In its pleadings, Germany had invoked the Demokratieprinzip (discussed later) and its traditional system of government supervision. As regards the Demokratieprinzip invoked by Germany, the Court followed an EU, rather than a national, interpret-ation of this principle and noted that the ‘principle does not preclude the existence of public authorities outside the classic hierarchical administration and more or

71 Ibler, n 64, pp 218– 20.72 See for instance Matthias Ruffert, ‘Regulierung im System des Verwaltungsrechts—

Grundstrukturen des Privatisierungsfolgerechts der Post und Telekommunikation’, (1999) 124 AöR 2, p 277 at footnote 176.

73 Halberstam also notes that the constitutional and interinstitutional dynamics in Germany have prevented independent agencies from being established in the same way as in other coun-tries (cf. Chapter III, at n 42) and that the birth of the Bundesnetzagentur followed from EU law requirements. See Daniel Halberstam, ‘The Promise of Comparative Administrative Law: A Constitutional Perspective on Independent Agencies’, in Rose- Ackerman and Lindseth (eds), Comparative Administrative Law, Cheltenham, Edward Elgar, 2010, p 197.

74 Markus Ludwigs, ‘Die Bundesnetzagentur auf dem Weg zur Independent Agency? Europarechtliche Anstöße und verfassungsrechtliche Grenzen’, (2011) 44 Die Verwaltung 1, p 41.

75 Of course this is not only so in Germany. The independence of the national energy regulator of Belgium has for instance also been questioned before the Constitutional Court in the light of the principles of democracy and separation of powers. The Court, in unconvincing manner, upheld the regulator’s independence, inter alia by referring to the primacy of EU law. See Merijn Chamon, ‘Verzelfstandiging in de nationale en Europese rechtsordes: nieuwe uitdaging van een meergelaagde administratie’, (2013) TBP 2– 3, pp 117– 18. For a discussion of the requirement of an independent energy regulator under Polish public law, see Waldemar Hoff, ‘The EU Model of Regulatory Authority and the Polish Constitutional System’, (2005) 9 Yearbook of Polish European Studies, pp 47– 66.


less independent of the government’.76 The Court also emphasized that the inde-pendence in question only related to the supervisor’s relationship with the execu-tive and not with the legislator. Still, one year earlier, the Court, in another case between the Commission and Germany, had sanctioned the possibility for EU legislation to grant powers to NRAs barring the national legislator from predefin-ing the NRA’s policy beyond what was laid down in EU legislation.77

The traditional German conception of the Demokratieprinzip is based on Article 20(2) GG, which provides that ‘[a] ll state authority is derived from the people. It shall be exercised by the people through elections and other votes and through specific legislative, executive and judicial bodies’. From this pro-vision, the German Constitutional Court has developed important jurispru-dence elaborating the Demokratieprinzip.78 The exercise of public power is then only legitimate (ie democratic) if there is an uninterrupted chain of legitimacy (ununterbrochene Legitimationskette) going from the people to the wielders of public power. In Mitbestimmungsgesetz Schleswig- Holstein the Court clarified this by reference to (i) the personal democratic legitimacy of public officials and (ii) the substantive legitimacy sanctioned and scrutinized by the people or Parliament.79

As Wiedemann points out, the Court’s insistence in Mitbestimmungsgesetz Schleswig- Holstein on the ultimate responsibility, and therefore authority, of the relevant government minister was heavily criticized and may have led to its taking a more progressive approach in Lippeverband,80 in which the Constitutional Court noted that the Demokratieprinzip should be open to evolution.81

This flexibility could help to reconcile (national) agencification with the Demokratieprinzip. After all, as Ludwigs notes, the requirement that the exercise of public power is substantively legitimate becomes problematic if independent (in the American sense) regulatory authorities are vested with powers.82

Because of this, it may be assumed that the issue of control— so central in the discussion on the US federal agencies— over Bundesoberbehörden or Anstalten will be the subject of a new debate in Germany, where it has lain dormant for many years because of the well- established hierarchical supervision by the ministries. This is so because the latter model is incompatible with the model promoted by the EU legislator, which emphasizes genuine independence— or more US- style independence, as Ludwigs would say— for regulatory authorities.

76 Case C- 518/ 07, Commission v. Germany, [2010] ECR I- 1885, para 42.77 See Case C- 424/ 07, Commission v. Germany, [2009] ECR I- 11431.78 Unsurprisingly, the Court of Justice has been criticized by German authors for its lack

of attention to the repercussions of its ruling in Short- selling on the Demokratieprinzip: see inter alia Christoph Ohler, ‘Rechtsetzungsbefugnisse der Europäischen Wertpapier- und Marktaufsichtsbehörde (ESMA)’, (2014) 69 JZ 5, pp 250– 2.

79 See Bundesverfassungsgericht, Lippeverband, [2002] 107 BVerfGE 59, 87– 8; Bundesverfassungsgericht, Mitbestimmungsgesetz Schleswig- Holstein, [1995] 93 BVerfGE 37, 67.

80 Richard Wiedemann, ‘Unabhängige Verwaltungsbehörden und die Rechtsprechung des BVerfG’, in Masing and Marcou (eds), Unabhängige Regulierungsbehörden: Organisationsrechtliche Herausforderungen in Frankreich und Deutschland, Tübingen, Mohr Siebeck, 2010, pp 46– 7.

81 Lippeverband, n 79, 91. 82 Ludwigs, n 74, p 47.


As the section on the US agencies has shown, the ultimate authority of an agency lies not only in the legislative mandate, but also in the degree of judicial control exercised by the (administrative) judge and the deference granted to the agency. In Germany, this question, just like an agency’s independence, is linked to the Demokratieprinzip, since a restricted judicial control affects the substantive legitimacy of an agency’s decision.83

2.4 Deference to agencies’ decisions

As a preliminary remark, it should be noted that the framework to assess the executive’s discretion in Germany is a bit more complicated than that in the US or in the EU. In Germany, a first major distinction is made between Tatbestand and Rechtsfolge, or discretion on whether the facts in a specific case meet the conditions set out in the norm and discretion on the legal effects. The former encompasses the definition of unbestimmten Rechtsbegriffen, that is, legal concepts which have not been sufficiently defined by the legislator. Here the courts exercise a full review of the executive’s decision under the ideal of the single correct decision. Exceptionally, the executive has a Beurteilungspielraum, that is, a margin of appreciation, which is the first type of possible execu-tive discretion. The other two traditional types of discretion fall under the aspect of the Rechtsfolge and are the allgemeine Verwaltungsermessen and the Planungsermessen, that is, the general administrative discretion and planning discretion.

At the same time, German administrative law, just like that of the other Member States, has not been immune from the influence of EU legislation, even if national administrative autonomy is the rule and EU intervention remains the exception in this area. Following pressures from EU legislation, the Federal Administrative Court has created a new type of discretion:84 the regulatory discretion which ignores the distinction between Tatbestand and Rechtsfolge.85

An important case in this regard is Arcor, in which the administrative court of Cologne had to rule on a decision of the Bundesnetzagentur which fixed certain rates which Deutsche Telekom could charge to Arcor for access to its infrastruc-ture. One of the questions which the administrative court referred to the Court of Justice was whether the Bundesnetzagentur was ‘entitled, when assessing cost- orientation in the context of its authority under [the relevant EU legislation], to a “margin of discretion” which is subject only to limited judicial control?’86 In its answer, the Court confirmed that NRAs have been granted broad discretion

83 Martin Eifert, ‘Die gerichtliche Kontrolle der Entscheidungen der Bundesnetzagentur’, (2010) 174 Zeitschrift für das gesamte Handelsrecht und Wirtschaftsrecht 4, p 455.

84 However, Ludwigs notes that EU legislation does not require national judges to be deferential to regulatory authorities, as is also apparent from the discussion of Arcor above, since it only deals with the NRA’s relationship with the legislator. See Ludwigs, n 74, pp 68– 9.

85 See Joachim Wieland, ‘Regulierungsermessen im Spannungsverhältnis zwischen deutschem und Unionsrecht’, (2011) 64 DÖV 18, p 706.

86 Case C- 55/ 06, Arcor AG & Co. KG, [2008] ECR I- 2931, para 39.


under EU telecoms legislation.87 At the same time it confirmed its traditional ap-proach on national procedural law, noting:

[I] t is a matter solely for the Member States, within the context of their procedural aut-onomy, to determine, in accordance with the principles of equivalence and effectiveness of judicial protection, the competent court, the nature of the dispute and, consequently, the detailed rules of judicial review with respect to decisions of the NRAs.88

However, in its ruling in the main proceedings, the Cologne administrative court used the Court’s answer to specifically emphasize the NRAs’ ‘only limit-edly scrutinisable margin of discretion’.89 The Federal Administrative Court, in the case on appeal, noted that the Bundesnetzagentur had a Beurteilungsspielraum which closely related to the regulatory discretion. As a result, it was (only) up to the judge to determine whether the contested decision ‘was reasonably and completely argued in view of the criteria explicitly or implicitly laid down in the legislative norm’.90 The latter two requirements could indeed resemble a Chevron test (discussed previously).91 The resulting complexity and the need to ensure the compatibility of Germany’s administrative law with that of the EU and the other Member States led Ludwigs to propose to come to a single regime of executive discretion, merging the existing types.92

However, even if such an evolution towards a single regime were to materialize, this would not mean German courts would defer just as readily as US courts to agency decisions. Wimmer notes that German legal tradition is rather suspi-cious of the possibility for the administration to adopt acts which are not fully reviewable by the courts.93 This stems from Article 19(4) of the Basic Law, which provides: ‘Should any person’s rights be violated by public authority, he may have recourse to the courts. If no other jurisdiction has been established, recourse shall be to the ordinary courts’. From its wording, this provision in itself does not rule out the possibility for courts to be highly deferential to decisions of the executive, but the jurisprudence of the Constitutional Court has added much substance to the provision, for instance in Behördliches Beschwerderecht.94

Under that jurisprudence Article 19(4) GG is understood as granting full jur-isdiction to the courts both on matters of law and fact, marginal review being the

87 Ibid, para 159. 88 Ibid, para 170.89 VG Köln, 1 K 3427/ 01, 27 August 2009, para 121. Translated from German by the author.

See also Ludwigs, n 74, p 66.90 BVerwG 6 C 13.10, 23 November 2011, para 38. Translated from German by the author.91 See Thorsten Attendorn, ‘Das “Regulierungsermessen” Ein deutscher “Sonderweg” bei

der gerichtlichen Kontrolle tk- rechtlicher Regulierungsentscheidungen?’, (2009) MultiMedia und Recht 4, pp 238– 41.

92 Markus Ludwigs, ‘Das Regulierungsermessen als Herausforderung für die letztentschei-dungsdogmatik im Verwaltungsrecht’, (2009) 64 JZ 6, pp 292– 4. See also Matthias Jestaedt, ‘Maßstäbe des Verwaltungshandelns’, in Erichsen (ed), Allgemeines Verwaltungsrecht, Berlin, de Gruyter, 2010, p 351.

93 Norbert Wimmer, ‘Kontrolldichte— Beobachtungen zum richterlichen Umgang mit Entscheidungsprärogativen’, (2010) 65 JZ 9, p 433.

94 See Bundesverfassungsgericht, Behördliches Beschwerderecht, [1973] 35 BVerfGE 263, 274.

Agencification in the US and Germany 315

exception. While the tension between EU law and this traditional jurisprudence is not exclusive to the question of national agencies’ discretion under EU law,95 it is clear that the Regulierungsermessen granted to the Bundesnetzagentur also sits uncomfortably with that jurisprudence, since it allows the agency to determine the regulatory objectives to pursue.96 As regards the latter, Wimmer questioned whether the legislator should not lay down a list of objectives and prioritize among them, insisting that the legislator should deal with the strategic questions himself and should continue to insist on a judicial scrutiny of how the legislation is being implemented by the executive.97

This should be juxtaposed with the ruling of the Court of Justice in Commission v. Germany, where the Court (and the Commission) excluded such a power for the national legislator and instead stressed that the German NRA should have been accorded greater powers.98 In a very critical comment on this case, Gärditz noted that ‘this European ideal could hardly be in more contrast with the passed down pattern of argumentation in German constitutional law’, insisting on the Demokratie- and Rechtsstaatsprinzip.99

3 Agencification in the US and Germany

These two brief excursions into the US and German federal systems have high-lighted a number of interesting issues. Firstly, in the federal system in which fed-eral administration is itself uncontested, the practice of establishing and empow-ering independent agencies has foremost sparked discussion and controversy in the light of the separation of powers. Secondly, in the federal system in which federal administration is an exception to the rule, federal agencies have predom-inantly been scrutinized under a federalist perspective, the agencies being iden-tified as possible threats to the autonomy of the federal entities. This is not to say that the notion of control is absent in the second type of federal system. As was observed, the mechanism to control such agencies was actually firmly established and has only recently been contested under pressures from European integration. Thirdly, the notion of control, which is of fundamental importance in both types of federal polity, is always linked to the notion of democracy: either because the Congress should lay down a sufficiently clear standard for the agency to enforce; because the President, in order to be accountable to the people, should have suf-ficient control over an agency; or because the Legitimationskette ultimately goes back to the people. Here it should be noted that in the EU, the democratic ques-tion still is in flux, which will have repercussions for the question of control over

95 See the observations of the German Bundesfinanzhof presented in Opinion of AG Jacobs in Case C- 269/ 90, Technische Universität München, [1991] ECR I- 5469, para 11.

96 Wimmer, n 93, p 435. 97 Ibid, p 438. See also Attendorn, n 91, p 241.98 See Case C- 424/ 07, n 77.99 Klaus Ferdinand Gärditz, ‘Anmerkung zur Rechtssache C- 424/ 07’, (2010) 65 JZ 4, p 200.

Translated from German by the author.


agencies. Lastly, the US case in particular shows first that the notion of control is a rather relative one and further that the debate on agencification is a never- ending story. It would be vain hope therefore to think that the debate in the EU could easily be settled.

4 Controlling EU Agencies

As was noted in the previous section, the question of the EU’s democratic underpinning has not been settled yet and this affects the question of control over EU agencies, since one major question is through which mechanism the agencies’ legitimacy is linked back to the people. As set out previously (see section II 2), neither the EU’s democratic deficit thesis nor the necessity of a European demos is subscribed to here. While the EU is found to be democratic, there evidently still is room for improvement and the EU’s democratic matur-ation is ongoing.

4.1 Controlling agencification

The assessment of whether agencification is controlled in the EU gives mixed results. Just like in the US, but unlike in Germany, there is no clear legal basis for EU agencification. The attempts to establish a clear framework in secondary legislation have only resulted in the Common Approach, which does not establish a sound framework (cf. section II 3.6). Of course this in itself is not different from the situation in the US, but the latter polity is far more mature than the EU.

A genuine framework could contribute both to controlling the process of agen-cification and to its transparency. While currently each of the institutions may have its own reasons for supporting agencification, there is no explicit raison d’ être for the phenomenon. A framework informed by a consensus on this issue would allow the EU citizen to engage more usefully in the political debate, enhancing procedural legitimacy.

However, two other issues which cause some friction in the US and Germany are absent from EU agencification, resulting in a process which is much more in control than agencification in the first two polities. While in the US the President is described as the head of the executive by the Constitution, it is Congress which establishes agencies, defines their powers, and provides for the level of presiden-tial control. A similar issue could be noted in the case of Germany, where the main line of contention is not so much that between the legislative and execu-tive (also because Germany has a parliamentary system) but that between Bund and Länder, whereby the Bundestag may establish and empower agencies without support from the Bundesrat. As a result, in the US and Germany, the two actors which stand to lose the most from federal agencification do not have an unassail-able veto power in the decision- making.

At EU level this is clearly different. After all, the Commission must first table a proposal and may in exceptional circumstances withdraw proposals on

Controlling EU Agencies 317

agencification.100 The Member States have individual veto powers under Article 352 TFEU and a collective veto power under the ordinary legislative procedure. As a result, the Commission is much more in control than the US President, and the Member States much more than the Länder. Even if EU agencification may be questionable from a legal perspective, it is founded on a broad political consensus and has largely remained politically uncontested (see section III 4.5).

4.2 Controlling individual agencies

Seen from a US perspective, EU agencies do not seem to pose too many problems. Because EU agencies have rather limited and well- defined mandates, the content, purpose, and scope of the authority ‘delegated’ to them is hardly problematic. This is so even for the agencies with the most far- reaching powers, such as the European Securities and Markets Authority (ESMA). Further, the representatives on the Board may be dismissed and replaced at will by the entity which they re-present,101 and the Director of an agency may be dismissed at will by the Board members (cf. section II 3.2.4).

The problem which EU agencies still pose results from their hybrid nature. If EU law is implemented by the Member States’ authorities, there is no direct control by the EU institutions. Any control is exercised by the Commission vis- à- vis the Member State.102 If however EU law is implemented by the Commission, other EU institutions and bodies may fully exercise control within the limits of their competences. Because agencies are established as independent bodies and because the Council has insisted on the link between the Member States and the agencies, the question of who should control the agencies has not been completely resolved. Obviously EU agencies are EU bodies and hence are subject to the control of the Ombudsman, the Court of Auditors, the EU Data Protection Supervisor, etc. However, whether they were subject to the control of the EU’s budgetary authority was only resolved during the second wave of agencification, and because agencifi-cation has developed ad hoc, so have the instruments for control.

4.2.1 Control from a federal perspectiveWhile national parliaments can scrutinize EU developments of their own accord,103 following the Lisbon Treaty, they have also become formal actors in the EU legislative

100 See Case C- 409/ 13, Council v. Commission, ECLI:EU:C:2015:217.101 The exception here is the EFSA. The members on the Board are appointed by the Council

and do not act as representatives, except for the representative of the Commission. Article 15 of the EFSA Board’s Rules of Procedure therefore provides that the Board may ask, by a two- thirds maj-ority, for the replacement of one of its members.

102 The reports adopted by the Court of Auditors on the implementation of EU law by the Member States may be seen as an exception, but these are more instruments for accountability than instruments of control. See for instance European Court of Auditors, Special Report 1/ 2013.

103 For studies undertaken by national parliaments on the phenomenon of EU agencies, see Johannes Saurer, ‘National Governance and Networked Accountability Structures: Member


process under Protocol 1 to the Treaties, and a subsidiarity scrutiny is foreseen under Protocol 2 to the Treaties. Since every agency is established through a legislative act and an agency’s mandate may only be fundamentally reviewed through a legislative act, this mechanism allows national parliaments to influence agencification, subject to the general and significant limitations of this procedure.104

Compared to the national parliaments, the Member States’ governments are obviously in much greater control, since a qualified majority of them is needed before an agency is established or before an agency’s powers are reviewed. The Council and the Member States also have a much greater role in possible non- legislative procedures, that is, the adoption of delegated and implementing acts. In this way, Member States retain sufficient control over the possible regulatory output of EU agencies.

In addition to this external control,105 the Member States may also exercise control from ‘within’ the agency, even if the size of the Boards make it difficult to coordinate action. Sometimes (see section III 3.1) the national representatives in EU agencies may themselves be regulators independent from the national govern-ment, in which case it would be difficult to claim that the (governments of the) Member States exercise control on the Board. These difficulties would make the internal control exercised by the Member States appear weaker than one would assume based on the Member States’ insistence on their representation on the Boards (cf. section III 4.1).

The difficulty of coordinating national representatives on agencies’ Boards also limits the Member States’ power to appoint and dismiss the Director, the Council and its bodies typically not being involved in the procedure.106 Furthermore, in-volving the Council in this procedure, either formally or informally, would move the hybrid agency instrument closer to direct administration, something which the Member States have previously rejected.

4.2.2 Control from an institutional balance perspectiveEven if the institutional balance may be no more than the notion contained in Article 13 TEU, it remains possible to derive from it a level of control nec-essary for the institutions and agencies to remain within the limits set in the Treaties.

State Oversight of EU Agencies’, in Rose- Ackerman and Lindseth (eds), Comparative Administrative Law, Cheltenham, Edward Elgar, 2010, pp 623– 5.

104 Its actual application also remains unsatisfactory for the national parliaments because the Commission applies the procedure restrictively. See Merijn Chamon, ‘Eerste toepassingen van de subsidiariteitstoets’, (2014) 62 SEW 6, p 278– 80.

105 Furthermore, as regards access to information, the discussion on the work programme and the annual report of the agencies (cf. section II 3.3) has shown that not all establishing acts of EU agencies prescribe that these documents are sent to the Member States for information.

106 The exceptions here are the CFSP agencies as well as Europol. In addition, for the FRA the Council may communicate its preferences to the Board on the Commission’s shortlist of candidates. See section II 3.2.2.


4.2.2.1 The CouncilThe Council, unlike the Member States, is kept informed of the functioning of the agencies through the work programmes and annual reports. In addition, the Council has to approve the agencies’ resources under the EU budget, allowing for ex ante and ex post control.

Whether the Council uses its instruments to effectively control agencies is a different matter. The chronically substantial carry- overs by the agencies would suggest the contrary. In addition, where Parliament has sometimes decided to withhold or postpone its discharge,107 the Council has always recommended the discharge be granted.

The Council’s seeming lack of concern about controlling the agencies may be explained by its strong position driving the process of agencification (cf. section III 4.1). Further, even if the self- financed agencies do not depend on the Council for their budget, these agencies do financially depend on the fees they charge and which are set pursuant to a comitology procedure involving the Member States. From this it follows that the Council’s interest in more thorough ongoing con-trol could be kindled when agencies would gain budgetary independence and/ or when the agencies’ mandates would depend on general clauses such as is the case in the US.

On this it may be noted that Short- selling effectively sanctioned the establish-ment of regulatory agencies at EU level, although the Court of Justice of the European Union (CJEU) has not (yet) adopted the US Supreme Court’s generous intelligible standard jurisprudence.

4.2.2.2 The CommissionThe position of the Commission in EU agencification is of course different (cf. section III 4.2). Whereas the Council or the Member States have not so much insisted on controlling agencies, the Commission has sought to ‘regain’ control of these bodies, albeit without much success. Following the Common Approach, the Commission has definitively given up on the idea of having parity in the agencies’ Boards and on bringing EU agencies under the umbrella of direct administration. The Commission itself has aptly described the difficult relationship between itself and the agencies as the latter being its ‘satellites’.108

107 For CEPOL the discharge has once been refused: Decision (EU) 2010/ 556 of the European Parliament, OJ 2010 L 252/ 232; Decision (EU) 2010/ 756 of the European Parliament, OJ 2010 L 320/ 11. Eventually, the accounts were closed by the Parliament: see Decision (EU) 2011/ 150 of the European Parliament, OJ 2011 L 62/ 31. For the EEA, EMA, and EFSA the discharge was postponed but ultimately granted. See Decision (EU) 2012/ 618 and Resolution of the European Parliament, OJ 2012 L 286/ 356; Decision (EU) 2012/ 620 and Resolution of the European Parliament, OJ 2012 L 286/ 367; Decision (EU) 2012/ 622 and Resolution of the European Parliament, OJ 2012 L 286/ 377. See Decision (EU) 2012/ 800 and Resolution of the European Parliament, OJ 2012 L 350/ 76; Decision (EU) 2012/ 802 and Resolution of the European Parliament, OJ 2012 L 350/ 82; Decision (EU) 2012/ 804 and Resolution of the European Parliament, OJ 2012 L 350/ 89.

108 European Commission, EU Agencies: Whatever You Do, We Work for You, Luxembourg, Office for Official Publications of the European Communities, 2007, p 1.


The different instruments available to the Commission to control EU agencies are manifold. Firstly, the Commission has the power of legislative proposal,109 and it also plays a role in the legislative procedure.110

A control instrument which allows the Commission a more targeted approach is its budgetary powers, since it will be the Commission proposing the agencies’ draft budgets to the Parliament and Council. The Commission therefore has some ex ante control as well as the possibility to ‘punish’ an agency in the ex post phase. While the Commission exercises control over the implementation of the agencies’ budgets under Regulation 1271/ 2013, this does not allow it to control the policy output of an agency. As regards administrative measures more generally, the Commission has a veto power over the agencies’ implementing rules in staff and financial matters.111

The Commission cannot formally control an agency through its involvement in the internal bodies of EU agencies. This results from the Director’s appointment procedure and from its minority position on the Board. In addition, with each new enlargement of the EU, the Commission’s weight on the Board decreases. For instance, the EMSA was originally established in 2002 with four Commission representatives holding 21 per cent of the votes, giving the Commission a veto power on issues to be decided by a four- fifths majority (such as the appointment of the Director).112 Following the 2004 enlargement the Commission’s influence on the Board was diluted and it can no longer block the decision appointing the Director.113 Lastly, while the early warning system provided in the Common Approach does not add anything to the powers of the Commission, it would seem to confirm that it is the Commission’s responsibility to ensure the agencies’ com-pliance with their mandate and general EU policy.

This raises the question of whether the Commission should not simply be em-powered to strike down a decision of the Board. A number of establishing regula-tions indeed provide for this possibility (cf. section V 6.1.1.3.1), but the legislator has not resorted to this mechanism consistently. In addition, its purpose is not so much to allow for Commission control but rather to allow wronged third parties to have an agency decision reviewed in absence of clear CJEU jurisdiction. The Lisbon Treaty having established this jurisdiction, the question of whether such a power should be vested in the Commission depends on how one conceives the executive power in the EU.

Evidently the notion of a unitary executive power is absent in the EU legal order.114 But even if the EU did not evolve to such a model,115 one could still argue that the

109 Of course, the CFSP agencies form an exception here, as well as the JHA agencies on which one quarter of the Member States may table a proposal pursuant to Article 76 TFEU.

110 See n 100. 111 See Chapter II, n 80. 112 See Chapter II, n 138.113 See European Commission, COM (2010) 611 final, p 9.114 See Deirdre Curtin, Executive Power of the European Union: Law, Practices, and the Living

Constitution, Oxford, OUP, 2009, p 104. As Curtin notes, the Commission itself also promotes the idea of a unitary executive. See for instance its intervention in the IGC of 2003: European Commission, COM (2003) 548 final, p 4.

115 In fact, Adamski reads the Short- selling ruling as an admission that in an EU of twenty- eight ‘a unitary executive is simply non- viable, mainly for political reasons’. See Dariusz Adamski, ‘The ESMA Doctrine: A Constitutional Revolution and the Economics of Delegation’, (2014) 39 ELRev 6, p 828.


Commission is the primary core executive.116 If such a position of the Commission were accepted, one could allow the Commission to take on further responsibility for (and control of ) the subsidiary actors. A more uniform system of administrative scrutiny by the Commission vis- à- vis the EU agencies should then be put in place. A general prerogative for the Commission to annul or suspend agency decisions could be part of this, as well as a greater involvement in the agency’s decision- making.

Evidently the question of whether such control mechanisms should be general-ized, vesting an important oversight power in the Commission, will also depend on how the EU agencies’ independence is conceived in the EU legal order. If the American notion of independence is adopted, much as EU legislation itself increasingly imposes on the Member States, such a far- reaching control by the Commission would be out of place. At the same time such independence would spark the same discussion on the legitimacy of these bodies, as one may now wit-ness in the Member States. It would then seem appropriate to strengthen the EU Parliament’s powers vis- à- vis the agency, as was also proposed by Ludwigs in rela-tion to the national agencies in Germany (cf. section V 2.3).

4.2.2.3 The ParliamentFrom an institutional balance perspective, it is clear that EU agencies are at least problematic, in the sense that the Parliament’s traditional powers of control over the EU’s executive are being undermined. If it is accepted that agencification is an alternative to granting administrative powers to the Commission, the conclusion must be that the Parliament loses power and influence.117

An analysis of the Parliament’s well- known scrutiny powers over the Commission has been undertaken elsewhere and is not required here.118 Suffice it to note that the Parliament’s approval is needed for the appointment of the Commission President and that of the Commission as a college. With this also comes the power to dismiss the Commission as a body. The latter power further interacts with the Parliament’s power of the budgetary discharge, a motion of censure against the Commission being the logical next step when discharge is refused.119 As regards the Parliament’s powers of scrutiny sensu stricto, it may ask the Commission, Council, and ECB questions and may organize debates with the Commission and Council. To ensure a smoothly functioning cooperation between the Parliament and the Commission, the two institutions have also con-cluded an interinstitutional agreement.120

116 Curtin, n 114, p 91.117 In cases where agencies are not an alternative to the Commission, this premise may of course

be questioned. See also section V 4.2.3.118 See David Judge and David Earnshaw, The European Parliament, New York, Pallgrave

Macmillan, 2008, pp 188– 228. On Parliament’s control in general, see Gregorio Garzón Clariana, ‘L’évolution du contrôle parlementaire de l’exécutif dans le droit de l’union européenne’, in Govaere and Hanf (eds), Scrutinizing Internal and External Dimensions of European Law— Les dimensions in-ternes et externes du droit européen à l’épreuve, Bruxelles, Peter Lang, 2013, Vol 1, pp 199– 210.

119 Judge and Earnshaw, n 118, p 202.120 See annex IV to the Framework Agreement, OJ 2010 L 304/ 47.


As was noted in the section on political limits, the Parliament tried to secure ex ante control over EU agencies by insisting on the codecision procedure. As regards ongoing control, the Parliament has foremost tried to impose on the Commission the responsibility for agencies. Since the Commission itself failed to secure greater control over EU agencies, it did not want to take on such respon-sibility, noting that agencies are (merely) its satellites (cf. section V 4.2.2.2). The result then is unsatisfactory as regards the possibility for the Parliament to exercise control over the agencies.

For instance, the Parliament only has a marginal role in the appointment of the main bodies of the agencies (cf. sections II 3.1.2 and 3.2.2). Similar problems may be noted with Parliament’s scrutiny powers sensu stricto. As was noted previ-ously (see section II 3.2.2), the establishing acts of some agencies do provide that the Director may be invited by the relevant Parliamentary Committee to attend a hearing, but the Common Approach has not codified this in a general prerogative of the Parliament. Of course, relations between the Parliament and the agencies may develop rather smoothly in absence of such clear procedures, but on a formal account the situation remains unsatisfactory.

In addition it should be noted that even if the Director represents his agency externally it would not make total sense to hear him and hold him accountable for the general policy of the agency, while only the regulations establishing the ESAs and the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) provide that the Chair of the Board may be invited by the Parliament to a hear-ing. Further, the detailed arrangements on the Parliament’s involvement with the Commission’s work programme are not in place when it comes to the agencies’ work programmes.

When it comes to the instrument of Parliamentary questions, the problems are clear from practice. MEPs do not have an opportunity to ask questions directly to EU agencies. As a result, they have to address them to the Commission, which first forwards them to the relevant agency and then communicates the answer back to the MEP. The end result of this procedure lacks transparency, since only the Commission’s but not the agency’s answer is made public.121 Furthermore, when an MEP is not satisfied with the agency’s answer and asks a follow- up ques-tion to the Commission, the latter will deny any responsibility for the quality of the agency’s answer, citing the agency’s independence.122 This is a recurring feature in general when it comes to the Commission being questioned about the policies followed by EU agencies.123 While the Commission is not wrong to deny responsibility, the overall result is untransparent and unsatisfactory.

121 See for instance the Answer of the European Commission to Parliamentary Question (E- 008719/ 2011) by MEP Georgios Papanikolaou, 28 October 2011; Answer of the European Commission to Parliamentary Question (E- 000612/ 2012) by MEP Arthur Zasada, 22 February 2012.

122 See for instance the Answer of the European Commission to Parliamentary Question (E- 0276/ 2009) by MEP Frank Vanhecke, 17 March 2009.

123 See for instance the Answer of the European Commission to Parliamentary Question (P- 004581/ 2012) by MEP Gilles Pargneaux, 1 June 2012; Answer of the European Commission


4.2.3 Strengthening control over EU agenciesThe tenor of the sections on the control exercised by the institutions raises the question of what new mechanisms could be introduced to secure a greater level of control over EU agencies. The answer to this question is not clear- cut, but in any scenario the Parliament’s powers should formally be strengthened. Depending on how the Commission’s role is conceived as part of the EU executive, its control powers could also be strengthened. Ultimately, the question of greater control should be answered by a reflection on the constitutional foundations of the EU. If, despite recent evolutions, a decision is made to strengthen the old community method, the Commission should see its position in the EU executive strength-ened. Another option is to further parliamentarize the EU, possibly establishing a primacy of the Parliament.

The plea to strengthen the powers of the Parliament (ultimately mirroring its powers vis- à- vis the Commission) could be criticized on the ground that the agen-cies and the Commission are fundamentally different actors. The agencies, being technical bodies, should not be under the same parliamentary scrutiny as the Commission. While this may be true, three remarks are in order. Firstly, ensuring the Parliament has a complete arsenal of control instruments does not mean Parliament will continuously use these instruments, but it does allow Parliament to intervene when necessary. Secondly, whether agencies are really ‘merely’ en-gaged in technical matters may simply be part of a mantra that might come under pressure in the future. Thirdly, if this critique should indeed be accepted, it would appear absolutely necessary to strengthen the Commission’s control over the agen-cies (and further strengthen Parliament’s power over the Commission) to ensure a strong Legitimationskette.

4.2.3.1 Strengthening Parliament’s powersFirstly, the extra requirements which should be imposed on the agencies in their relations with the Parliament should not be laid down in bilateral agreements between the agencies and the Parliament. Since agencies are executive bodies of a second order, these requirements should be imposed on the agencies in their establishing acts (or ideally in a horizontal act).124

4.2.3.1.1 Information requirementsMinimally, a formal obligation should be provided for the Directors of EU agen-cies to make a statement before the relevant Parliamentary Committee whenever

to Parliamentary Question (E- 006447/ 2012) by MEP Maria Bizzotto, 1 August 2012; Answer of the European Commission to Parliamentary Question (E- 006563/ 2012) by MEP Sergio Paulo Fransesco Silvestris, 17 August 2012. Similarly for questions to the Council, see Answer of the Council to Parliamentary Question (E- 3282/ 06) by MEP Angelika Beer, 4 October 2006; Answer of the Council to Parliamentary Question (E- 005663/ 2011) by MEP Angelika Werthmann, 26 July 2011.

124 See however, Article 45(7) SRB, which foresees an agreement between the SRB and Parliament.


the latter so requests.125 In more exceptional circumstances the Chair of the Board could also be called for a hearing by the committee. These instruments, together with complete and automatic access to the information held by the agency,126 could then be a substitute for the Parliament’s representation on some of the agencies’ Boards, which now confuses the Parliament’s role vis- à- vis the Boards.

As to transparency for the citizen, a clear obligation on the part of the agen-cies to answer Parliamentary questions should also be provided for. Again here it would be important to lay down these principles unilaterally. In this sense it would be different from the arrangement which the Parliament and the ECB have informally worked out, whereby Parliament amended its Rules of Procedure (cf. current rule 118) and the ECB ‘agreed’ to answer such questions in absence of any formal obligation.127 As regards the agencies, the Parliament obviously should not have to rely on their affability to allow Parliamentary scrutiny.

4.2.3.1.2 Extending Parliament’s reach to agencies’ Directors?So far, the changes proposed are rather uncontroversial. This would be different as regards an upgraded role for the Parliament in procedures for the appointment and dismissal of the Director. The Director would thus have to be confirmed by the Parliament before he can take office. As regards the dismissal, the Parliament could have a formal right to make requests to the appointing authority. If the ap-pointing authorities remain the Boards, the Parliament would not be able to force the Boards to act on such a request. However, a refusal by a Board could be one of the scenarios in which the Parliament could summon the Chair of the Board. If, on the other hand, the Commission became the appointing authority of the Directors (as it now is for the first generation agencies), the Parliament could, in theory, force the Commission to dismiss a Director.

4.2.3.1.3 Extending Parliament’s reach to agencies’ Boards?The question of whether the Parliament should have a greater role in nomin-ating and dismissing the Board is even more controversial. Still, a greater role for the Parliament than is currently foreseen could also be explored, since an im-portant part of the EU’s democratic quality should rest on an enhanced role of the European Parliament (cf. section II 2).

Since the Parliament has no say over the EU’s indirect administration, enhan-cing its role vis- à- vis the Boards of EU agencies would signify a shift towards rec-ognizing agencification as an instance of direct administration, something which

125 Currently the Parliament’s Rules of Procedure (rule 126) provide that it may submit requests to agencies, but only ‘in cases where the Parliament has the right’ to do so (emphasis added).

126 As regards confidential information, further arrangements could be worked out similarly to the Parliament’s scrutiny of the ECB in the Single Supervisory Mechanism. See Annex II to the Framework Agreement on relations between the European Parliament and the European Commission, OJ 2010 L 304/ 47 and Interinstitutional Agreement between the European Parliament and the European Central Bank, OJ 2013 L 320/ 1.

127 See European Central Bank, Monthly Bulletin, November 2002, p 54.


the Member States have opposed. Still, if an enhanced role of the Parliament is pursued this could be done indirectly, by enhancing the Commission’s position vis- à- vis the Boards. It could also be done directly by allowing the Parliament to veto a Board member. This would give the Parliament a greater say than its mere advisory role in the appointment of the EFSA Board, but it would not go as far as requiring a Parliamentary approval of the Board.

4.2.3.1.4 Completing Parliament’s budgetary powersAs regards the budgetary procedure, it was noted that as of yet the fully self- financed agencies are not subject to Parliamentary discharge. Two solutions have been proposed to ensure greater democratic accountability as regards the dis-charge to these agencies.128 Thus, discharge would still be given by the compe-tent agency bodies, but the Parliament’s involvement in this procedure would be enhanced. Alternatively, the discharge would simply be granted (or refused) by the Parliament itself. The second solution should be preferred and the Parliament could even take the bold step of unilaterally claiming this power by amending its Rules of Procedure.

Aside from the interinstitutional politics, the problem of the Parliament’s control ultimately goes to the democratic foundations of the EU. If it were accepted that the EU should further move towards a parliamentary system and/ or that the EU’s democratic legitimacy is primarily derived from the Parliament, it cannot be denied that the democratic legitimacy of all execu-tive actors (with the exception of the (European) Council) depends on the Parliament. Rejecting such a role for the Parliament would have the EU’s democratic legitimacy depend on the Member States. However, EU struc-tures, even EU agencies, are quite well detached from the national level, and it appears unconvincing that the Member States could ensure credible demo-cratic control over EU agencies.

4.2.3.2 Strengthening the Commission’s control over agenciesUnless Parliament’s control is directly vested, it would appear necessary to also strengthen the Commission’s hold over EU agencies. Although this would also sit well with the evolution towards a parliamentary model at EU level, it would appear much more controversial than the strengthening of the Parliament.

The Commission could be strengthened by generalizing certain arrange-ments which may today already be found in some establishing regulations. It would be useful for instance to have the Board presided over by a representa-tive of the Commission. As regards the work programme, a balance between Commission control and agency autonomy could be found in the procedures applicable in the EMCDDA, European Railway Agency (ERA), and European Maritime Safety Agency (EMSA). As regards the Director, the Commission

128 European Parliament, ‘The Income of Self- financed Agencies and the EU Budget’, DG for Internal Policies, 2013, pp 66– 70.


could be the appointing (and thus dismissing) authority, as is now the case for the Cedefop and Eurofound. Combining these measures and generalizing them for all agencies would result in a significantly greater control of the Commission while maintaining agency autonomy. At the same time, the existing composi-tions of the Boards need not be altered, allowing all Member States to be dir-ectly involved in steering the agency. As noted, one further way of strengthening the Commission’s position in EU agencies’ Boards is to provide that Executive Boards should be established, reserving a position for the Commission in those bodies (cf. section II 3.1.8).

4.2.3.3 The impact of strengthened control on EU agenciesProviding the institutions with more formal control instruments vis- à- vis the agen-cies would not necessarily impact much on the autonomy of the agencies, since the intensity of the ongoing control over EU agencies would not necessarily change.

On the other hand, strengthening and formalizing the institutions’ control could result in further ‘burdening’ the agencies. When devising control mecha-nisms, account should therefore be taken of the extra costs which these mecha-nisms impose on agencies,129 and at the same time it again puts into question the proportionality (cf. section IV 3.3) of creating small agencies (with limited mandates).

5 Concluding Remarks on the Political Control of Agencification

The overall process of EU agencification has been found to be more controlled than the processes of agencification in the US and Germany and as a result, the process is less contested. However, this does not detract from the fact that the failure to come to a binding horizontal framework in secondary law remains to be deplored. Ideally, even if an enabling clause is to be written in primary law, a binding and sufficiently detailed framework should be established in secondary organic law.

As regards the individual agencies, control has not been found to be too prob-lematic, in large part because of their limited mandates. However, the hybrid nature of these bodies was found to have some problematic repercussions as to the precise role to be played by each institution in controlling agencies. In addition it was noted that this question is further complicated because of the interaction with the discussion on the nature of EU democracy, a debate which has not been settled yet. Lastly of course, the issue of control will become more important fol-lowing Short- selling: while the need to control agencies which simply gather infor-mation may seem minimal, this is different for agencies which have been granted significant powers.

129 Similarly for the US independent agencies: see Custos, n 4, pp 632– 3.

Judicial Scrutiny of EU Agencies’ Acts 327

Because the agencies have been established as a result of the Member States’ (in Council) insistence on their involvement in the EU administration (instead of relying on direct administration through the Commission), one would expect that this also implies that they exercise sufficient control. However, the Member States themselves are in no position to control and take on responsibility for the agencies. Since EU agencies are EU bodies, control should foremost lie with the EU institutions. From this perspective the Parliament and the Commission were identified as the primary institutions to exercise control: the first being the as-sembly representing the citizens and exercising democratic control over the execu-tive in general; the second as the core executive with a special responsibility for the subsidiary executive bodies.

6 Judicial Scrutiny of EU Agencies’ Acts

Because it differs from political control, the analysis of the judicial control over EU agencies has been incorporated in the present section dealing with the pos-ition of the EU agencies in EU procedural law.130

While the following sections are not confined to judicial protection against agencies’ acts, the latter constitutes a major part of what follows. Indeed, the legitimacy of administrative action also depends on the possibility to have that action reviewed before an independent judiciary, a central concern in Meroni and Romano. Since the EU is based on the rule of law,131 and the Court of Justice has the task of ensuring the law is observed in the application and interpretation of the Treaties,132 its responsibility to ensure a complete system of legal remedies and procedures133 also extends to the acts of the agencies.

To secure judicial review of agencies’ acts pre- Lisbon, different techniques have been worked out. They will be analysed in turn to further verify whether they should or could remain to be relied upon post- Lisbon. Special attention is thereby devoted to the peculiar system of the internal Boards of Appeal characterizing the decision- making agencies. As in a previous section (cf. section II 3), an un-warranted level of heterogeneity in relation to these bodies and their functioning may be noted. Evidently, the soundness of the solutions introduced by the Lisbon Treaty will also be gauged.

The Treaty provisions on the action for damages, which were not updated in light of agencification, and the separate legal personality of EU agencies raise the question of whether the EU itself could be held liable for agencies’ (in)action.

130 A short summary of some of these findings has previously been published. See Merijn Chamon, ‘EU Risk Regulators and EU Procedural Law’, (2014) 5 European Journal of Risk Regulation 3, pp 324– 37.

131 See the Preamble to the TEU and Article 2 TEU.132 Case 70/ 88, Parliament v. Council, [1990] ECR I- 2041, para 23.133 Case C- 50/ 00 P, Unión de Pequeños Agricultores v. Council, [2002] ECR I- 6677, para 40.


A major blind spot in the Lisbon Treaty relates to the possible active locus standi of EU agencies in proceedings before the Court, an issue which will also be discussed and which interrelates with the finality of the agencification process.

6.1 Passive locus standi of EU agencies

Pre- Lisbon, the Treaty provisions on legal remedies were completely silent as re-gards the agencies, Articles 230 and 232 EC only allowing the (in)action of the Parliament, Council, Commission, and European Central Bank (ECB) to be scrutinized by the Court. In what follows only the action for annulment will be commented upon.

6.1.1 Action for annulmentA strict reading of Article 230 EC did not allow any applicant to challenge acts of the agencies.134 The remedies to this lacuna were: (i) extending Les Verts to agen-cies; (ii) imputing agency decisions to a delegating authority; (iii) providing for legal remedies in secondary legislation; and (iv) amending the Treaties.

6.1.1.1 Filling the gap by broadening the scope of Les VertsPre- Lisbon, the agencies’ position was similar to that of the Parliament in the ori-ginal Rome Treaties when Article 173 EEC still only allowed scrutiny of Council and Commission acts. This lacuna led to the famous Les Verts case,135 in which the Court found:

The European Economic Community is a community based on the rule of law, inasmuch as neither its Member States nor its institutions can avoid a review of the question whether the measures adopted by them are in conformity with the basic constitutional charter, the Treaty. […] An interpretation of Article 173 of the Treaty which excluded measures adopted by the European Parliament from those which could be contested would lead to a result contrary both to the spirit of the Treaty as expressed in Article 164 [current Article 19 TEU] and to its system.136

Lenaerts was the first to suggest a transposition of Les Verts to the EU agencies,137 but Remmert questioned whether Les Verts could be applied beyond the realms of the institutions proper.138 Lenaerts’ suggestion gained weight following Commission v. EIB, where the Court’s jurisdiction to review acts of the EIB’s Management Committee was questioned, since Article 237(b) EC only allowed actions against the EIB Board of Governors. The Court relied on Les Verts and remarked: ‘[A] lthough it is not a European Community institution, the EIB none the less is a Community

134 Koen Lenaerts, ‘Regulating the Regulatory Process: “Delegation of Powers” in the European Community’, (1993) 18 ELRev 1.

135 Case 294/ 83, Parti écologiste ‘Les Verts’ v. Parliament, [1986] ECR 1339.136 Ibid, para 23. 137 Lenaerts, n 134, pp 45– 6.138 Barbara Remmert, ‘Die Gründung von Einrichtungen der mittelbaren Gemeinschaftsverwaltung’,

(2003) 37 Europarecht 1, p 140.


body established and endowed with legal personality by the EC Treaty and it is on that account that the EIB is subject to judicial review by the Court’.139 While this strengthened Lenaerts’ argument, it did not undermine Remmert’s objection but rather focused the problem: EU agencies are not established and endowed with legal personality by primary law.

Apart from applying a Les Verts reasoning, the Court could also have relied on its Chernobyl case law. Where in Chernobyl the Court used the institutional bal-ance to enable the Parliament to safeguard its prerogatives, the Court could have invoked the institutional balance to safeguard its own prerogatives as in Pergan Hilfstoffe,140 since the ordinary legislator should not be able to undermine the Court’s prerogative to uphold the law by empowering agencies that are immune from judicial scrutiny.

However, in Sogelma the Court of First Instance (CFI) ultimately opted for extending Les Verts to include measures adopted by the European Agency for Reconstruction (EAR).141 In its answer to the EAR’s plea of inadmissibility, the Court referred to Les Verts, adding that the EAR not being an EU institution was not relevant.142 Of course, the imminent entry into force of the Lisbon Treaty probably also played an important part in the Court’s decision.143 Still, the fact that Les Verts was extended to the EU agencies was remarkable because only a few days following Sogelma, the CFI, albeit a different chamber, declined to extend Les Verts to the Ombudsman, since the latter was not an institution under Article 7 EC.144 This is odd, since this observation equally applied to the EAR in Sogelma and Article 263 TFEU as introduced by the Lisbon Treaty would also allow ac-tions against the Ombudsman.

The reliance on Les Verts in Sogelma was further remarkable because the Court could also have solved the case by imputing the EAR’s decision to the delegating authority, the Commission. In fact the Court in its judgment contemplated this possibility but found it inapplicable in casu. However, analysis of the legal context in which the EAR operated points to another conclusion. Article 1 of the regulation establishing the EAR provided that ‘The Commission may dele-gate to an Agency implementation of the Community assistance provided for in [the CARDS regulation]’. To this end the Commission concluded a frame-work agreement with the EAR,145 and the actual Commission decisions on the

139 Case C- 15/ 00, Commission v. EIB, [2003] ECR I- 7281, para 75.140 See Chapter IV, n 726.141 Case T- 411/ 06, Sogelma v. EAR, [2008] ECR II- 2771. For an earlier attempt (by an appli-

cant), see Case T- 148/ 97, Keeling v. OHIM, [1998] II- 2217, paras 22– 5.142 Case T- 411/ 06, Sogelma v. EAR, [2008] ECR II- 2771, para 37.143 Daniel Riedel, ‘Rechtsschutz gegen Akte Europäischer Agenturen’, (2009) 20 EuZW 16,

p 568. For an argument in this vein dating from before Sogelma, see Valérie Michel, ‘Entre orthodoxie bienvenue et manque d’audace regrettable: de la recevabilité du recours contre un avis’, (2008) Europe 255, pp 13– 14.

144 Case T- 196/ 08, Srinivasan v. Ombudsman, ECLI:EU:T:2008:470, para 13.145 Framework Agreement between the Commission and the EAR, Implementing Modalities for

the Management of Financial Transfers between the Commission and the Agency, on file with the author.


Annual Action Programmes explicitly provided that the implementation of the Programme was delegated (by the Commission) to the EAR.

Despite this clear legal framework, the Court in Sogelma ruled that the dec-ision could not be imputed to the Commission as in Schering- Plough,146 since the EAR’s powers were not mere advisory powers. This of course shows either a confused reading of Schering- Plough or the problems inherent in that judgment, since if in Sogelma the EAR had exercised advisory powers, the Court could have declined jurisdiction because of a lack of binding legal effects.147

While the Court’s reasoning in Sogelma on the impossibility to impute the EAR’s decisions to the Commission was unconvincing, the Court should not be faulted for dismissing that option in favour of broadening the scope of Les Verts. After all, the solution of imputing agency decisions to the Commission has a number of fundamental shortcomings which argue against it being used (dis-cussed later). Although Sogelma anticipated the entry into force of the Lisbon Treaty it remained relevant for some time afterwards,148 following the maxim tempus regit actum.149

Lastly, in Spain v. Eurojust,150 the Court was faced with a twofold problem: firstly, Spain had brought its action (against a third pillar body) under Article 230 EC instead of under Article 35 EU. Secondly, that latter Article only provided that (framework) decisions could be challenged. AG Poiares Maduro quickly dis-missed the possibility to bring the action under Article 230 EC,151 and suggested that Article 35 EU could be read similarly to Article 173 EEC in Les Verts.152

It is hard to find ‘jurisdictional alchemy’ in the AG’s solution, as one com-mentator asserted;153 or at least it was not more alchemistic than the Court’s solution in Les Verts. Either way, the Court did not follow the AG’s suggestion and confined itself to reviewing Spain’s plea based on Article 230 EC. Because acts of Eurojust were not mentioned in Article 230 EC, the Court declared the application inadmissible.154 Of course, the Treaty of Lisbon has now addressed this issue.

146 Case T- 133/ 03, Schering- Plough Ltd v. Commission and EMEA, ECLI:EU:T:2007:365.147 See eg Case T- 311/ 06, FMC Chemical and Arysta Lifesciences v. EFSA, [2008] ECR II- 88;

Case T- 312/ 06, FMC Chemical v. EFSA, [2008] ECR II- 89; Case T- 397/ 06, Dow AgroSciences v. EFSA [2008] ECR II- 90.

148 See Case T- 70/ 05, Evropaïki Dynamiki v. EMSA, [2010] ECR II- 313, para 65. Similarly, Case T- 69/ 05, Evropaïki Dynamiki v. EFSA, ECLI:EU:T:2007:314.

149 Case T- 539/ 08, Etimine SA and Ab Etiproducts Oy v. Commission, [2010] ECR II- 4017, para 76.

150 Case C- 160/ 03, Spain v. Eurojust, [2005] ECR I- 2077.151 Opinion of AG Poiares Maduro in Case C- 160/ 03, Spain v. Eurojust, [2005] ECR I- 2077,

para 14.152 Ibid, para 20.153 Richard Creech, (2006) 2 European Constitutional Law Review 1, p 148.154 Case C- 160/ 03, n 150, para 40. The Court did note that the contested measures could be

challenged under Article 91 of the Staff Regulations. These proceedings would be brought by nat-ural persons, allowing the intervention of Member States such as Spain. See ibid, paras 42– 3. This led to the bizarre result that traditionally privileged parties could not challenge a decision, whereas non- privileged parties could.


6.1.1.2 Filling the gap by imputing agencies’ decisions to the CommissionA second way of solving the lacuna in the system of judicial protection is to re-define the measure by attributing it to an institution whose acts are challengeable.

However, this legal fiction is only really tenable if the act in question has been adopted in the exercise of delegated powers. The technique was adopted for the first time by the Court in SNUPAT, where decisions of the same Brussels agen-cies as in Meroni were at issue. The Court first remarked that the Caisse had been set up by the High Authority and that it held part of its powers. Secondly, the Court noted that the decisions of the Caisse were final decisions and that the High Authority could have provided for the possibility of an administrative appeal, but that it had chosen not to do so.155 As a result, it concluded:

Therefore, it must be accepted— and to do otherwise would be to deprive the undertak-ings of the protection afforded them by Article 33 of the ECSC Treaty— that the decisions adopted by the [Caisse] rank as decisions of the High Authority and, as such, are open to applications for annulment under the conditions laid down in Article 33.156

More than thirty years later a similar case came before the CFI in DIR International Film. In addition to the discussion of this case in a previous section (see IV 5.5.1.3), it may be noted that the Commission decided on the guidelines which the European Film Distribution Office (EFDO) had to follow. Likewise, before the EFDO took the actual decisions on the subsidies, the Commission made its binding views known.157 As a result, the Commission in DIR International Film accepted that the contested decision was imputable to it and the Court corres-pondingly accepted jurisdiction.158

Both in SNUPAT and in DIR International Film, imputing decisions adopted by private parties to the Commission was an elegant and logically sound solution. In subsequent cases, however, the Court also applied this solution to acts of EU agencies, in a much less logically sound way.

Thomae v. Commission159 concerned a decision of the European Medicines Agency (EMA) and, because the agency generally only has an advisory capacity, the Commission taking the formal decision granting or denying authorizations, no provision had been made in its establishing regulation for judicial review.

However, in further (material) legislation, the EMA had received further powers. Thus when a marketing authorization has been granted, it can still be amended, for example when it has to be renewed,160 but also when the applicant requests major and minor variations to the authorization. This was the case in Thomae v. Commission, where the applicant had requested an authorization to

155 Joined Cases 32/ 58 & 33/ 58, SNUPAT v. High Authority, [1959] ECR 127, pp 137– 8.156 Ibid, p 138.157 Joined Cases T- 369/ 94 & T- 85/ 95, DIR International Film e.a. v. Commission, [1998] ECR

II- 357, para 52.158 The Court’s judgment was successfully appealed in Case C- 164/ 98 P, albeit unrelated to

this point.159 Case T- 123/ 00, Dr Karl Thomae GmbH v. Commission, [2002] ECR II- 5193.160 See Articles 14 and 39 of Regulation (EC) 726/ 2004, OJ 2004 L 136/ 1.


market the product under different trade names. The procedure was laid down in a Commission regulation, implementing the regulation establishing the EMA.161 Under the procedure, variations are submitted to the EMA, which either allows the variation (resulting in an amendment of the original authorization by the Commission) or disallows it, in which case the applicant can amend its request.

In Thomae v. Commission, the EMA had rejected the variation since it found that a product authorized for the whole internal market could not subsequently be marketed under different trade names. Evidently, this decision precluded the applicant from amending its application in line with the EMA’s observations. While the Court ruled favourably for the applicant, the more interesting aspect of the case for the present section is the question of the action’s admissibility. The procedure in the implementing regulation had been set up as such that it did not result in a challengeable Commission decision in the event of a negative appraisal. Such a negative decision was in all respects an EMA decision against which no legal remedies were available.

The applicant had therefore decided to bring the Commission before the CFI to contest the EMA’s decision. Remarkably enough, the Court did not refer to SNUPAT and DIR International Film and gave no reason whatsoever why the action was admissible, even if it was clearly not directed against the body that had adopted the contested decision.

While imputing the decision to the Commission could be supported, since the EMA adopted a decision following a procedure laid down in a Commission regulation, one could also dispute that the Commission had delegated powers, similarly to what was at issue in Inet Hellas (cf. section IV 5.5.1.3). After all, the Commission merely implemented the provisions of Article 15(3) and (4) of Regulation 2309/ 93,162 which specified that applicants should submit their appli-cation for a variation to the EMA and that the Commission would adopt arrange-ments for the examination of such applications. Granted, the Commission could have established a review procedure as hinted at by the Court in SNUPAT, but the Commission’s ‘minimalist’ role is also a general feature of the procedure.163

It may further be noted that the Commission’s regulation was adopted fol-lowing scrutiny by the representatives of the Member States in a comitology committee.164 Under the new EMA Regulation (726/ 2004), and following the entry into force of the Lisbon Treaty, the specific arrangements will be laid down in a delegated regulation.165 Following the wording of Article 290 TFEU, it could then be said that the Commission merely supplemented the existing rules contained in the regulation, in full respect of the essential elements laid down therein, including perhaps the role of the EMA.

161 Regulation (EC) 542/ 95 of the Commission, OJ 1995 L 55/ 15. This regulation has been repealed and ultimately replaced by Regulation (EC) 1234/ 2008 of the Commission, OJ 2008 L 334/ 7.

162 Current Articles 16 and 41 of Regulation (EC) 726/ 2004.163 See Article 5(2) of Regulation (EC) 542/ 95. 164 See n 161.165 Regulation (EU) 1235/ 2010 of the European Parliament and of the Council, OJ 2010 L 348/ 1.


These elements taken together might indicate that there is no real delegation of powers, but only the Commission further building on the regime established by the legislator. This difference in legal context between Thomae v. Commission and SNUPAT and DIR International Film would suggest that the decision of the EMA could not be imputed to the Commission as the Court did.

In Schering- Plough the same issue was again brought before the Court,166 but the applicant directed the action against both the Commission and the EMA, given the legal uncertainty characterizing the relationship between them. The Court therefore had to rule on the admissibility. Referring back to Thomae v. Commission, it proceeded by looking at the admissibility of the action in so far as it was directed against the EMA. Here the Court could have endorsed the ar-gument put forward by the EMA, by simply observing that Article 230 EC only mentioned the Commission and not the EMA. The Court decided differently and remarked that the basic Regulation only provided for advisory powers for the EMA; the refusal of Schering- Plough’s application could then not be a refusal by the EMA but should be deemed to emanate from the Commission,167 the action directed against the EMA being inadmissible.

This part of the Court’s judgment raises a number of issues. For one, without special scrutiny, the Court simply assumed the basic Regulation only granted ad-visory powers to the EMA. Secondly, it is unclear whether the contested decision could be imputed to the Commission solely by virtue of the fact that the EMA did not have binding powers, or whether the Court also believed the powers in question had been delegated by the Commission. The reasoning in SNUPAT and DIR International Film depended on the delegation, but in Schering- Plough the Court seemed to emphasize the fact that certain bodies cannot adopt binding decisions per se if the acts establishing such bodies do not provide so. This leads to the problem noted in relation to the Court’s reasoning in Sogelma (discussed previously) where the Court emphasized the nature of the powers exercised by the EAR and EMA, even if the Commission had delegated powers to the EAR.168 This issue also remains somewhat relevant following the Lisbon Treaty, since the technique of imputing decisions has not only been used by the Court in order to remedy lacunae in legal protection.169

As was mentioned in the previous section, the technique of imputing a con-tested act to a body other than its author has some shortcomings, meaning it should only be used in exceptional circumstances. In general, the technique is an exception to the rule that proceedings should be brought against the author of the act. Invoking such an exception should therefore be justified, for example by relying on the principle of effective judicial protection, and should also be

166 Case T- 133/ 03, n 146. 167 Ibid, para 22.168 This issue could have been clarified in CSL Behring v. Commission and EMA, but the Court

simply dismissed the case on the merits, without ruling on the admissibility. See Case T- 264/ 07, CSL Behring v. Commission and EMA, [2010] ECR II- 4469.

169 See Case T- 49/ 04, Hassan v. Council and Commission, [2006] ECR II- 52, para 59. This issue could have been addressed in Case T- 48/ 14 but that case was discontinued.


proportional, for example in the light of the principle of legal certainty. This is because the technique blurs the lines of accountability.

In SNUPAT and DIR International Film the exception could indeed be jus-tified, given the nature of the bodies to which the Commission had genuinely delegated powers. However, for the cases related to the EMA this seems much less the case. The problems involved in imputing decisions are further illustrated by the provision of Article 233 EC (current Article 266 TFEU), which raises the question of how the Commission could comply with the requirement contained therein in the absence of a right of injunction and without breaching the agency’s autonomy.

6.1.1.3 Filling the gap by providing for review mechanisms in the agency’s establishing regulation

A third way to ensure legal protection is to provide for legal remedies in the estab-lishing regulations of the agencies. However, the institutions have not been wholly consistent in implementing this technique, as it may even be found in the regula-tions of agencies without decision- making powers. The establishing act may then provide that an appeal may be lodged before the Commission, directly before the Court, or before an internal board of review before the Court may be seized. Yet, directly broadening the Court’s jurisdiction under secondary law is questionable since it amounts to an amendment of primary law through a procedure different from that provided for in Article 48 TEU.170

6.1.1.3.1 Legality review by the CommissionA first technique is to provide for an appeal procedure before the Commission.

In those cases, the regulation provides some parties with a right to have the le-gality of the agency’s acts reviewed by the Commission. The Commission must then take a decision within a specified time limit, whereby a failure to do so amounts to a dismissal of the case. Although this makes the whole system ‘water-tight’, it has been noted that the possibility of an implied dismissal does not create many incentives for the Commission to actually undertake a thorough examination.171

Whatever the outcome of the Commission’s examination, the resulting deci-sion is challengeable before the Court, creating a two- staged review which ultim-ately ensures access to the judge. Fischer- Appelt notes that because the regulations provide the Commission’s examination is on the legality of the agency decision, the supervision exercised by the Commission is not of a political nature.172

170 Opinion No 8/ 2001 of the Court of Auditors, OJ 2001 C 345/ 1, para 22; Ellen Vos, ‘Agencies and the European Union’, in Zwart and Verhey (eds), Agencies in European and Comparative Perspective, Antwerpen, Intersentia, 2003, pp 140– 1. In fine this was also remarked by the CFI (in a different context): see Case T- 77/ 01, Diputación Foral de Álava e.a. v. Commission, [2002] ECR II- 81, para 31.

171 Dorothee Fischer- Appelt, Agenturen der Europäischen Gemeinschaft, Berlin, Duncker & Humblot, 1999, p 312.

172 Ibid, p 311. Similarly, see Lenaerts, n 134, p 45.


In practice, the technique has only been used for a handful of agencies.173 The regulations of the first generation agencies were also the first to provide for this kind of review, even if they are non- decision- making agencies. Thus Article 18 of the Cedefop174 and Article 22 of the Eurofound175 regulations provide that the Member States, the members of the Board, and third parties directly and person-ally involved may ask the Commission to examine the legality of the agencies’ acts.176 Parties must do so within fifteen days and the Commission must take a decision within one month.

This technique, although in itself a very elegant and legally sound solution to the problem, has not been followed for other agencies. Only in respect of five other agencies can a similar procedure be found.177 Interestingly, the Commission Legal Service had advised including the same provision in the EEA regulation,178 referring to the current Article 13 TEU, noting that the EU should fulfil its tasks through its (then) four institutions (see section IV 0) and to Meroni, noting that the Court’s jurisprudence requires that the institutions remain responsible for the EU’s action and that the legal protection of private parties should remain unaffected.179

Theory aside, these procedures have hardly been used in practice since no ad-ministrative appeals have been lodged against acts of Cedefop, Eurofound, and the European Centre for Disease Prevention and Control (ECDC).180 Until re-cently, the same was true for the European Food Safety Authority (EFSA), al-though some four procedures have now been initiated,181 and the Commission’s decision in one of them also figures in a case before the General Court.182 As concerns the Community Plant Variety Office (CPVO), the Commission did once threaten the Administrative Board with starting this procedure. Following this threat, the Board amended its decision. In the case of the EU- OSHA, the procedure has been initiated once by a legal person following the rejection of its

173 In the original proposal for the EMA a similar procedure was also provided for, but it was not retained in the final act: see Article 59 in European Commission, COM (90) 283 final.

174 Regulation (EEC) 337/ 75, OJ 1975 L 39/ 1.175 Regulation (EEC) 1365/ 75, OJ 1975 L 139/ 1.176 It may be noted that the Parliament cannot lodge an appeal before the Commission. Although

this may be explained by the fact that these agencies were established pursuant to Article 253 EEC, today this is an anomaly. Under the ECDC Regulation, for instance, Parliament is represented on the Board and Board members have a right of appeal to the Commission.

177 See Articles 22 EU- OSHA, 44 CPVO, 122 OHIM, and 28 ECDC. For the OHIM, this appeal procedure has now been repealed by Regulation (EU) 2015/ 2424 of the European Parliament and of the Council, OJ 2015 L 341/ 21. For the EFSA, see Article 36 of Regulation (EC) 1829/ 2003, OJ 2003 L 268/ 1. See also Article 14 of Regulation (EC) 1935/ 2004, OJ 2004 L 338/ 4. These Articles explicitly provide that the Commission, in its decision, may require the EFSA to withdraw or undo the contested act. The case of the CPVO is rather special: Apart from the internal Board of Appeal (discussed presently), its establishing regulation also provides for a separate review by the Commission in relation to the acts of the President that are otherwise uncontrolled and the acts of the Administrative Council in relation to the CPVO’s budget.

178 See Service Juridique de la Commission, JUR (89) D/ 3370.179 See Service Juridique de la Commission, JUR (89) D/ 3901.180 Situation as of 13 December 2011. Replies of the Commission following a request for docu-

ments by the author, on file with author.181 Information as of 6 November 2014. 182 See n 342.


tender by the EU- OSHA. The Commission confirmed the EU- OSHA’s decision following its examination,183 and the applicant did not appeal the Commission’s decision before the General Court.

a. Conclusion The technique of entrusting the Commission with a form of administrative oversight over the functioning of an agency is an elegant and legally sound solution, but also has certain drawbacks. A first is that, in so far as the parties referred to mirror the original Article 173 EEC (or Article 230 EC), they also crystallize that Article in secondary legislation, while Article 263 TFEU has broadened access to the judge, especially for natural and legal persons. That the Parliament is not represented on most agencies’ Boards while it is a privileged party under Article 263 TFEU is evidently also problematic. Furthermore, the provisions in secondary legislation grant a procedural right to the members of the Boards themselves, not to the entities which these members represent.

Secondly, Article 263(5) TFEU provides a special exception for natural and legal persons challenging agencies’ acts, raising the question of whether the proce-dures providing in a remedy before the Commission also come under the scope of this exception. This will be discussed later (cf. section V 6.1.1.4.1.a).

A final (potential) drawback is more of a political than of a legal nature. While the agencies have been established as independent bodies,184 a review by the Commission may undermine that independence even if the review is purely legal.185 In general terms it is confusing to see the institutions put much effort and resources into the establishment of new EU bodies with distinct legal personality, only to give potentially very far- reaching powers of oversight to the Commission. At the same time it should be noted that this oversight may be seen as necessary and positive. This is so, for instance, under a strict reading of Meroni as advanced by the Commission Legal Service during the negotiations on the EEA (as dis-cussed previously). In addition, if in the future a model of EU administration is chosen whereby the Commission is the head of a unitary executive (at EU level), such an oversight would flow naturally from the Commission’s function. In any event, the current provisions in secondary legislation require updating, especially in the light of the new Article 263 TFEU.

6.1.1.3.2 Legality review directly by the CourtA second way of ensuring full legal protection against the acts of agencies is to provide for the possibility of a direct appeal before the Court in the establishing regulation. Provisions to this end may be seen in the establishing acts of many and very different types of agencies. The technique is very straightforward and

183 Commission letter dated 15 January 2010 addressed to the applicant, Mr. Snoeij, confirming the original EU- OSHA decision, on file with the author.

184 For an emphasis on this independence, see the fifth consideration of the preamble to the Cedefop Regulation, Articles 8 and 10 of the Eurofound Regulation, consideration 18 of the pre-amble to the OHIM Regulation, consideration 14 of the preamble to the ECDC Regulation.

185 See n 172.


foresees a provision in secondary law referring directly to the Treaty provision on the action for annulment or mirroring that provision.186

The technique was first applied in the Regulation establishing the EMCDDA, of which Article 17 read: ‘The Court of Justice shall have jurisdiction in actions brought against the Centre under the conditions provided for in Article 173 of the Treaty.’187 This provision has been preserved in Article 22 of the current recast Regulation, the reference to Article 173 EEC being replaced by a reference to Article 230 EC.188 According to the General Court this reference must be read as a reference to ‘the provisions governing the annulment action’,189 that is, Article 263 TFEU, while the reference to ‘the Court of Justice’ must be interpreted as a reference to the institution, not the specific Court.190 These issues of interpret-ation already show the inherent problems of a system of judicial protection set up by the legislator in parallel to the system laid down in the Treaties.

A reference to Article 230 EC has also been included in the acts setting up the EMA,191 European Institute of Innovation and Technology (EIT),192 and European Chemicals Agency (ECHA).193 In addition, the regulation of the latter agency also provides: ‘The Agency shall be required to take the necessary meas-ures to comply with the judgment of the Court of First Instance or the Court of Justice.’ It thus reproduces the former Article 233 EC (current Article 266 TFEU). The addition of this third paragraph might appear to be superfluous but its addition results from the agencies not being EU institutions. Before the entry into force of the Lisbon Treaty the ECB was not an institution of the EU either. As a result, the last provision of Article 233 EC read: ‘This Article shall also apply to the ECB.’ This further clarified that the institutions referred to in the first paragraph of the Article were indeed to be understood as the institutions stricto sensu.

This solution adopted by the legislator in the ECHA regulation has been rep-eated in the establishing regulations of the more recent agencies with compar-able powers to the ECHA, the Agency for the Cooperation of Energy Regulators (ACER), and the ESAs. The references to Article 263 TFEU in the ESAs’ Regulations then seem completely superfluous but may be explained by the fact that the original proposals for the ESAs’ Regulations were made in September 2009,194 three months before the entry into force of the Lisbon Treaty, and

186 The latter solution may be found in the CPVO and EASA Regulations. See Articles 73 CPVO and 50 EASA.

187 Regulation (EEC) 302/ 93 of the Council, OJ 1993 L 36/ 1.188 Regulation (EC) 1920/ 2006, OJ 2006 L 376/ 1.189 Case T- 63/ 06, Evropaïki Dynamiki v. EMCDDA, [2010] ECR II- 177, para 1.190 Ibid, paras 32– 3. The General Court had already ruled as such in Evropaïki Dynamiki

v. EMSA (see n 148) but in the EMCDDA case it did not refer to the EMSA case.191 See Article 73a as amended by Regulation (EC) 1901/ 2006. In Nycomed Danmark v. EMA

the General Court heard a case against the EMA involving a decision adopted under Regulation 1901/ 2006. See Case T- 52/ 09, Nycomed Danmark v. EMA, [2011] ECR II- 8133.

192 See Article 12(5) EIT, under the heading ‘liability’. 193 See Article 94(1) ECHA.194 See European Commission, COM (2009) 501 final; European Commission, COM

(2009) 502 final; European Commission, COM (2009) 503 final.


still referred to Article 230 EC. Still, as will be discussed later (cf. section V 6.1.1.3.3.b), the reference to Article 263 TFEU is not only superfluous but also creates legal uncertainty.

a. Conclusion The most straightforward solution, that is, including a mini- Article 230 EC in the establishing regulation by either copying the provisions of that Article or by referring directly to it, has been found to be problematic as a technique in two regards. Firstly, broadening the jurisdiction of the Court through secondary law is a doubtful practice from a legal point of view. Secondly, the flu-idity of the provisions in primary law to which the provisions in secondary law refer causes legal uncertainty. Now that Article 263 TFEU explicitly provides for action against agencies’ decisions and foresees that secondary law may make specific arrangements for actions of private parties, the legislator should omit other refer-ences to direct actions before the Courts in its regulations establishing agencies.

6.1.1.3.3 Legality review by internal organsThe last and most elaborate technique used by the legislator is the one where in-ternal review bodies are established within the agencies that decide on adminis-trative appeals brought by private parties. First, the characteristic features of these bodies will be looked at.

a. Features of the Boards of Appeal The following agencies have been set up with internal Boards of Appeal (BoAs): the Office for Harmonization in the Internal Market (OHIM), CPVO, European Aviation Safety Agency (EASA), ECHA, ACER, ESAs, and Single Resolution Board (SRB). In addition, under the fourth railway package the ERA would also house a BoA.195 The Board of Appeal established for the ESAs is in fact a single joint Board of Appeal. This institutional curiosity will be dealt with later.

Although the Boards of Appeal are all provided for in the establishing Regulations, the latter do not explicitly define the nature of these bodies. According to Barents, they are not administrative courts, but administrative review bodies.196 Fischer- Appelt on the other hand sees them as quasi- judicial bodies.197 In the Baby- Dry case, the General Court noted that there is a con-tinuity in terms of their functions198 between the body within the agency making the first decision and the BoA,199 even if some BoAs have not grasped all the consequences of this (cf. section V 6.1.1.4.1.b). The Court of Justice later also juxtaposed the review procedures within the agency, of which the BoA forms an integral part, and the judicial review exercised by the Courts.200 The

195 See European Commission, COM (2013) 27 final.196 René Barents, Procedures en procesvoering voor het Hof van Justitie en het Gerecht van eerste

aanleg van de EG, Deventer, Kluwer, 2005, p 45.197 Fischer- Appelt, n 171, p 314.198 Opinion of AG Sharpston in Case C- 29/ 05 P, OHIM v. Kaul GmbH, [2007] ECR I- 2213,

para 29.199 Case T- 163/ 98, The Procter & Gamble Company v. OHIM, [1999] ECR II- 2383, para 38. This

was also confirmed for non- OHIM Boards of Appeal: see Case T- 102/ 13, Heli- Flight GmbH & Co. KG v. EASA, ECLI:EU:T:2014:1064, para 27.

200 Case C- 29/ 05 P, OHIM v. Kaul GmbH, [2007] ECR I- 2213, para 51.


General Court also noted this in Calypso v. Calpico when it ruled that a party could not rely on the right to a fair hearing under Article 6 ECHR, since pro-ceedings before the Boards of Appeal are administrative and not judicial in nature.201 This would also mean that proceedings before the Boards of Appeal do not qualify as legal proceedings under Article 228 TFEU, which would otherwise bar the Ombudsman from opening enquiries. Navin- Jones rightly concludes from this that while the BoAs may not be courts, it would go against their innate function and purpose to qualify them as merely administrative bodies.202

The different Boards of Appeal all fulfil the same basic function, yet a lot of differences exist in their organization and functioning. As regards the ACER and ESAs, however, there are signs of a genuine model, since the provisions on their Boards of Appeal are virtually identical.203 As regards the composition of the BoAs the standard is three members,204 but ACER and ESAs’ Boards of Appeal are composed of six members, while that of the SRB has five members. Since the Boards fulfil a legal review function in a sector- specific context, the qualifications of the members reflect this, with both legal and technical expertise present on the Board.205 For the older agencies it is further provided that the Commission has the power to determine further organizational measures through implementing acts. For the ACER, ESAs, and SRB it is provided that they themselves will adopt their own Rules of Procedure.

For organizational reasons most, but not all, Boards are assisted by their own registry.206 Decisions of the Boards may be taken by a majority vote;207 in the case of the ACER and ESAs this means at least four out of six votes. An additional re-quirement for the ESAs’ Board is that such a majority should include the vote of at least one of the members of the Board appointed by the agency whose decision is contested.208

201 Case T- 273/ 02, Krüger GmbH & Co. KG v. OHIM, [2005] ECR II- 1271, para 62. See also Case T- 63/ 01, The Procter & Gamble Company v. OHIM, [2002] ECR II- 5255, paras 22– 3.

202 Marcus Navin- Jones, ‘A Legal Review of EU Boards of Appeal in Particular the European Chemicals Agency Board of Appeal’, (2015) 21 EPL 1, p 145.

203 The ACER BoA’s decisions shall be published while those of the ESAs’ BoA shall both be reasoned and published. A more important difference relates to the power vested in the BoA to take decisions in lieu of the agency, discussed further on.

204 In the case of the OHIM there is also an enlarged BoA composed of nine members, or the possibility for a single- member BoA.

205 See Articles 135(1) OHIM, 11(2) of the CPVO Implementing Regulation 874/ 2009, 2(1) of the EASA Implementing Regulation 104/ 2004, 1(1) of the ECHA Implementing Regulation 1238/ 2007, 58(2) ESA, and 85(2) SRB. The exception here is ACER as Article 18(1) ACER only refers to expertise in the energy sector.

206 This is normally provided in the implementing regulations. See Article 5 OHIM IR, Article 5(1) ECHA Implementing Regulation 771/ 2008. Sometimes it is provided the Director will set up a registry: see Article 12(1) CPVO IR, Article 5(1) EASA IR. The registry for the ACER is established under the BoA’s Rules of Procedure. The ESAs’ BoA does not have a registry but its functions are performed by the Joint Committee: see Article 58 (8) ESAs.

207 See Article 13 OHIM IR, Article 11(4) CPVO IR, Article 7(1) EASA IR, Article 20 ECHA IR 771/ 2008, Article 18(1) ACER, Article 58(6) ESA.

208 See Article 58(6) of the ESAs Regulations.


The rules on the appointment of the members of the Boards are more diver-gent. In the case of the OHIM and CPVO a differentiation is made between the President/ chairmen and the other members of the Board.209 For the Boards of the other agencies there is no such differentiation and a general rule applies whereby the Management Board appoints the members of the Board of Appeal, follow-ing a proposal of the Commission,210 except in the case of the SRB where the Commission is not involved.211 Because the Board of Appeal of the ESAs is a joint body, the three agencies only nominate one- third each of its members.212

The terms of the members of all the Boards is five years and they are all renew-able for five years.213 The regulations are silent on the number of times a term may be renewed, except for those of the ECHA, ESAs, and SRB which foresee that terms may only be renewed once.214 That fact that the members of the Boards of Appeal are supposed to be independent yet depend on the Management Boards of the agencies for their re- appointment is criticised by Thomas.215 Even if the Boards of Appeal do not review the decisions of the Management Boards but only those of the agencies’ experts, it would indeed seem advisable to keep the Board of Appeal members as independent as possible from all other bodies within the agency.

b. Procedure before the Board of Appeal Logically, the technique of establishing Boards of Appeal has only been used for those agencies which have been granted important decision- making powers directly affecting the legal position and in-terests of private parties. The OHIM and the CPVO were the first such agencies.

In 2010, 179,954 applications were filed at the OHIM and 2,886 applica-tions were filed at the CPVO.216 In the same year 2,423 appeals were lodged against decisions of the OHIM and 19 against decisions of CPVO.217 It is quite clear that it would be impossible for the General Court to handle all these cases. Because of the high number of appeals before the OHIM, more than one BoA has been established within the OHIM.218 The CPVO only has

209 The Chairmen of the OHIM BoAs are appointed by the Council from a shortlist drawn up by the Administrative Board while for the CPVO the decision is made by the Commission follow-ing an opinion of the Administrative Board. For both agencies, the other members are appointed by the Administrative Board, in the case of the CPVO following a proposal of the Office itself. See Articles 136 OHIM and 47 CPVO.

210 See Articles 41(3) EASA, 89(2) ECHA, 18(2) ACER, 58(3) ESAs. In the case of the ECHA, ACER, and the ESAs, the Commission drafts the proposal following a public call for expressions of interest. In the case of ACER and the ESAs it is furthermore provided that before the Management Board appoints the BoA members, it will consult the Board of Regulators/ Supervisors.

211 See Article 85(2) SRB. 212 See Article 58(3) ESAs.213 See Articles 136(1) (2) OHIM, 47(1) (2) CPVO, 42(1) EASA, 90(1) ECHA, 18(3) ACER,

58(4) ESAs.214 See Articles 90(1) ECHA, 58(4) ESAs, and 85(2) SRB.215 David Thomas, ‘European Chemical Agency Board of Appeal decisions in Honeywell and

Dow Chemicals’, (2013) 20 MJECL 4, p 612.216 OHIM, Annual Report 2010, pp 7– 8; CPVO, Annual Report 2010, p 32.217 OHIM, Annual Report 2010, p 8; CPVO, Annual Report 2010, p 87.218 The Rules of Procedure of the OHIM Boards of Appeal allow the Presidium of the Boards

of Appeal to establish new Boards. See Article 1(6) of Regulation (EC) 216/ 96 of the Commission, OJ 1996 L 28/ 11.


one BoA,219 although the Administrative Council may decide to establish add-itional BoAs.220

The jurisdiction of the Boards of Appeal of the OHIM and CPVO is defined differently. The jurisdiction of the former is defined ratione personae, that is, depending on which agency body adopted the decision.221 For the CPVO the jurisdiction is defined ratione materiae, that is, depending on the legal basis in the Regulation of the decision,222 and the latter example has been followed for the other agencies.223 The difference is not unimportant since it means the OHIM BoA would only have to verify whether it has jurisdiction in relation to the author of a contested act, while the CPVO BoA must identify the legal basis of the contested act.224 Here it may also be noted that the Joint BoA of the ESAs and that of the SRB are the only ones which also have jurisdiction to rule on appeals against refusals for access to documents under Regulation 1049/ 2011.225

For the oldest agencies appeals have suspensory effect by default,226 but in the case of the CPVO, the agency may order the decision not to be suspended. This possibility does not exist for the OHIM. For the other agencies, except the ECHA,227 the appeal generally does not have suspensive effect, although the EASA (but not the EASA BoA) may decide to suspend the application of the con-tested decision,228 while in the case of the ACER, the ESAs, and the SRB the Board of Appeal takes this decision.229

The persons entitled to bring proceedings before the BoAs also differ between the OHIM and CPVO Regulations. Article 59 OHIM provides that ‘[a] ny party to proceedings adversely affected by a decision may appeal. Any other parties to the proceedings shall be parties to the appeal proceedings as of right.’ However, for the CPVO the legislator borrowed the wording of the Treaty, and the CPVO example was again followed for the more recent agencies.230 Article 68 CPVO provides that ‘[a]ny natural or legal person may appeal, subject to Article 82, against a decision, addressed to that person, or against a decision which, although in the form of a decision addressed to another person, is of direct and individual

219 Also because of the possibility of interlocutory revision (discussed below), not every appeal is remitted to the CPVO BoA.

220 See Article 11(1) of the CPVO IR 874/ 2009.221 See Article 58 OHIM: those bodies are the Examiners, the Opposition Divisions, the

Administration of Trade Marks and Legal Division, and the Cancellation Divisions.222 See Article 67(1) CPVO.223 See Articles 44 EASA, 91 ECHA, 19 ACER, 60 ESA, 85(3) SRB.224 See also Case T- 36/ 09, dm- drogerie markt GmbH & Co. KG v. OHIM, [2011] ECR II- 6079,

where the General Court noted that the original decision of the Opposition Division did not have any legal base in the regulation. This had not precluded the OHIM BoA from accepting jurisdic-tion. Differently see Case T-660/14, SV Capital v. EBA, ECLI:EU:T:2015:608, para 72.

225 Other BoAs do not have this jurisdiction: see for instance EASA Board of Appeal, Case AP/ 04/ 2013, Robinson Helicopter Company, para 100.

226 Except for appeals brought before the BoA of CPVO against decisions taken pursuant to Articles 29 or 100(2) CPVO.

227 See Article 91(2) ECHA.228 See Article 44(2) EASA. 229 See Articles 19(3) ACER, 60(3) ESA, 85(6) SRB.230 See Articles 45 EASA, 92 ECHA, 19 ACER, 60 ESA, 85(3) SRB.


concern to the former’. Thus even private persons which were not a party to the proceedings before the agency may appeal a decision if they can show they are directly and individually concerned by the decision.231

The other requirements for filing an appeal before the BoAs of the OHIM and the CPVO are similar. In both cases, appellants have two months to file an appeal following notification (or publication).232 The appellant then has another two months to complete its written statement. For most other agencies there is a single period of two months,233 but for the ECHA it is three months,234 and for the SRB six weeks.235 The ACER and ESAs’ regulations further impose a very strict time limit on the BoA itself, having to decide on an appeal within two months after it was lodged; for the SRB this is one month.

Because a Board of Appeal is more accessible than the ordinary (EU) judge, fees are payable when appeals are lodged, to limit frivolous action and to fund the functioning of the Board of Appeal system.236 The fees are not simply a re-flection of the cost of proceedings and they vary greatly between the agencies.237 In the case of the EASA and ECHA a multiplier is also applied depending on the type of applicant (eg SME or multinational). As a rule, fees are refundable when interlocutory revision is granted or when the appeal is successful. However, for the OHIM a substantial procedural violation needs to be shown as well in the case of a successful appeal.238 Apart from these differences it is clear this system was inspired by the regime set up by the European Patent Convention, the appeal procedure of which is strikingly analogous to those of the CPVO and the OHIM.239

In the appeals against OHIM and CPVO decisions, the agency concerned may rectify the decision if it considers the appeal to be admissible and well- founded, before the appeal is remitted to the BoA. For both agencies a differentiation is made between ex parte and inter partes proceedings. However, in the case of the CPVO there is an absolute bar on interlocutory revision in inter partes cases,240 an

231 The parallelism between Article 263 TFEU and the provisions referred to in the footnote above was also noted by the Boards of Appeal: see inter alia EASA Board of Appeal, Case AP/ 03/ 2012, Luck, paras 43– 5; ESAs Joint Board of Appeal, Case 2013- 8, SV Capital OÜ v. EBA, para 24.

232 See Articles 60 OHIM and 69 CPVO.233 See Articles 46 EASA, 19(2) ACER, 60(2) ESA. 234 See Article 92(3) ECHA.235 See Article 85(3) SRB.236 However, fees are not foreseen for appeals against decisions of the ACER, ESAs, and the SRB.237 See Regulation (EC) 2869/ 95 of the Commission, OJ 1995 L 303/ 33 (OHIM); Regulation

(EC) 1238/ 95 of the Commission, OJ 1995 L 121/ 31 (CPVO). For the EASA the basic fee is ad-justed with a multiplier depending on the applicant: see part III to the Annex of Regulation (EU) 319/ 2014, OJ 2014 L 93/ 58. For the ECHA the fees also depend on the type of applicant: see Annex VIII to Regulation (EC) 340/ 2008 of the Commission, OJ 2008 L 107/ 6.

238 Article 51 of Regulation (EC) 2868/ 95 of the Commission, OJ 1995 L 303/ 1. Costs need not be claimed in the form of order but specific procedural violations need to be raised. See Joined Cases T- 124/ 02 & T- 156/ 02, The Sunrider Corp. v. OHIM, [2004] ECR II- 1149, paras 69– 70.

239 See Ian Muir, Matthias Brandi- Dohrn, and Stephan Gruber, European Patent Law: Law and Procedure under the EPC and PCT, Oxford, OUP, 2002, pp 254– 73.

240 See Article 70(1) CPVO.


example also followed in the EASA and ECHA Regulations.241 Article 62 OHIM permits such a revision if the other party or parties agree.

During the proceedings, parties as a rule have the right to make oral representa-tions, except in the case of appeals against OHIM decisions where oral representa-tions will depend on whether the Board of Appeal finds this useful.242

A major distinctive feature of the Boards of Appeal system is that the Boards are competent to remake the original decision when an appeal is found admissible.243 However, they may also decide to remit the case to the original department for fur-ther action. The latter is then bound by the ratio decidendi of the BoA’s decision. The exceptions here are the Joint BoA for the ESAs and the Appeal Panel of the SRB.244 The former can only confirm a contested decision or remit it to the ESA concerned, precisely because it is a joint body of the ESAs.245

This characteristic again shows the attractiveness of the technique of establishing BoAs within agencies. In relation to the OHIM it was noted that BoAs were neces-sary given the sheer number of appeals. For other agencies,246 that number is more limited, but the BoA instrument remains interesting because it allows a review of a decision by experts in a timeframe which judicial authorities could not match.247 Further, the parties’ representatives before the BoAs do not need to be lawyers.248 The model in the ACER and ESAs’ regulations further provides that the decisions shall be made public. Such an expression provision is lacking for the other agencies, even if this requirement may still be imposed in the Board of Appeal’s Rules of Procedure.

After an appeal has been decided by a Board of Appeal, its decision, but not the original decision,249 can be challenged before the Court of Justice.250

241 See Article 47 EASA. For the ECHA Article 93(1) of the Regulation generally provides the Director may rectify the decision without a reference to ex parte or inter partes proceedings. For the ACER and ESA this possibility is not provided for.

242 See Articles 77(1) OHIM, 71 CPVO, 48 EASA, 13(1) ECHA IR 771/ 2008, 19(4) ACER, 60(4) ESA, 85(7) SRB.

243 See Articles 64 OHIM, 72 CPVO, 49 EASA, 93(3) ECHA, 19(5) ACER.244 See Article 85(8) SRB. 245 See Article 60(5) ESA.246 In its original ECHA Staff model, however, the Commission projected a total of 549 cases in

2010, while in reality only one appeal was lodged. See European Commission, REACH Operational Staff Model Summary— Agency, on file with the author.

247 Further, Navin- Jones remarks that following the Lanxess case, a Board of Appeal may en-tertain claims that do not challenge the legality of a decision, something a Court could not do. See Navin- Jones, n 202, pp 149– 50. Whether this conclusion is correct remains to be seen. After all, in Lanxess, the BoA ignored the plea of inadmissibility (raising this issue) and held that the action was unfounded in any event. See ECHA Board of Appeal, Case A- 004- 2012, Lanxess, paras 54– 61.

248 See Lucas Bergkamp, Nicolas Herbatschek, Suriya Jayanti, and Paul Craig, ‘Dispute Resolution and Legal Remedies’, in Bergkamp (ed), The European Union REACH Regulation for Chemicals— Law and Practice, Oxford, OUP, 2013, p 265; Marco Bronckers and Yves Van Gerven, ‘Legal Remedies under the EC’s New Chemicals Legislation REACH’, (2009) 46 CMLRev 6, pp 1844– 5.

249 In Heli- Flight the appellant had challenged the original decision before the General Court. The latter only held the action admissible in so far as directed against the BoA decision: see Case T- 102/ 13, n 199, paras 30– 2.

250 See Articles 73 CPVO and 65 OHIM. The original Article 73 CPVO has been amended to be more in line with the provision in the OHIM Regulation. See the explanatory memorandum in European Commission, COM (95) 144 final, p 2.


Evidently, the review of the Court is confined to the review of the legality of the Board’s decision and it must assess the case as it was brought before the Board of Appeal.251 Apart from resulting from a general limitation on the Court’s competence, this is part of the logic of the whole review procedure which aims to put the brunt of the proceedings at the first stages of the review for reasons of procedural efficiency.252

Because an agency itself is not a party to the decision of the BoA, it cannot itself appeal a decision of the BoA. However, it need not support the BoA’s decision before the General Court if it believes the BoA erred in its assessment.253

Article 266 TFEU is reproduced in the establishing regulations since the agencies must take the necessary measures following a judgment of the Court of Justice.254 If the Court annuls the contested decision, the case will normally be referred back to the Board of Appeal that adopted the annulled decision, but the Court may also refer the case back to the department which made the first decision.255 Because of the obligation on the agency to comply with the Court’s ruling, the Court itself is not entitled to issue directions to the agency.256 As a result, although the Court may for example annul a decision of a BoA rejecting an application for a trade mark or plant protection right, it cannot order the agency to grant such an application.257

When the General Court is seized of a case against a decision of a Board of Appeal, it does not act as the court of last instance. This means that a further appeal on points of law against a decision of the General Court is open before the Court of Justice under the conditions laid down in Title IV of the ECJ Rules of Procedure.

As regards appeals against decisions of the Boards of Appeal, the ESAs were the first decision- making agencies established post- Lisbon. It would have been logical therefore if the relevant provisions in the establishing regulation were adapted to the new provisions in Article 263 TFEU. However, this is not the case, Article 61 of the Regulations inter alia reading:

1. Proceedings may be brought before the Court of Justice of the European Union, in accordance with Article 263 TFEU, contesting a decision taken by the Board of Appeal or, in cases where there is no right of appeal before the Board of Appeal, by the Authority.

2. Member States and the Union institutions, as well as any natural or legal person, may institute proceedings before the Court of Justice of the European Union against decisions of the Authority, in accordance with Article 263 TFEU.

251 Case C- 214/ 05 P, Sergio Rossi SpA v. OHIM, [2006] ECR I- 7057, para 50.252 See also the amendments to the General Court’s Rules of Procedure, OJ 1995 L 172/ 3.253 Charles Gielen and Verena von Bomhard, Concise European Trade Mark and Design

Law, Alphen aan de Rijn, Kluwer Law International, 2011, p 199.254 See Articles 73(6) CPVO, 65(6) OHIM, 50(3) EASA, 94(3) ECHA, 20(3) ACER, 61(4) ESA.255 Case T- 106/ 00, Streamserve Inc. v. OHIM, [2002] ECR II- 723, para 18.256 Case T- 331/ 99, Mitsubishi HiTec Paper Bielefeld GmbH v. OHIM, [2001] ECR II- 433, para 33.257 Because there are several OHIM Boards of Appeal, Article 1(d) of the rules of procedure

provides further guidance on a referral following a CJEU ruling.


This Article resembles Articles 94 ECHA and 20 ACER, and the original Article as proposed by the Commission was even identical to the latter two.258 However, the Commission’s proposal was amended on this point by the Parliament, adding the current paragraph 2. As was noted previously (cf. sec-tion V 6.1.1.3.2), this paragraph seems superfluous: it simply reproduces the first part of paragraph 1, which in itself is largely superfluous because it repeats Article 263 TFEU. In addition, paragraph 2 also provides that all ‘Union insti-tutions’ may bring proceedings, at the same time referring back to Article 263 TFEU which does not grant standing to all ‘Union institutions’. Furthermore, paragraph 2 essentially makes Article 263 TFEU applicable, but that Article already applies by virtue of primary law. Apart from being superfluous, para-graph 2 also confuses the system of judicial scrutiny as set up following the Lisbon Treaty in so far as proceedings are brought by non- privileged parties. For these proceedings Article 263(5) TFEU firstly refers back to the specific conditions and arrangements laid down in the establishing regulation of the agency concerned, but in the cases of the ESAs, these refer back to Article 263 TFEU.

Strikingly, Article 86 of the Regulation establishing the SRB has been drafted identically to Articles 61 of the ESAs’ Regulations. While the Commission in its original proposal (rightly) did not include a provision on access to the Court,259 the Council still introduced it,260 this to ensure that Member States that do not participate in the Single Resolution Mechanism could still challenge the SRB’s decisions. However, exactly this scenario is already governed by Article 263 TFEU.261

c. Insufficient attention to the Boards of Appeal in the Common Approach on Decentralised Agencies The Common Approach does not devote much attention to the Boards of Appeal and their functioning. In fact, they are only mentioned under the heading ‘other internal bodies’ where the imparti-ality and independence of the members of the Boards is stressed. Therefore, the agencies are called upon to exchange best practices. Unfortunately, the Common Approach also contains no further principles on procedures before the Boards.

The discussion in the preceding sections has shown that there is considerable variation on a number of issues. Ideally these elements should be covered by a generally applicable framework on agencies.

In relation to the Joint Board of Appeal of the ESAs, there is also a more fun-damental question pertaining to the nature of this body. Since there is no specific

258 See Article 47 of the original proposals: European Commission, COM (2009) 501– 503 final.259 See European Commission, COM (2013) 520 final.260 See Article 77a of the Presidency Compromise, 17410/ 13.261 Still, this might be explained by the fact that at the timing of the adoption of the Regulation,

a case was pending before the General Court in which the UK’s capacity to bring proceedings against an act of the ECB was questioned, since the UK does not fully participate in the EMU. The General Court confirmed the UK’s standing. See Case T- 496/ 11, UK v. ECB, ECLI:EU:T:2015:133, paras 69– 76.


BoA for each of these agencies, the Joint BoA is specialized in matters of financial regulation, raising the question of whether it might in the future not be established as a specialized court of the EU under Article 257 TFEU. Indeed the Joint Board of Appeal already shares a lot of the features characterizing such a specialised court, apart from the presence of non- legal experts on the Joint Board of Appeal, which conflicts with Article 257(4) TFEU. If the competences of the ESAs together with those of the ECB and the SRB in the framework of the Banking Union were fur-ther enlarged, this might justify replacing the Joint Board of Appeal with a proper specialized Tribunal.

d. Conclusion The agencies’ Boards of Appeal are a good instrument to lessen the already significant caseload of the General Court and to allow for a flexible and efficient review mechanism to the benefit of private parties. Although this technique should in principle be supported, the overall system as implemented lacks coherence.

6.1.1.4 Filling the gap by amending the TreatiesThe fourth and final technique to address a lacuna in the Treaties’ system of ju-dicial protection is to align that system with the EU’s evolved institutional archi-tecture. This is something which was remarked upon by Louis as early as 1973 on the occasion of the establishment of the European Monetary Cooperation Fund (EMCF), the ECB’s precursor.262

Although some of the techniques discussed previously also had their merits, it is clear that this technique is the sole preferable for reasons of legal certainty, co-herence of the system of judicial review, respect for the hierarchy of norms in the EU legal order, and respect for the institutional balance.

6.1.1.4.1 An explicit base for judicial scrutiny of agency decisions: Article 263 TFEU

Paragraphs 1 and 5 of Article 263 TFEU (former Article 230 EC) now read as follows:

1. The Court of Justice of the European Union shall review the legality of legisla-tive acts, of acts of the Council, of the Commission and of the European Central Bank, other than recommendations and opinions, and of acts of the European Parliament and of the European Council intended to produce legal effects vis- à- vis third parties. It shall also review the legality of acts of bodies, offices or agencies of the Union intended to produce legal effects vis- à- vis third parties.

…5. Acts setting up bodies, offices and agencies of the Union may lay down specific

conditions and arrangements concerning actions brought by natural or legal per-sons against acts of these bodies, offices or agencies intended to produce legal effects in relation to them.

262 Jean- Victor Louis, ‘Le Fonds Européen de Cooperation Monetaire’, (1973) 9 CDE 3, pp 294– 5.


The amendments to Article 230 EC in Article 263 TFEU are completely identical to the changes which Article III- 365 of the Constitution would have introduced. The Treaty establishing a Constitution for Europe found its origins in the Laeken Declaration of 2001, which called for a European Convention. The latter pro-duced a draft text which served as a basis for negotiations in a classic intergovern-mental conference (IGC). During that conference the final text was agreed upon, and it was ultimately signed in Rome in 2004.

In the Convention, the deliberations on the necessary amendments to Article 230 EC were entrusted to the special Discussion Circle on the Court of Justice, in which it was noted that the regulations establishing the existing agencies con-tained clauses granting jurisdiction to the Court and, according to the Chairman, most of the Discussion Circle’s members supported a simplification by providing for the Court’s jurisdiction in an amended Article 230 EC.263 To get a better view of the existing ‘ad hoc’ system established by the different regulations, the Secretariat of the Convention prepared a working document setting out the dif-ferent solutions adopted by the legislator.264 The inconsistent approach was noted and a position was suggested in the working document that relied heavily on the position taken by the Commission in its 2002 Communication,265 which formed the basis of the ultimately failed interinstitutional agreement.

The working document suggested that Article 230 EC should be amended with an explicit reference to acts of agencies. Noteworthy in this regard is that, according to the Secretariat, such a reference would not ‘prevent the legislator from providing specific arrangements on a case- by- case basis’,266 but for reasons of legal certainty ‘a paragraph could be added to Article 230 EC to clarify that the legislator may establish specific arrangements for bringing proceedings against agencies’.267 In the end, the suggestions of the Secretariat were all taken over in the draft final report:

[T] he discussion circle recommends that Article 230 TEC be amended so as to cover, in addition to legal acts adopted by the institutions, those of the Union’s bodies and agencies […]; the act establishing the agency might also lay down specific arrangements for the exercise of control of the agency or body in question.268

The draft final report elicited quite a number of amendments from the members of the Discussion Circle, but only one member commented on the proposal made in relation to the agencies. Baroness Scotland of Ashtal suggested that the issue should be dealt with following a detailed examination of the individual position of each agency and following a consultation with each agency. More importantly,

263 Secretariat of the European Convention, Report on the Meeting of the Discussion Circle on the Court of Justice on 3 March 2003, 13 March 2003, CONV 619/ 03.

264 Secretariat of the European Convention, Right of Appeal against Agencies Created by Secondary Legislation, 10 March 2003, Working Document 9 of the Discussion Circle on the Court of Justice.

265 European Commission, COM (2002) 718 final.266 Working Document 9 of the Discussion Circle on the Court of Justice, para 6, see note 264.267 Ibid.268 Secretariat of the European Convention, Draft final report of Mr António Vitorino, 11 March

2003, Working Document 8 of the Discussion Circle on the Court of Justice, para 26.


she noted that the question of legal review of agency decisions could not simply be regarded as a ‘tidying up’ exercise.269 In the end, however, the suggestion in the draft was retained in the final report.270

The Convention took over the suggestion in its Draft Treaty as Article III- 270,271 which remained unamended on this point in the Treaty establishing a Constitution for Europe as Article III- 365. The latter was copied into the Lisbon Treaty as Article 263 TFEU.

The new Article 263 TFEU then raises a number of questions. The first is ob-viously whether the gap in legal protection is effectively bridged; a corollary to this question being whether the solution provided for by Article 263 TFEU will also work better than the more ad hoc solutions adopted by the legislator and the Court discussed previously. This will also provide further clarification on whether the majority in the Discussion Circle was right in viewing the amendments as a simple tidying- up exercise.

a. Is the lacuna in the system of legal protection ‘filled’? At first sight the answer to the first question is yes, if ‘lacuna’ refers to the gap which existed in legal pro-tection vis- à- vis the agencies compared to the other institutions or bodies under the former first pillar. This is so since Article 263 TFEU applies the same require-ments to actions against agencies, both as regards (i) the type of acts which may be challenged and (ii) the parties entitled to bring proceedings. Concerning the type of acts, Article 263 TFEU imposes the same conditions as those which apply to the other institutions without a significant independent law- making function: the Parliament and the European Council. These institutions’ and the agencies’ acts have to produce legal effects vis- à- vis third parties.272

However, this general conclusion is undermined by one important element which has formed the vantage point of this study. This is the problem of the Treaties’ silence on the precise role and function of the agencies in the EU ex-ecutive. Because of legal and political issues, there has been some reticence on the part of the legislator to grant de iure binding powers on the agencies save the power to take individually binding decisions.273 What have been granted to the agencies in return are de iure advisory powers which are de facto binding. These acts do not intend to produce legal effects vis- à- vis third parties, but only impose an obligation on the authority to which the advice is given. As a result, much of

269 Secretariat of the European Convention, UK Comments to the Draft Final Report, 14 March 2003, Working Document 19 of the Discussion Circle on the Court of Justice, p 3 at para 6.

270 Secretariat of the European Convention, Final Report of the Discussion Circle on the Court of Justice, 25 March 2003, CONV 636/ 03, para 26.

271 Secretariat of the European Convention, Draft Treaty Establishing a Constitution for Europe, 18 July 2003, CONV 850/ 03.

272 On this requirement, see Case T- 258/ 06, Germany v. Commission, [2010] ECR II- 2027, para 29. Further, by reasoned order the CFI has observed that the Parliament’s acts intended to produce legal effects should also be understood as acts that could exceed the limits which have been set on the competence of that institution: see Case T- 353/ 00 R, Le Pen v. Parliament, [2001] ECR II- 125, para 59.

273 See Merijn Chamon, ‘Le recours à la soft law comme moyen d’éluder les obstacles constitu-tionnels au développement des agences de l’UE’, (2014) RUE 576, pp 155– 7.


the acts of agencies do not qualify as acts intended to produce legal effects vis- à- vis third parties and would thus fall outside the scope of Article 263(1) TFEU. This issue will be discussed in more detail later (see section V 6.1.1.5).

Still, paragraphs 2 to 4 of Article 263 TFEU determine which parties are privi-leged, semi- privileged, or non- privileged and since their capacity is not deter-mined by the author of the contested act, access to the Court in actions brought against agencies is not more restricted than in actions brought against the insti-tutions. An important exception to this is paragraph 5, which provides for the possibility that acts establishing agencies may establish specific conditions and arrangements for challenges brought by natural and legal persons.

A number of remarks are in place here. Even though this paragraph was nec-essary in order to safeguard the existing multi- staged review provided for in the regulations of the decision- making agencies, it is phrased very broadly. To be sure, the specific conditions and arrangements may, not shall, be laid down, leaving the legislator the choice whether or not to use this option, which further only exists in relation to actions brought by natural and legal persons. Article 263(5) TFEU therefore does not affect the privileged and semi- privileged parties. However, within these limits, it seems that the secondary legislator has been granted almost carte blanche as regards the nature and scope of these specific conditions and ar-rangements. The only real limit under primary law would then be Article 47 of the Charter, which provides for the right to an effective remedy.

Paragraph 5 thus directs natural or legal persons back to the act establishing the agency, without itself imposing an obligation on such persons to exhaust all the review mechanisms provided for in the establishing regulation before seizing the Court itself. Whether or not such an obligation exists will depend on the wording of the provisions in the regulations themselves. As a result, in the case of Cedefop, Eurofound, EU- OSHA, EFSA, and ECDC, where a procedure before the Commission exists to check the legality of the acts of these agencies,274 these remedies would not need to be exhausted since it is only provided that the Commission may be asked to review a decision. On the other hand, exhausting the remedies before the Boards of Appeal would be required. This follows both from the logic of the multi- staged review itself and because the regulations first direct natural and legal persons to the Board of Appeal.

b. Proceedings under Article 263 TFEU in practice and in relation to proceedings before the Boards of Appeal Several cases have been lodged against the ECHA related to the procedure to include certain chemical products on the list of Substances of Very High Concern of Annex XIV to the ECHA Regulation.275 The Commission decision to put a substance on the list is subject to a comitology procedure but Article 59 of the ECHA Regulation provides that before inclusion

274 See nn 174, 175, and 177.275 See Case T- 93/ 10, Bilbaína de Alquitranes e.a. v. ECHA, ECLI:EU:T:2013:106; Case T-

94/ 10, Rütgers Germany e.a. v. ECHA, ECLI:EU:T:2013:107; Case T- 95/ 10, Cindu Chemicals e.a. v. ECHA, ECLI:EU:T:2013:108; Case T- 96/ 10, Rütgers Germany GmbH e.a. v. ECHA, ECLI:EU:T:2013:109.


on the list, the ECHA will establish a candidate list. A number of producers and suppliers of substances which had been put on the candidate list lodged actions before the General Court, the decision under Article 59 of the Regulation not being a decision that is challengeable before the Board of Appeal.

The first question which needed to be solved was whether the decision produced legal effects. Since the inclusion on the candidate list gives rise to a number of infor-mation obligations, the Court found that the ECHA’s decision was indeed challenge-able, applying a broader notion of binding legal effects than in its previous case law (cf. section V 6.1.1.5).276 The ECHA then argued that its decision was not a regu-latory act in the sense of Article 263 TFEU and could therefore not be challenged by the applicants. Specifically, the ECHA argued that under its mandate it cannot adopt regulatory acts, since the regulatory power is reserved to the Commission. The General Court, referring to Microban,277 rejected this argument.278 Evidently the agency invoked the idea that EU ‘regulatory’ agencies are in fact not competent, following Meroni, to regulate an economic sector.279 Unsurprisingly, however, the Court did not allow such a restricted notion of ‘regulation’ to influence its interpret-ation of the concept of ‘regulatory act’ under Article 263 TFEU.

As noted previously (see section IV 5.4.2.2), Rütgers is also one of those cases in which the General Court confirms the broad discretion which agencies such as the ECHA may exercise. This jurisprudence sits uneasily with the Court’s ob-servation in Short- selling that agencies’ powers should be ‘amenable to judicial review in the light of the objectives established by the [legislator]’.280 To what does this judicial review amount if the legislator’s objectives are rather general and the Court is required to defer to the agency’s judgements? One way of strengthening the Court’s control over the EU agencies would be a further proceduralization of their decision- making, allowing the Court a more useful scrutiny of the external legality of agencies’ decisions given the limited scrutiny of the internal legality (see also section V 6.1.1.5).

As further already hinted at (see section V 6.1.1.3.3.a), the Rütgers case law was also wrongly invoked, as noted by Thomas,281 by the ECHA BoA in the Dow Benelux case before it. In assessing the ECHA’s decision, the BoA noted that ‘[t] he agency’s discretionary powers have also been recognised by the Court of Justice of the European Union’282 to limit its own review of the decision. Of course, the BoA here ignored the functional continuity between the agency and the BoA. Rather than relying on Rütgers to defer to the agency’s decision, the BoA should have relied on that case to exercise an extensive review because, under the Court’s jurisprudence, the BoA itself also forms part of the agency and enjoys broad discretion.

276 Case T- 96/ 10, n 275, para 34.277 See Case T- 262/ 10, Microban v. Commission, [2011] ECR II- 7697.278 Case T- 96/ 10, n 275, para 60. 279 Ibid, para 61.280 Case C- 270/ 12, UK v. Parliament and Council, ECLI:EU:C:2014:18, para 53.281 Thomas, n 215, pp 620– 2.282 ECHA Board of Appeal, Case A- 001- 2012, Dow Benelux, para 110.


Ultimately, this raises the question of how the Board of Appeal’s review relates to that exercised by the judge. Following the principle of functional continuity in the procedure before both the agency and the BoA, a BoA should not defer to assessments made by the agency. However, the BoA itself only has very limited resources compared to the agency, and it will therefore have to start from the latter’s assessment rather than undertaking a fresh one from scratch. Still, this assessment should be deeper than the assessment made by an ordinary judge. The ECHA BoA seems to have described this relation cogently in Honeywell when it dealt with its proportionality review:

[Here] the Board of Appeal underlines the clear differences between itself and the EU Courts. In particular, the latter refrain from substituting their own assessment for that of the EU institution whose decision is being reviewed. However, under Article 93(3) of the REACH Regulation, the Board of Appeal ‘may exercise any power which lies within the competence of the Agency […] Moreover, in conducting its administrative review of Agency decisions, the Board of Appeal possesses certain technical and scientific expertise which allows it to enter further into the technical assessment made by the Agency than would be possible by the EU Courts. As a result, when examining whether a decision adopted by the Agency is proportionate, the Board of Appeal considers that it should not be limited by the need to establish that the decision is ‘manifestly inappropriate’ to the objective pursued.283

As a result, an appellant would have the opportunity to present the BoA (but not the Court) with an argument built on detailed and technical reasoning, which the BoA should then use to scrutinize the decision of the agency. The added value of a BoA would then lie in its capacity to take a deep look into the reasoning of the agency. Of course, the appellant still has to show that the agency has erred by presenting its own alternative reasoning, and a BoA cannot simply use the ‘power which lies within the competence of the agency’ to take a favourable decision for the appellant. The latter still has to make its case. In Heli- Flight the EASA Board of Appeal noted in this regard: ‘The [BoA] cannot see that the elements presented [by the appellant] give it a basis for finding that the Agency is wrong.’284 It should be noted that this is different from the approach taken by the ECHA BoA in Dow Benelux, where it noted ‘that the Appellant has not managed to adequately rebut the fact that its read- across proposal for the endpoint on pre- natal develop-mental toxicity contains a level of uncertainty considered to be unacceptable by the Agency’.285 As noted by Thomas, the fact that the BoA saw no role for itself to ascertain whether the agency’s (preferred) level of uncertainty was justified, and whether the agency was right to find that the alternative proposed by the appellant

283 ECHA Board of Appeal, Case A- 005- 2011, Honeywell Belgium N.V., para 117.284 EASA Board of Appeal, Case AP/ 01/ 2012, Heli- Flight GmbH & Co. KG, para 79. The deci-

sion was appealed before the General Court: see Case T- 102/ 13, n 199.285 ECHA Board of Appeal, Case A- 001- 2012, n 282, para 120 (emphasis added). For a discus-

sion of the contrasting approaches of the ECHA BoA on its own powers, see also Navin- Jones, n 202, pp 158– 65. A similar issue seems apparent in Heli- Flight: see EASA Board of Appeal, Case AP/ 01/ 2012, n 284, para 63.


went beyond that level, made the scientific judgement of the agency virtually unchallengeable.286

Again, the fact that BoAs are composed of technical experts means that, unlike the courts, a BoA does not have to take the scientific judgements of an agency as a given and, if it does, this puts into doubt the added value of an extra level of review by BoAs. Here again a better example of the added value of a BoA may be found in the Honeywell case, in which the BoA found that the ECHA had not exercised discretion where it should have, yet it continued by noting: ‘Whilst the Board of Appeal could annul the Contested Decision on the grounds given above it will nonetheless consider certain of the Appellant’s other claims to assist the Agency in its preparation of a new decision.’287

The preceding shows how the review exercised by BoAs may have a significant added value, provided that the BoAs fulfil their role and do not act as mini-ature courts or as courts of first instance before appeals may be brought to the General Court.

c. Evaluation of the Treaty amendment An evaluation of the Treaty amend-ment is closely linked to an evaluation of the work of the Convention on this point. Although there do not appear to be fundamental problems in practice fol-lowing the entry into force of the Lisbon Treaty in the light of Article 47 of the Charter, it would seem that Baroness Scotland of Ashtal was still right that the amendment of Article 230 EC was not simply an issue of ‘tidying up’.

Obviously, the problematic aspect to Article 263 TFEU is the fifth paragraph, which not only safeguards the system of Boards of Appeal but also leaves a very large margin of discretion to the legislator. The amendment to Article 263(1) TFEU theoretically closes the net and secures that the same level of judicial protection is guaranteed vis- à- vis the agencies as compared to the institutions proper,288 but when the net is effectively closed depends on paragraph 5. Whether anything will be tidied up then depends on whether the legislator, when it exer-cises its powers under Article 263(5) TFEU, is tidy and consistent. As has been noted, this is not completely the case. As a result, even following the Treaty of Lisbon there still remains a need for a framework.

6.1.1.5 The challenge of reviewing non- binding agency actsSince agencies rely more on non- binding instruments than the institutions proper, an important issue that is especially relevant to them is whether their non- binding acts can be challenged before the Union judge. Previously it was noted that in a number of cases against the EFSA,289 the Court had indeed

286 Thomas, n 215, p 620.287 ECHA Board of Appeal, Case A- 005- 2011, n 283, para 100.288 Whether this is a sufficient level of judicial protection is a different matter which will not be

addressed here. In her analysis of AG Jääskinen’s Opinion in Short- selling (cf. Chapter IV, n 576), Skowron noted that the remedies available under German administrative procedural law are much more extensive than those offered under EU law. See Magdalena Skowron, ‘Die Zukunft eu-ropäischer Agenturen auf dem Prüfstand’, (2014) 49 Europarecht 2, p 256.

289 See n 147.


found that the agency’s scientific opinions could not be challenged. Allemano and Mathieu note that this jurisprudence of the Court cannot be upheld, given the increasing importance of the EU agencies (and their non- binding acts).290 In Olivieri,291 the Court still found the EMA’s opinions to be un-challengeable preparatory acts, but in Artegodan it recognized that it should also review an EMA scientific opinion when a (binding) Commission act is being challenged that is based on such an opinion.292 Of course, the Court can only review such a scientific opinion in a limited way, inter alia on its procedural consistency. The usefulness of the Artegodan exception would then be enhanced by a further proceduralization of agencies’ decision- making. As Chiti notes, this proceduralization is generally underdeveloped in the adop-tion of soft law by agencies, but also for some agencies involved in procedures leading to binding (implementing) acts.293 A further proceduralization would then elaborate the standard against which the Court can scrutinize agencies’ acts under the Artegodan exception. Evidently, a high degree of proceduraliza-tion would also allow greater scrutiny by the Court of binding acts, since the Court’s scrutiny is already limited because of the discretion exercised by EU agencies (cf. section IV 5.4.2.2).

As Gundel notes, the problem related to the (reduced) possibility to challenge non- binding acts under Article 263 TFEU is a general issue which is not limited to the EU agencies.294 Indeed, as was noted previously (see section V 6.1.1.4.1.a), the differences between agencies and institutions have largely been evened out under Article 263 TFEU, but this is one of those issues which are relevant to both the agencies and the institutions. At the same time, as Alemanno and Mathieu have rightly noted, the EFSA’s lack of risk management powers (rather than risk assessment powers) follows from a restrictive interpretation of Meroni.295 More generally it has been noted that the legislator has extensively granted soft law powers to EU agencies in order to circumvent the objections which result from a strict interpretation of Meroni.296

290 Alberto Alemanno and Stéphanie Mahieu, ‘The European Food Safety Authority before European Courts’, (2008) 3 European Food and Feed Law Review 5, p 322.

291 Case T- 326/ 99, Olivieri v. Commission and EMEA, [2003] ECR II- 6053, paras 51– 5.292 Case T- 74/ 00, Artegodan v. Commission [2000] ECR II- 2583, para 200. Further, following

AG Darmon’s opinion in Liberal Democrats and the General Court’s ruling in Szomborg, the in-action of an agency to produce an opinion should also be challengeable under Article 265 TFEU if such an opinion is a necessary precondition for another authority to adopt a binding act. See Opinion of AG Darmon in Case C- 41/ 92, The Liberal Democrats v. Parliament, [1993] ECR I- 3153, para 80. See also Opinion of AG Gulmann in Joined Cases C- 15/ 91 & C- 108/ 91, Buckl e.a. v. Commission, [1992] ECR I- 6074, note 20; Case T- 228/ 08, Szomborg v. Commission, [2009] ECR II- 224, para 19. Also confirmed in Case T- 292/ 09, Mugraby v. Council and Commission, [2011] ECR II- 255, para 36.

293 Edoardo Chiti, ‘European Agencies’ Rulemaking: Powers, Procedures and Assessment’, (2013) 19 ELJ 1, pp 100– 2.

294 Jörg Gundel, ‘Der Rechtsschutz gegen Handlungen der EG- Agenturen— endlich geklärt?’, (2009) 44 Europarecht 3, p 391.

295 Alemanno and Mahieu, n 290, p 327.296 Chamon, n 273.


Alternatively, should the Commission’s control function be strengthened vis- à- vis the agencies a more direct procedure before the Commission, to challenge non- binding agency acts, could also be envisaged. A third option would be to broaden the competence of the Boards of Appeal in EU agencies, allowing non- binding or preparatory acts to be challenged.297

At the level of the Court, the interpretation of the criterion that an act should produce legal effects, that is, is capable of affecting the interests of the applicant by bringing about a distinct change in its legal position,298 may also be subject to change. In Pfizer for instance the CFI held an action inadmissible, since the contested decision only required Pfizer to provide information to the EMA.299 However, in a number of more recent cases, the General Court has held that in-formation obligations do constitute legal obligations (cf. section V 6.1.1.4.1.b).300

6.1.1.6 Concluding remarksWrapping up the different subsections above, a number of remarks seem in order. The Lisbon Treaty resolved the lacuna in legal protection under Article 230 EC concerning EU agencies’ acts in the proper way, compared to the unsatisfactory solutions which had been applied pre- Lisbon. However, the theory of the imput-able act has not been affected by Article 263 TFEU and could theoretically still be applied. Article 263 TFEU now also provides the same level of scrutiny vis- à- vis the agencies as compared to the institutions. Still, a large margin is left to the legislator under Article 263(5) TFEU regarding actions brought by natural and legal persons. In addition, even if it is doubtful whether Article 263 TFEU would be the appropriate place, the Treaty amendment does not address the issue of non- binding acts on which agencies rely much more than the institutions.

As regards the techniques used to ensure review of agency measures before the entry into force of the Lisbon Treaty, the lack of uniformity in the functioning of the different Boards of Appeal was noted. The legislator should act to rationalize these differences. As regards the Commission’s review powers, it was noted that the answer to the question of whether this technique should be maintained ultim-ately depends on the definition of the Commission’s role in the EU administration and how agencies relate to the Commission. If the Commission is to develop as the head of the (unitary) EU executive, such review powers are an interesting way of ensuring continued Commission supervision of the EU agencies’ functioning.

6.1.2 Action for damagesDifferent to the procedures under Articles 263 and 265 TFEU, the Lisbon Treaty has not amended the provision on the action for damages to allow

297 Chamon, n 130, p 334.298 Koen Lenaerts, Dirk Arts, and Ignace Maselis, Procedural Law of the European Union,

London, Sweet & Maxwell, 2006, p 211.299 Case T- 123/ 03, Pfizer Ltd. v. Commission, [2004] ECR II- 1631, para 30.300 Case T- 94/ 10, n 275, para 33. See also Case T- 1/ 10, SNF SAS v. ECHA, [2011] ECR II- 6576,

para 42.


injured parties to receive compensation for harm caused by an EU agency’s wrongful conduct.

6.1.2.1 Competence of the Court of Justice in absence of an explicit Treaty provision

The system of non- contractual liability in the Treaties is laid down in Articles 268 iuncto 340 TFEU, which simply refer to the non- contractual liability of the Union and the ECB and the personal liability of Union servants.

At first sight, therefore, the non- contractual liability of EU agencies cannot be established, given that they are not institutions to which Article 340(2) TFEU refers. In Krikorian for instance, the CFI noted that acts of the European Council could not give rise to the non- contractual liability of the European Community since the European Council was not an institution of the EC.301 In other case law the Court has relaxed this requirement in relation to claims for damages against EU bodies that are not institutions in the formal sense under Article 13 TEU. Because the EIB had acted on behalf of the Community and because the EIB had been established by the Treaty, the Court concluded in SGEEM and Etroy v. EIB:

It would be contrary to the intention of the authors of the Treaty if, when it acts through a Community body established by the Treaty and authorized to act in its name and on its behalf, the Community could escape the consequences of the provisions of Article 178 [current Article 268 TFEU] and the second paragraph of Article 215 [current Article 340 TFEU] of the Treaty.302

In Lamberts v. Ombudsman the CFI noted that despite the wide discretion which the Ombudsman has in the exercise of his function, he is always under an obliga-tion to follow up on an appeal by an EU citizen, and that the action for damages only depends on the existence of wrongful conduct. On its jurisdiction to decide upon the non- contractual liability of the Community because of wrongful con-duct of the Ombudsman, the CFI noted, referring to SGEEM and Etroy v. EIB:

The Ombudsman is clearly a body established by the Treaty, which conferred on him the powers set out in Article 195 (1) EC [current Article 228 (1) TFEU]. The right of citizens to have recourse to the Ombudsman is an integral part of citizenship of the Union, as provided for in Part Two of the EC Treaty.303

From these two cases, two requirements may be identified which need to be met by an EU body not mentioned in Article 13 TEU before its non- contractual lia-bility may be established: (i) the body must act on behalf of the Community and (ii) the Treaty authors would have intended Article 340 TFEU to refer also to that

301 Case T- 346/ 03, Krikorian e.a. v. Parliament, Council and Commission, [2003] ECR II- 6037, para 17. On appeal the Court of Justice confirmed the decision ‘sans qu’il [fût] nécessaire d’examiner le rôle du Conseil européen dans l’Union européenne’. See Case C- 18/ 04 P, Krikorian e.a. v. Parliament, Council and Commission, ECLI:EU:C:2004:691, para 30.

302 Case C- 370/ 89, SGEEM and Etroy v. EIB, [1992] ECR I- 6211, para 15.303 Case T- 209/ 00, Lamberts v. Ombudsman, [2002] ECR II- 765, para 50. The appeal to this case

before the Court of Justice was dismissed, see Case C- 134/ 02 P, Ombudsman v. Lamberts, [2004] ECR I- 2803.


body (rather than only those of Article 13 TEU), created by primary law. It should be noted that the Court referred to bodies established by, rather than under, the Treaties. The demarcation line might therefore lie between the secondary and ter-tiary organizational structure of the EU.304 Application of the first requirement would then necessitate an enquiry as to whether the body in question acted on behalf of the EU or on its own behalf.

Although Wathelet and Wildemeersch see this case law as a possible reason why Article 340 TFEU was not amended analogous to Articles 263 and 265 TFEU, with a reference to the bodies, offices, and agencies of the EU,305 it seems difficult to derive a competence of the Court to review the non- contractual liability of EU agencies therefrom. As mentioned previously, most EU agencies should be situ-ated in the tertiary organizational structure of the EU and therefore they do not meet the second requirement. The first requirement is a different matter. The fact that the EIB has a separate legal personality,306 unlike the EU institutions, did not prevent the Court from concluding it had acted on behalf of the EU.307 EU agencies could therefore also act on behalf of the EU. Still the two requirements cannot be cumulatively met by most of the EU agencies, meaning the EU could not incur liability for these EU agencies’ wrongful conduct.

Apart from the previous case law, there is also case law on the non- contractual liability arising from the wrongful conduct of the Brussels Agencies. Similar to the technique adopted by the Court in SNUPAT (cf. section V 6.1.1.2), the Court also found in Worms v. High Authority that, notwithstanding the Office being a body under private law, its acts could ‘be considered as directly giving rise to the liability of the High Authority [sic]’.308 The argument properly construed is that because the High Authority had entrusted a number of tasks to the Office, any wrongful conduct of that body, subject to the condition that the Office acted in a public capacity, could give rise to the liability of the Community under Article 40 ECSC.

An argument based on this case law would indeed be much more persuasive than arguing the Court could adopt a flexible interpretation of the reference to the institutions in Article 340 TFEU which would even include EU agencies. An argument based on Worms simply avoids the question of the position of EU agencies in the EU’s institutional architecture, because in any case the EU agen-cies have not merely been entrusted with tasks by the other EU institutions (just like the Office), but have also been established by the EU institutions. Given the

304 See the distinction made by Hilf, Chapter I, n 64.305 Melchior Wathelet and Jonathan Wildemeersch, Contentieux européen, Bruxelles,

Larcier, 2010, p 262.306 Article 308 TFEU provides the ‘European Investment Bank shall have legal personality’.

This provision has remained unaltered following the different Treaty revisions.307 The AG noted that the Community could be held liable for the conduct of the EIB, subject

to the condition that the actual manifestation of the liability, ie the payment of the compensation, is made out of the EIB’s own funds and not out of the general budget. See Opinion of AG Gulmann in Case C- 370/ 89, SGEEM and Etroy v. EIB, [1992] ECR I- 6211, para 13.

308 Case 18/ 60, Louis Worms v. High Authority, [1962] ECR 195, p 204.


much closer bond between the EU agencies and the EU institutions than that be-tween the Office and the High Authority, extending Worms to EU agencies would be unproblematic— the question to which EU institution the wrongful conduct should then be imputed being less problematic than the question to which EU institution an illegal act of an EU agency should be imputed if the same tech-nique as in SNUPAT is used (discussed previously). After all, Article 340 TFEU provides for the liability of the EU and not of the liability of its individual institu-tions, the Court being quite lenient in declaring admissible those actions which have been brought against an institution which allegedly acted unlawfully instead of against the EU.309

6.1.2.2 Provisions in secondary law conferring competence on the Court of Justice

From the perspective of Worms, it is quite strange that the Community and later the Union legislator have consistently included provisions on the non- contractual liability of EU agencies in their establishing regulations. This is so for every single EU agency established under the former first pillar, the establishing acts of which typically contain a provision mirroring Article 340(2) TFEU. However, this solu-tion does not mirror the solutions in Worms, SGEEM and Etroy v. EIB, or Lamberts v. Ombudsman since those cases involved the liability of the Community, while the provisions in the regulations provide for the liability of the agency itself.

In this sense, the situation of the EU agencies is similar to the situation of the ECB following the entry into force of the Lisbon Treaty. Under the Nice Treaty, Article 288(3) EC provided that the rules on the Community’s non- contractual liability would ‘apply under the same conditions to damage caused by the ECB or by its servants in the performance of their duties’.310

The special reference to the ECB was necessary because before the entry into force of the Lisbon Treaty, the ECB was not a formal institution under Article 7 EC. The current Article 340(3) TFEU however introduces a clearer distinction between the liability of the ECB and that of the Union. The agencies are in a similar situation since their conduct gives rise to their own liability, not that of the Union, and any compensation due is payable from the EU agencies’ own funds. However, although Article 46 of the Statute of the Court of Justice was amended by the Treaty of Lisbon to provide for the distinction between the liability of the ECB and that of the EU,311 the liability of the agencies has not been dealt with similarly.

It was noted previously that the introduction of a provision in the establish-ing regulations on the non- contractual liability of the EU agencies is standard

309 See for instance Case 353/ 88, Briantex Sas and Antonio Di Domenico v. EEC and Commission, [1989] ECR 3623, para 7.

310 See also Andrea Biondi and Martin Farley, The Right to Damages in European Law, Alphen aan den Rijn, Wolters Kluwer Law & Business, 2009, pp 89– 90.

311 While the first paragraph of Article 46, dealing with proceedings against the EU, was not amended in substance, the Lisbon Treaty did add a second paragraph, reading: ‘This Article shall also apply to proceedings against the European Central Bank regarding non- contractual liability.’


practice as regards the former first pillar agencies. The matter is different for the Justice and Home Affairs (JHA) and Common Foreign and Security Policy (CFSP) agencies. The acts setting up the CFSP agencies are completely silent on their non- contractual liability. As regards the JHA agencies, there is no provision at all on the non- contractual liability of the European Police College (CEPOL) in its establishing decision. Following the revision of Eurojust’s mandate in 2008, the regime for non- contractual liability of Eurojust was introduced in Article 27c, which provides Eurojust shall make good any damage caused through fault by the College. However, jurisdiction is given to the national courts, determined by reference to the Brussels I Regulation.312

In the Europol decision the distinction is made between liability arising from the unauthorized or incorrect processing of data and Europol’s other liability. As regards the former, the regime is geared towards the liability of the Member States and not of Europol itself. An injured party should therefore start proceedings against the Member State in which the event which gave rise to the damage oc-curred, the competent judge being determined by that state’s law. If that damage had actually occurred because of the fault of another Member State or Europol itself, these are then bound to repay the first Member State. Should disputes arise on whether and/ or how much of the compensation should be repaid, Europol’s Management Board will settle the matter with a two- thirds majority. As regards the other liability of Europol, the regime is the same as that of Eurojust.313

Given the constitutional background against which these agencies have ori-ginally been established, it is understandable that the regimes put in place for these JHA agencies differ from those of the former first pillar agencies. However, following the entry into force of the Lisbon Treaty, there is no longer any reason why the Court of Justice should not have jurisdiction as regards these agencies’ liability.

6.1.2.3 Non- contractual liability of EU agencies before the Courts and Boards of Appeal

As regards actual proceedings before the Courts in which compensation is claimed for wrongful conduct of EU agencies, there is limited practice so far. These cases generally fall into three different categories: staff cases, cases brought following unsuccessful tenders, and cases concerning alleged wrongful conduct of an EU agency in the exercise of its policy tasks. Only the latter will be commented upon.

6.1.2.3.1 Alleged wrongful conduct in the exercise of policy tasksThe general rules on the action for damages pursuant to Article 340 TFEU will not be discussed in detail here. What is foremost important is that under the classic requirements for liability, that is, a wrongful act or omission, damage

312 Regulation (EC) 44/ 2001 of the Council, OJ 2001 L 12/ 1.313 The original Europol convention referred to the Brussels Convention rather than the Brussels

I Regulation.


incurred, and a causal relationship between these two,314 the analysis of the first requirement makes a distinction between whether the acting authority exercised discretion or whether it was bound: whether a breach of a rule of law intended to confer rights on individuals is sufficiently serious thus depends on the author-ity’s margin of discretion. If the authority had no such margin (or a considerably reduced one), the infringement of a rule of law itself may be sufficient for the EU to incur liability. However, where an authority does exercise discretion, the (al-leged) injured party will have to show that that authority manifestly and gravely disregarded the limits on its discretion.315

To be sure, the issue of discretion is not only relevant for determining the liability of the EU. In an action for annulment, the Union judge will also show reluctance when reviewing the authority’s assessment of the economic facts and circumstances which led to the adoption of the contested act.316 Similarly, under the action for failure to act, the authority must be under a duty to act, that is, not having any discretion on whether to act.317 Whereas discretion may have different meaning in one legal system,318 as illustrated by the pre-vious discussion on German law, and a differentiation was for instance also proposed by AG Van Gerven (cf. section IV 5.5.1.2), the Court itself does not make a clear differentiation in its jurisprudence.319 If the original Meroni doctrine were applied to the EU agencies, one would assume that a manifest error on the part of an EU agency should never be shown. Yet the section on Meroni has shown that the EU Courts recognize that EU agencies may exer-cise discretionary powers, even more so following Short- selling. As a result, should EU agencies become genuine regulators they would find themselves in a similar position as the US regulatory agencies, who are (partially protected) from damage claims under the discretionary function exception laid down by the Federal Tort Claims Act.320

For the EU agencies, and in the current state of agencification, cases on com-pensation for damages in the first place involve EU agencies with tasks relevant to specific private parties, that is, those that are involved in the decision- making on applications filed by private parties, either as an expert body or as the final decision- maker.

314 Lenaerts, Arts, and Maselis, n 298, p 369. 315 Ibid, p 387.316 Ibid, p 307.317 In addition, compelling an authority to act under Article 265 TFEU will not necessarily

result in the adoption of the act sought. Even if an authority is under a duty to act, it is perfectly possible that it has discretion as to the content of the act which it is required to adopt.

318 Alexander Fritzsche, ‘Discretion, Scope of Judicial Review and Institutional Balance in European Law’, (2010) 47 CMLRev 2, p 362.

319 Ibid, p 364.320 See 28 United States Code § 2680 (a). Further see Jay Lapat and James Notter, ‘Inspecting

the Mine Inspector: Why the Discretionary Function Exception Does Not Bar Government Liability for Negligent Mine Inspections’, (2006) 23 Hofstra Labor & Employment Law Journal 2, pp 418– 20.To illustrate, the discretionary function test was applied to the air safety regulator and to the FDA in cases such as Varig Airlines and Berkovitz. See US Supreme Court, United States v. Varig Airlines, [1984] 467 United States Reports 797; US Supreme Court, Berkovitz v. United States, [1988] 486 United States Reports 531.


In the three cases against the EFSA related to the inclusion of certain sub-stances in Annex 1 of Directive 91/ 414,321 the applicants also claimed damages resulting from the loss of profit because their clients would no longer buy the products in question and because of the damage suffered to their reputations. The EFSA pleaded the inadmissibility of the claims for damages because the measure contested by the applicants did not have legal effects. Unfortunately, the Court did not reach this point in its reasoning, since it found the claims to be inadmissible as the applicants had not quantified or assessed the damage al-legedly suffered.

Assuming (qualitative) agencification will continue, following Short- selling, more cases involving agencies will come before the Court. One issue which would merit further attention is the question of comparative responsibility involving an EU agency. An agency such as the EASA, for instance, certifies aeronautical products, while the ERA, assuming the fourth railway package is adopted, re-ceived the power to authorize vehicles being put on the network. In the event of an aeroplane or train accident in the future, and making abstraction of the issue of discretion, the responsibility of these agencies might be raised in cases before national judges which are asked to rule on the liability of the different actors involved: infrastructure manager, pilot/ driver, airline/ railroad company, regula-tors, etc. In these cases of joint liability, the Court requires that the liability of the national authorities be determined before deciding on the EU authority’s responsibility.322 For the moment, the provisions in secondary legislation pro-vide for an exclusive competence of the Court but, given the silence in primary law as regards the agencies’ liability, this could also change in the future.323 In any event, if EU agencies are conceived as forming part of a multi- layered (agencified) administration (cf. section III 3.1), the possibility of holding regu-latory authorities liable might require harmonization on this issue between the Member States and the EU. Van Praag for instance has noted that EU legislative harmonization in the financial sector has reached such a level that the different regimes for non- contractual liability of financial supervisory authorities ought to be aligned.324

321 See cases at n 147.322 Lenaerts, Arts, and Maselis, n 298, p 384.323 In such a case, the agencies’ immunity under the Protocol on Privileges and Immunities

would need to be addressed as well. See for instance Article 49 of Regulation 1901/ 2006, which pro-vides: ‘Without prejudice to the Protocol on the Privileges and Immunities of the European Communities, each Member State shall determine the penalties to be applied for infringement of the provisions of this Regulation or the implementing measures adopted pursuant to it in relation to medicinal prod-ucts authorised through the procedures laid down in Directive 2001/ 83/ EC’ (emphasis added), thus implying that the Member States cannot impose penalties should the EMA (or the Commission) infringe EU law.

324 Emanuel van Praag, ‘Aansprakelijkheid van financiële toezichthouders naar Europees recht’, (2014) 62 SEW 5, p 217. Some EU Member States for instance provide for an immunity for their supervisors while others apply the normal liability rules. See Robert Dijkstra, ‘Liability of Financial Regulators: Defensive Conduct or Careful Supervision?’, (2009) 10 Journal of Banking Regulation 4, p 269. Evidently, EU legislation does not provide for immunity for the ESAs or the EU agencies in general.


6.1.2.3.2 Action for damages and proceedings before the Boards of AppealThe action for damages under Articles 268 iuncto 340 TFEU is an autonomous procedure which does not require a prior ruling finding the illegality of the act giving rise to damages. In its early case law, however, the Court did make the action for damages dependent upon an earlier finding of illegality.325 In a way, following the Lisbon Treaty, the system of Boards of Appeal might have reintro-duced this requirement. Here it need only be recalled that before proceedings may be brought under Article 263 TFEU against EU agencies, procedures before the Board of Appeal (if they are provided for) must be exhausted. So far, how-ever, the Boards of Appeals’ jurisdiction is limited to hearing annulment cases. Since compensation for damages are (exclusively) granted by the Court, it could occur that a party applies for compensation before the Court for an (alleged wrongful) act which should first be challenged before the Board of Appeal. This could of course undermine the system of Boards of Appeal and the requirement of Article 263(5) TFEU. So far this has not led to serious problems, even if the ESA Board of Appeal has noted in a case that it ‘does not however have power to order the payment of monetary compensation to an appellant’.326 Whether this issue should be addressed is a different matter. After all, proceedings before the Boards of Appeal are dealt with much more expediently than cases before the General Court. A party could therefore apply for damages before the Court and simultaneously request the annulment or reversal of the decision before the Board of Appeal. Of course, the possibility to broaden the latter’s jurisdiction could also be explored.

6.1.2.4 Concluding remarksFollowing the Treaty of Lisbon, the scope of Articles 268 iuncto 340 TFEU was not extended to the EU agencies. Applying the Court’s Worms jurisprudence to EU agencies then seems more fruitful than applying SGEEM and Etroy v. EIB and Lamberts v. Ombudsman, to bring the agencies under the regime of non- contractual liability as provided for in primary law.

The legislator has however opted for a different regime by including specific provisions on the liability of the agencies (rather than the EU) in the establishing regulations, but this only for the former first pillar agencies. Following the Treaty of Lisbon, this practice should be reviewed.

Today, the jurisprudence on the agencies’ non- contractual liability other than that resulting from staff cases or tendering procedures gone wrong remains lim-ited. Assuming (qualitative) agencification will inevitably push forward, it is to be expected that this will change. For the moment, the Boards of Appeal established in some agencies have no jurisdiction to award damages but, given the silence in primary law, the legislator might also bring this remedy under the competence of a Board of Appeal.

325 Case 25/ 62, Plaumann & Co. v. Commission, [1963] ECR 95, p 108.326 ESAs Joint Board of Appeal, Case 2013- 14, Global Private Rating Company ‘Standard Rating'

Ltd v. ESMA, para 183.


6.2 Active locus standi of EU agencies

While the Treaty of Lisbon at first sight has mended the lacuna in Articles 230 (and 232 EC), the Les Verts, Comitology, and Chernobyl cases as regards the stand-ing of the Parliament show there are two sides to the coins of Articles 263 and 265 TFEU, that is, the questions of both passive and active locus standi. In Comitology, the Court rejected the Parliament’s argument on a parallelism between the active and passive participation in judicial proceedings,327 relying instead on the institu-tional balance in Chernobyl to grant the Parliament standing.

Following Les Verts and Chernobyl, the Maastricht Treaty amended the then Article 173 EEC, allowing acts of the Parliament intended to produce legal ef-fects vis- à- vis third parties to be challenged and at the same time allowing the Parliament (and also the ECB) to bring proceedings to protect its prerogatives.328 The Amsterdam Treaty further allowed the Court of Auditors to bring proceedings to safeguard its prerogatives. After the Treaty of Nice upgraded the Parliament to a (fully) privileged party under Article 230 EC, the Treaty of Lisbon introduced a new semi- privileged party, allowing the Committee of the Regions to bring pro-ceedings to safeguard its prerogatives.

It should therefore be noted that three semi- privileged parties under Article 263 TFEU are expressly granted active locus standi, with only the ECB being identified in name as a possible defendant in proceedings under that Article.

Of course, the reason for this is that the other two parties are not tasked under the Treaties to adopt binding acts. In the Ismeri case, however, the issue arose as to whether proceedings under Article 173 EEC could be brought against the Court of Auditors. Because the latter itself did not raise a plea of inadmissibility on this ground, it was ignored at first instance and on appeal, but Advocate General Colomer did dedicate part of his analysis to this question, concluding that following Les Verts, ‘the activities of all institutions [are] subject to review by the Court of Justice in so far as they may produce legal effects vis- à- vis third parties’.329 In the AG’s reasoning, Article 13 TEU played an important role. Since the Court of Auditors had been included in the list of institutions of Article 4 EEC, the AG found that an action for annulment of an act of the Court of Auditors could not be ruled out simply because Article 173 EEC did not refer to that institution. Since the Committee of the Regions is mentioned by Article 13 TEU, without being qualified as an institution, the AG’s reasoning would also extend to this body.

The reverse analogy with the situation of the EU agencies is clear: Articles 263 and 265 TFEU expressly provide for their passive locus standi, but are silent on the

327 Case 302/ 87, Parliament v. Council, [1988] ECR 5615, para 21.328 The Maastricht Treaty further allowed proceedings to be brought against the ECB and also

allowed the ECB to bring proceedings to safeguard its prerogatives.329 Opinion of AG Ruiz- Jarabo Colomer in Case C- 315/ 99 P, Ismeri Europa Srl v. Court of

Auditors, [2001] ECR I- 5281, para 44. Before AG Colomer, other AGs had concluded in favour of allowing an action for annulment of an act of the Court of Auditors: see the cases cited by AG Colomer in footnote 32 to his opinion in Ismeri.


active locus standi. Was this a conscious omission by the Treaty authors? From the working documents of the European Convention this does not appear to be the case. As regards judicial proceedings, the Discussion Circle on the Court of Justice foremost focused on the standing of private parties under the strict Plaumann test and on the possibility to challenge acts of agencies. Only one amendment was tabled which would have granted active locus standi to agencies, although this consequence was not the amendment’s principal aim. Then MEP Kaufmann inter alia suggested that paragraphs one and three of Article 263 TFEU should be sim-plified as follows:

(1) The Court of Justice of the European Union shall review the legality of the acts of the institutions, as well as those of the bodies, offices or agencies of the Union intended to produce legal effects vis- à- vis third parties.330

[(2) It shall for this purpose have jurisdiction in actions brought by a Member State, the European Parliament, the Council or the Commission on grounds of lack of competence, infringement of an essential procedural requirement, infringement of the Treaties or of any rule of law relating to their applica-tion, or misuse of powers.]

(3) The Court shall have jurisdiction under the same conditions in actions brought by the other Institutions as well as the bodies, offices or agencies of the Union for the purpose of protecting their prerogatives.331

MEP Kaufmann presented this amendment as a mere simplification exercise, but its effects, had it been accepted, would evidently have been greater. The amend-ment is interesting because it does raise the question of whether agencies should be granted active locus standi to safeguard their prerogatives, and this under both Articles 263 and 265 TFEU.

This question cannot be simply answered in the affirmative, since the Court, in Comitology, rejected the Parliament’s argument on a parallelism between active and passive locus standi under Article 263 TFEU. The argument which it used in-stead, in Chernobyl, built on the institutional balance. So did AG Colomer in his opinion in Ismeri, albeit indirectly, as he referred to the current Article 13 TEU which enshrines the institutional balance (see section IV 7.2.4). Indeed, under both lines of reasoning the answer would be negative. Agencies could not dir-ectly invoke the institutional balance analogously to Chernobyl since they do not come under the institutional balance.332 This could only be different for those EU agencies which are part of the secondary organizational layer (Europol, Eurojust, EDA) if the Court could identify some prerogatives. Similarly, an argument analogous to that of AG Colomer would fail because the agencies are not men-tioned in Article 13 TEU and such a defect could not be covered by reference to

330 See Secretariat of the European Convention, Suggestion for Amendment of Article: III- 266 by Ms. Dr. Sylvia- Yvonne Kaufmann, http:// european- convention.eu.int/ docs/ Treaty/ pdf/ 882/ Art III 266 Kaufmann DE.pdf. Translated from German by the author.

331 Ibid. Translated from German by the author.332 See the rejection of Schniga’s extension of the institutional balance to the third organizational

layer of the EU (cf. section IV 7.2.3.1).

http://european-convention.eu.int/docs/Treaty/pdf/882/Art%20III%20266%20Kaufmann%20DE.pdf





cases such as SGEEM v. EIB and Lamberts v. Ombudsman (cf. section V 6.1.2.1), as agencies are not ‘established’ by the Treaties either.

The amendment proposed by MEP Kaufmann further shows the difficulty of granting agencies active locus standi. As agencies would only be semi- privileged parties, they would only be able to bring proceedings ‘to safeguard their preroga-tives’. But what are the prerogatives of bodies which are only established by sec-ondary legislation and for which the Treaties do not foresee a special function or role?

It may further be questioned whether it would be expedient, from a political perspective, to upgrade the standing of EU agencies. As AG Colomer noted in his opinion in Ismeri, Article 263 TFEU has two functions. One is to safeguard the rights of applicants and the other to monitor the institutions’ compliance with EU law.333 Granting active locus standi to agencies would then fall under the latter function. One could indeed argue that it would be useful to allow the experts to intervene when the institutions adopt decisions which are not in line with scien-tific or technical advice. On the other hand, it cannot be doubted that the institu-tions’ legitimacy is greater than that of the agencies and that it is pursuant to that political legitimacy that the institutions may make political choices which deviate from the scientific and technical advice.334 If agencies work in support of the EU institutions, an agency should not be able to judicially contest the political choices which the EU institutions make when elaborating EU policy.

Seen from this perspective, the fact that no agency has claimed the right to bring proceedings so far shows that agencies are sticking to their role and are co- operating with the Commission and other institutions rather than confronting them.

Since there is no necessary link between the actions under Articles 263 and 265 TFEU,335 one may wonder whether the same comments apply to the action for failure to act. Curiously, in her study on this procedure, Cazet lists the agencies among the requérants privilégiés, noting that ‘[l] e Conseil européen, la Banque central européenne, les organes et organismes de l’Union ont été rajoutés [à cette catégorie] à la faveur du traité de Lisbonne’.336 However, Article 265 TFEU only provides that ‘[t]his Article shall apply, under the same conditions, to bodies, of-fices and agencies of the Union which fail to act’ (emphasis added). This would suggest that the action extends to the agencies only if and in so far as they have failed to act, but would not extend to agencies as parties able to bring proceedings.

Would it then be advisable to insert a provision in Article 265 TFEU, similar to the one introduced by the Treaty of Maastricht for the ECB, to the effect that the Court of Justice would have jurisdiction in actions or proceedings brought by the bodies, offices and agencies of the Union ‘in the areas falling within their field

333 Opinion of AG Ruiz- Jarabo Colomer in Case C- 315/ 99 P, n 329, para 37.334 This in itself should also be taken into account by the Court in the event that agencies would

be able to bring proceedings against the institutions.335 Lenaerts, Arts, and Maselis, n 298, p 332.336 Safia Cazet, Le recours en carence en droit de l’Union européenne, Bruxelles, Bruylant,

2012, p 124.


of competence’? This reference to fields of competence immediately draws the at-tention to a major difference between the EU agencies and the ECB. Unlike the latter, the EU agencies have no competences under the Treaties, their field of com-petence being defined by secondary legislation. Again this problem boils down to the fact that agencies do not come under the institutional balance and hence have no proper field of competence or prerogatives.

In their fields of competence, however, the agencies are often involved in com-posite proceedings involving cooperation between different actors, and this is also the case at EU level. As a result it is not inconceivable that another EU body or an EU institution’s failure to act would mean that the agency cannot fulfil its role in the procedure. At first sight, in such situations an action for failure to act brought by the agency could prove useful. However, it would seem doubtful here whether the agency would have an actual interest in bringing proceedings. After all, the agency itself is a tool rather than an actor with an independent interest. If a procedure remained in limbo because of the inaction of a body or institution, the interest of a private party or (another) institution would have that party or in-stitution bring proceedings, but not the agency.

This would only be different in a limited number of cases where the inaction of an institution or body threatens to paralyse the functioning of an agency so that it could not fulfil its mandate anymore or where an agency has been established as the primary authority within its field of competence, similar to the ECB’s role in monetary policy. As was noted, the latter seems to be happening gradually in the financial sector, where the three ESAs have been established with significant powers.

Among these is the power to adopt binding decisions vis- à- vis national author-ities. Yet before an ESA can do so, the institutions are usually required to act.337 It is up to the Council for instance to conclude that there is an emergency situation, and it is up to the Commission to conclude that a national authority has acted in breach of EU law, following which the ESA can act. In both cases the relevant ESA may invite the Council or Commission to take those decisions, raising the question of what would happen if these institutions do not act. Should the ESA, with its specific and important task in regulating the financial markets, not have the power to have such inaction scrutinized by the Court? After all, recitals 11 to the preambles of the ESAs’ establishing regulations provide that ‘[t] he Authority should protect public values such as the integrity and stability of the financial system, the transparency of markets and financial products and the protection of investors’. Should an ESA then not have standing when it believes the inaction of an institution338 threatens the ‘integrity and stability of the financial system’, given that the ESA is under an obligation to protect this integrity and stability?

337 It may be recalled here that the ESMA’s contested power in Short- selling is an ‘emergency intervention’ power which the agency exercises out of its own motion.

338 For the sake of the argument, abstraction is made of the likely event that the Court would rule that the Commission’s or Council’s decision to act lies within their discretion and hence no obligation to act could be found.


Ultimately this is again a political question and depends on where we see the EU agencies within the EU’s institutional architecture. Are they mere support-ing bodies for the institutions? Or should they be qualified as the primary actors (besides the legislator) at EU level in their specific policy field? Since the latter option would redefine the EU’s institutional architecture, a recognition of this in primary law would appear absolutely essential.

Given the wording of Article 340 TFEU and the separate legal personality of the EU agencies, there do not seem to be significant legal hurdles in an agency claiming damages from the EU or another EU agency. Obviously, whether this would be sensible or even realistic from a practical perspective is a different matter.

Lastly, to illustrate that the active locus standi of agencies (as semi- privileged parties) under Articles 263 and 265 TFEU would be far from straightforward, the rules on interventions by EU agencies may be recalled. In Rubinum a Commission decision to suspend the authorization for a feed additive, following a negative opinion from the EFSA, was contested. The EFSA had therefore requested per-mission to intervene in the procedure, but the General Court referred to Article 40 of the Court’s Statute which puts agencies on par with natural and legal per-sons when it comes to the possibility to intervene.339 The General Court found that the EFSA should still argue its interest, even if it had delivered an opinion in the administrative procedure and even if the issue wholly fell within its mandate as defined by the EFSA Regulation, since ‘cette circonstance ne le dispense pas d’établir son intérêt à la solution du litige’,340 further explicitly rejecting any ana-logy with the European Data Protection Supervisor341 since the EFSA Regulation ‘n’attribue pas à l’Autorité européenne de sécurité des aliments un tel droit’.342 Still, it is significant that the General Court rejected such a right because sec-ondary law did not provide for it, not because the EFSA is not provided for under primary law.

6.2.1 Concluding remarksMany valid practical reasons for granting EU agencies active locus standi under Articles 263 and 265 TFEU may be imagined. One could also argue that even if granting active locus standi would not bring much benefits, it would not bring harm in practice either. Yet even if this may be true, the underlying issue is one of principle. The capacity to bring actions under Articles 263 and 265 TFEU is not defined at random but rather in a deliberate manner.

As noted by AG Colomer, the system of judicial protection aims (i) to safe-guard the rights of private parties and (ii) to monitor the institutions’ com-pliance. This evidently requires a broad category of bodies which are granted

339 Still, the EU agencies, unlike ordinary private parties, are not barred from intervening in cases between institutions and/ or Member States.

340 Order of the President in Case T- 201/ 13, Rubinum v. Commission, ECLI:EU:T:2013:562, para 11.

341 Order of the Court in Case C- 317/ 04, Parliament v. Council, [2005] ECR I- 2457, paras 14– 19.342 Case T- 201/ 13, Rubinum v. Commission, ECLI:EU:T:2013:562, para 12.


passive locus standi, but this should not necessarily be effectuated by allowing all the bodies or institutions of the EU (or any private party through an actio popu-laris) to bring proceedings. The category of bodies and institutions to which this right has been granted has developed gradually, in line with the development of the institutional architecture and the institutions’ competences as defined by the Treaties.

As long as there is no clear basis in the Treaties for agencification, then the EU agencies could not claim active locus standi under Articles 263 and 265 TFEU as (semi- )privileged parties.

6.3 Conclusion

The Lisbon Treaty adequately addressed the issue of passive locus standi of agen-cies. While Les Verts was also extended to EU agencies in Sogelma, the solution offered by the Lisbon Treaty remains preferable for evident reasons. As a result, recourse to other solutions which had been worked out to secure judicial scru-tiny of agencies’ acts should be discontinued in favour of relying on Article 263 TFEU. The only exceptions which could be made here are the possibility for the Commission to exercise an administrative review of agencies’ acts, depending on how the Commission’s role is conceptualized in the EU executive. Secondly, since Article 263(5) TFEU refers back to the specific procedures provided in the establishing acts of the agencies, this solution evidently may still be employed as well.

The system of Boards of Appeal acts as an important filter before access is granted to the General Court. In addition, it allows affected parties to challenge decisions on technical grounds which could not be invoked before an ordinary judge. Yet, while the Boards of Appeal of all agencies fulfil the same function, a large degree of heterogeneity characterizes the provisions on their organization and functioning. Furthermore, the Boards of Appeal themselves are still looking for their proper place in between the agency’s administration and the Court of Justice.

As regards the judicial scrutiny of agencies’ acts, the General Court has em-phasized its limited review in light of the discretion exercised by the agencies. This strand of case law sits uneasily with Meroni, especially following the Court’s renewed confirmation, in Short- selling, that agencies cannot exercise discretionary powers. Still, Short- selling will not stop the General Court from recognizing and sanctioning the large degree of discretion exercised by EU agencies.

Given the extended scope of Articles 263 and 265 TFEU, it is remarkable that Articles 268 iuncto 340 TFEU have not been amended to include the EU agen-cies. From a theoretical perspective, imputing the agencies’ wrongful conduct to one of the institutions, analogous to Worms, is a more sound solution than trying to bring the agencies themselves within the scope of Article 340 TFEU, analo-gous to SGEEM and Etroy v. EIB and Lamberts v. Ombudsman. In practice, the legislator has chosen to deal with the matter explicitly by granting jurisdiction, in secondary law, to the Court of Justice. The fact that this matter is not governed


by provisions of primary law would also allow the legislator to bring this remedy under the jurisdiction of an agency’s Board of Appeal.

The amendments resulting in the passive locus standi of EU agencies under Articles 263 and 265 TFEU stand in contrast to the lack of active locus standi for EU agencies post- Lisbon. Since the Court has rejected a parallelism between the two, the capacity to bring proceedings cannot be inferred from the agencies’ con-firmed passive locus standi. Remarkably, this matter was seemingly ignored by the Convention leading to the Constitution for Europe. Ultimately the question of whether EU agencies should be entitled to bring proceedings should follow from a decision on the agencies’ place within the EU’s institutional architecture, which is a question that can only be resolved by a new Convention and IGC.

VIConclusions and Proposal

1 The EU Agency Instrument— Conclusions

The main conclusions to be drawn from the preceding chapters are that: (i) agen-cification is an atypical form of administrative integration and administrative cap-acity building; (ii) agencification is a political solution, favourable to the Member States, to functional pressures; (iii) while agencification has largely been tailored to the needs of the Member States, the agency instrument is seen as legitimate by all actors involved; (iv) agencification as an atypical phenomenon has developed in absence of a framework, resulting in an unjustifiable diversity in the rules on agencies’ functioning; (v) agencification (as evolved up until today) is question-able under primary law, but the Court has not stepped in and instead has resigned itself to institutional practice; (vi) the Treaty of Lisbon has mended the most sig-nificant lacunae in EU procedural law related to EU agencies, but the agencies’ position in that procedural law can only be definitely settled once the question on agencification’s finality is settled; (vii) since important legal limits are not en-forced, the onus is even more on the political institutions to ensure the legitimacy of agencification.

Agencification is an atypical form of administrative integration and adminis-trative capacity- building. The Treaties lay down that EU law is, as a rule, imple-mented by the Member States. When EU actors are involved, the Treaties prescribe that the actors responsible are the Commission and, exceptionally, the Council. Agencies do not fit nicely in this dichotomy between direct and indirect admin-istration, but from a functional and procedural perspective (cf. section I 2.2) it is clear that they are part of a trend towards greater uniformity in the European Administrative Space, resulting in administrative integration. Agencification is then also an atypical form of administrative capacity- building, since expertise at EU level was usually housed within the Commission or drawn from third parties under the auspices of the Commission or Council (cf. section III 1).

Agencification is a political solution, favourable to the Member States, to functional pressures. Following the SEA and Europe 1992, indirect administra-tion has become increasingly insufficient to deliver the EU’s policies. Granting more powers to the Commission could have addressed this problem but it would negatively affect the Member States’ administrative autonomy. Agencies are in- between solutions allowing for greater coordination between the Member States, ensuring better application of EU law without giving up as much autonomy as

Conclusions and Proposal370

would be the case were the Commission to take on implementing tasks. In par-allel, agencies allow the EU to pool expertise without expanding the Commission.

Agencification has largely been tailored to the needs of the Member States but the agency instrument is seen as legitimate by all actors involved. If the Commission had taken a principled stance, agencification would have halted fol-lowing the first wave. However, the Commission has allowed a fragmentation of the EU’s executive and an undermining of its own position as part of a trade- off with greater and deeper European integration. This pragmatism was also taken over by the Parliament when it became co- legislator in agencification. This con-sensus among the institutions has also meant that the first real challenges to agencification, in ENISA and Short- selling, before the Court were brought by a Member State that was outvoted in Council.

Agencification as an atypical phenomenon has developed in absence of a frame-work, resulting in an unjustifiable diversity in the rules on agencies’ functioning. In agencification, the institutions have acted based on their own political interests and pragmatism much more than on considerations related to good governance or a concern about the effects of agencification on EU governance in general. While the negative effects of the lack of such a framework were already apparent in the early 1990s, agencification has continued in an ad hoc fashion. The at-tempt to come to a framework faltered, resulting in the non- binding Common Approach— which is not even a codification of existing practices, let alone a re-flection of a genuine common understanding on agencification shared by the three institutions. A framework could do away with irrational differences between agencies’ statutes and give expression to a common understanding on the purpose of EU agencies, bolstering the EU’s legitimacy.

Agencification (as evolved up until today) is questionable under primary law, but the Court has not stepped in and instead has resigned itself to institutional practice. While agencification does not appear problematic in the light of the principle of conferred powers, and may even be seen as an expression of the sub-sidiarity principle, allowing Member States to remain involved in the executive function, the assessment is less positive for some individual agencies. Here a ten-sion may also be noted between the principles of subsidiarity and proportionality. Under subsidiarity, establishing agencies with modest powers is questionable. The same is also true under a proportionality perspective should a heavy organizational form such as an agency be chosen to exercise modest powers. At the same time, such modest powers in themselves are least likely to be found disproportionate.

However, the greatest legal obstacle to agencification may be found in the idea of institutional balance, whereby the problem is not so much that agencies would violate that balance, but that the legislator does so when it grants powers which the Treaties assign to the Commission. In second order there are the limits which flow from the Court’s jurisprudence in Meroni and Romano. In Short- selling the Court dealt with these issues unconvincingly, since it simply ignored the insti-tutional balance and incorrectly noted that Romano did not add anything to Meroni. In a questionable way it reduced the latter to a single requirement and narrowed down the notion of discretionary powers, resulting in a Meroni- light,

The EU Agency Instrument—Conclusions 371

disregarding the notion of control. Although this sanctioning of agencification by the Court could be interpreted as obviating the need for a Treaty legal basis, it in fact underscores the necessity thereof. After all, the gap between primary law and the reality of agencification has again widened.

Agencification initially resulted in important lacunae in the system of legal protection as laid down by the Treaties, since the latter assumed a closed number of bodies competent to adopt acts having legal effects. While parallels could be drawn between the agencies and bodies such as the Parliament, European Investment Bank (EIB), European Central Bank (ECB), and Ombudsman, a fun-damental obstacle to such analogies remained the fact that agencies were and are not provided for under primary law. In the absence of a framework, different tech-niques have been used to secure adequate judicial protection, which should now, post- Lisbon, be re- evaluated. The lack of a framework has also meant that the technique of the internal Board of Appeal has been implemented differently for different agencies. Here again the Common Approach was found to be a missed opportunity to contemplate more profoundly the organization, functioning, and jurisdiction of these bodies. While the Lisbon Treaty addressed the most im-portant lacunae, a number of issues remain unresolved because of the unresolved primary law status of EU agencies. The precise place of the EU agencies under the EU’s procedural law was therefore found to depend on the finality of EU agencification.

Agencification requires legitimacy, which depends even more on the political institutions given the questionability of the phenomenon from a legal perspec-tive. Since the sanctioning of agencification by the Court in Short- selling has not resulted in a greater legal legitimacy of the phenomenon, the onus falls on the political institutions, in the absence of a Treaty legal basis, to secure the legit-imacy of the phenomenon. This could be achieved by adopting a clear framework, which would contribute to controlling agencification and agencies, respecting the prerogatives of all the institutions involved. Control indeed appears crucial, even more since EU agencification is intertwined with national agencification in cer-tain sectors (cf. section III 3.1). However, comparative excursions to other pol-ities show that EU agencification and agencies are rather under control. While advances may and should still be made, the excursions also show how pursuing a completely satisfactory solution would be in vain. Still, any serious solution to this challenge only seems possible if a consensus may be reached on the purpose of agencies and their place in the EU’s institutional framework. This question in turn depends on even more fundamental questions related to the Commission’s pos-ition in the EU executive and the democratic finality of the EU. These questions partially transcend the authority of the institutions, including the Court, which is why a Treaty legal basis for agencification appears necessary and appropriate.

In the spirit of these conclusions, the following section will take a look at the proposals made in legal doctrine and during the IGCs to include a legal basis in the Treaties for agencies. In the light of these proposals and drawing insights from the previous chapters of this study, a legal basis for agencies in the Lisbon Treaty is then proposed.


2 The EU Agency Instrument— Proposal

A pragmatist might inquire about the purpose of the present section, given that (i) there is a broad consensus among the political institutions on agencification, (ii) the Court has refrained from intervening, and (iii) the process of agencifica-tion is more or less controlled.

In line with the conclusions in the preceding section, the answer to this ques-tion follows from the legitimacy of the EU and its functioning, which, as far as the EU agencies go, are undermined by the absence of a clear legal basis. This in turn has resulted in a lack of clearly enforceable limits to the process and the lack of a genuine common approach. The Court’s ruling in Short- selling was a symptom of this and led Van Gestel to note that what the case ‘has made painfully clear is that we are lacking a solid normative framework in the Treaties’.1 Following two decades of agencification, the experimental phase of the phenomenon should also be behind us and a consolidation phase should be commenced. As was found, the Common Approach might be a first timid attempt thereto, but it is far from sufficient.

A legal basis in the Treaties would then in itself firmly anchor agencification in pri-mary law and, depending on how the provision is drafted, may result in both clearer limits and a horizontal instrument governing agencification. The following sections should therefore be situated in the academic debate on a possible anchoring in pri-mary law of agencification. Similar to the position taken here, in 2000 Vos found that ‘[f ] or reasons of legal certainty and coherence, a legal basis for the creation of agencies and the delegation of authority should […] be introduced in the Treaty’.2

While suggestions for a legal basis had been made earlier,3 the academic debate has really picked up since the second wave of agencification. In 1993 Lenaerts noted that a change to the current Article 13 TEU might be required.4 Majone elaborated this idea by suggesting the possibility of a ‘minor treaty revision’ adding ‘a sentence empowering the Council and the European Parliament to establish new bodies that may be needed for the efficient functioning of the internal market’5 to Article 13 TEU. Geradin proposed two other solutions: the first was an amend-ment to Article 13 TEU which would list the agencies ‘next to the Council, the Commission, the EP, the ECJ and the Court of Auditors in its list of EU

1 Rob van Gestel, ‘European Regulatory Agencies Adrift? Case C- 270/ 12’, (2014) 21 MJECL 1, p 195.

2 Ellen Vos, ‘Reforming the European Commission: What Role to Play for EU Agencies?’, (2000) 37 CMLRev 5, p 1124. Similarly, see Ronald van Ooik, ‘The Growing Importance of Agencies in the EU: Shifting Governance and the Institutional Balance’, in Curtin and Wessel (eds), Good Governance and the European Union: Reflections on Concepts, Institutions and Substance, Antwerpen, Intersentia, 2005, p 132.

3 See for instance Herman Maas, ‘Delegatie van bevoegdheden in de Europese Gemeenschappen’, (1967) 15 SEW 1, pp 17– 8; see section IV 5.2.

4 Koen Lenaerts, ‘Regulating the Regulatory Process: “Delegation of Powers” in the European Community’, (1993) 18 ELRev 1, p 42.

5 Giandomenico Majone, ‘The Credibility Crisis of Community Regulation’, (2000) 38 JCMS 2, p 276.

The EU Agency Instrument—Proposal 373

bodies’;6 the second was ‘the insertion of a specific legal basis in the Treaty for the creation of agencies, as well as a reference to the procedure to be followed for their creation’.7 According to Geradin, the second option would be the easiest to secure, while the first option would be preferable since it would (allegedly) integrate the agencies in the institutional balance, allowing agencies to exercise discretionary powers. At the same time it would also, dubiously, place the agencies at the same level as the institutions.8 Rodrigues also explored Geradin’s second option, that is, the addition of a new legal basis allowing the legislator to set up agencies, and con-sidered an amendment of the then Article 211 EC (currently Article 291 TFEU), explicitly allowing the Commission to subdelegate its powers to an agency.9

The most concrete proposal was made by Helfritz, who suggested including an Article 267a EC. The Article would have been rather detailed, consisting of eleven paragraphs. Most importantly, it prescribed (i) that agencies be established pursuant to the codecision procedure; (ii) that the Commission would have the competence to overrule agencies’ decisions; (iii) that the decision to establish Boards of Appeal would be left to the agencies themselves; (iv) a fixed composition for the Board, ap-pointed solely by the Council; (v) (in general terms) which kinds of powers could be granted to agencies; (vi) the rules regarding the agencies’ staff and budget; and (vii) that agencies may only be established and abolished by a regulation.10

Recently, Scholten proposed a shorter Article made up of four paragraphs. Her proposal does not make clear where in the Treaties the Article would be inserted, but it prescribes (i) that agencies would be established pursuant to the ordinary legislative procedure, clearly indicating the reasons for the necessity of the agency; (ii) that only executive powers ‘involving no policy- making discretion’ could be delegated and that agencies be accountable to the Parliament and Council; (iii) a horizontal framework for agencies be established pursuant to the ordinary legisla-tive procedure; and (iv) that the national parliaments be informed of the elabor-ation of this framework.11

2.1 Proposals in the IGCs and the Convention

Calls to provide a legal basis were first made for the Amsterdam Treaty when agencification really started to gain pace in the mid- 1990s.12 However, it was only

6 Damien Geradin, ‘The Development of European Regulatory Agencies: What the EU Should Learn from the American Experience’, (2004) 11 CJEUL 1, p 15.

7 Ibid. 8 Ibid.9 Stéphane Rodrigues, ‘Quelques considérations juridiques en faveur d’un statut pérenne des

autorités européennes de régulation’, (2004) 60 Actualité Juridique Droit Administratif 22, pp 1181– 2.10 Vark Helfritz, Verselbständigte Verwaltungseinheiten der Europäischen Union, Berlin,

Weissensee Verlag, 2000, pp 205– 7.11 Miroslava Scholten, The Political Accountability of EU Agencies: Learning from the US

Experience, Maastricht, Universitaire Pers Maastricht, 2014, p 303.12 See Giandomenico Majone, ‘From the Positive to the Regulatory State: Causes and

Consequences of Changes in the Mode of Governance’, (1997) 17 Journal of Public Policy 2, p 165; Isabelle Muller, ‘Les Organismes Communautaires Décentralisés’, (22 janvier 1997) Les Petites Affiches 10, p 30.


during the 2000 intergovernmental conference (IGC) that the solution of a proper legal basis was genuinely discussed. The starting point for the IGC to come to a legal basis was its scrutiny of the use of the then Article 308 EC (current Article 352 TFEU). The members of the conference noted that in some areas or for some purposes Article 308 EC had been relied on rather extensively, raising the ques-tion of whether proper legal bases should not be created, allowing Article 308 EC to retain its ‘exceptional’ character. These proposals were dealt with under the broader discussion of expanding the scope of qualified majority voting, meaning the contemplated new legal bases would also move decision- making away from a unanimity requirement in the Council.

On the agencies, the presidency noted that:

Article 308 of the TEC has been resorted to here in the past in order to establish a Community body forming a separate legal entity in an area in which the Treaty did not provide any specific legal basis. Such agencies are likely to be set up more frequently in future, in view of the difficulties experienced in administering programmes and the need to improve management and supervision of such tasks.13

A few months later the presidency noted that there was some dissent as to whether new legal bases should indeed be introduced, all the while noting a willingness to examine the matter.14 The proposal was to add a new paragraph to Article 7 EC (now Article 13 TEU) reading: ‘Where this appears necessary in order to carry out any of the activities provided for in Article 3, the Council, acting in ac-cordance with the [codecision procedure], shall establish an agency having legal personality and determine the rules applicable thereto.’15 Following further dis-cussions, the idea was advanced to add a new Article 256a EC to the Treaty. The new Article would have read:

Where this appears necessary in order to carry out any of the activities provided for in Article 3, the Council, acting in accordance with the [codecision procedure], shall estab-lish an agency having legal personality and confer on it powers to implement the rules which the Council lays down, without prejudice to Article 202 TEC.

The rules governing languages applying to each of the agencies established on the basis of this Article shall be covered by Article 290.16

Evidently, neither of the two proposals made it to the final version of the Nice Treaty, but the proposed Article 256a EC in particular is rather elucidating and will be returned to later. The issue also resurfaced during the Convention on the Future of Europe. The Commission in its communication on the EU’s institu-tional architecture noted that ‘[t] he constitutional treaty should […] include a

13 Conference of the Representatives of the Governments of the Member States, Possible Extension of Qualified- majority Voting, 22 February 2000, CONFER 4711/ 00, p 2.

14 Conference of the Representatives of the Governments of the Member States, Presidency Report to the Feira European Council, 14 June 2000, CONFER 4750/ 00, p 32.

15 Conference of the Representatives of the Governments of the Member States, Extension of Qualified Majority Voting, 3 July 2000, CONFER 4753/ 00, p 18.

16 Conference of the Representatives of the Governments of the Member States, Extension of Qualified Majority Voting, 14 September 2000, CONFER 4770/ 00 ADD 1, p 40.


provision on the criteria for the establishment, running and monitoring— in pol-itical, legal and budgetary terms— of […] agencies’.17 The Convention’s Working Group on complementary competences then noted that Article 308 EC had been extensively relied on in certain areas (such as establishing agencies) and that it ‘agreed on the need to recommend new specific legal bases in the Treaty for such policy areas if the Union wished to pursue policies in these fields’.18 However, the working group’s recommendation was not picked up and the only other note-worthy proposals tabled during the Convention originated from (groups of) in-dividual members that did not make it to the final version of the Constitutional Treaty either.19

A first such amendment, tabled by a group headed by MEP Duff, related to the current Article 291 TFEU and proposed the following second paragraph:

Where uniform conditions for the implementation of the Union’s binding acts are needed, those acts may confer implementing powers on the Commission and in the cases provided for in Article (CFSP), on the Council. Specialised agencies possessing legal personality may also be entrusted with implementing tasks. Such tasks must be strictly defined and purely ex-ecutive in a technical sense.20

The Convention members argued that ‘[d] iscretionary powers must be restricted to the Union institutions themselves. Nevertheless, it would be useful to make explicit the basic principles of the ECJ’s case law on the delegation of executive powers as found in Meroni.’21

The Freiburg Draft for a European Constitution, written under the auspices of Prof Schwarze, was also circulated in the Convention. Similarly to the proposal of MEP Duff, it suggested to add a provision to the Article on implementing powers specifying that ‘[t] he Union may only create Union agencies with public powers (Agencies) if the Treaty on the Policies of the Union expressly authorizes it to do so’.22 Since the latter Treaty would have been part three (on Community policies) of the (Nice) EC Treaty, it would not readily contain such express authorizations, but the Freiburg draft equally provided that the Treaty on the Policies would be revisable following a simplified procedure.23

17 European Commission, COM (2002) 728 final, p 14.18 Secretariat of the European Convention, Final Report of Working Group V, 4 November 2002,

CONV 375/ 1/ 02 REV 1, p 15.19 Vos specifically deplored the lack of references to EU agencies in the Constitution’s Articles

dealing with implementing and delegated acts. See Ellen Vos, ‘Editorial: White and Black Smoke Coming from the Justus Lipsius Building’, (2004) 11 MJECL 3, p 230.

20 Secretariat of the European Convention, Suggestion for Amendment of: Article 28 by MEP Duff et al, http:// european- convention.eu.int/ docs/ Treaty/ pdf/ 28/ Art28Duff.pdf, p 1 (emphasis in original).

21 Ibid.22 Secretariat of the European Convention, Contribution submitted by Mr Erwin Teufel, member

of the Convention: ‘Freiburg Draft of a European Constitutional Treaty’, 20 January 2003, CONV 395/ 03, p 12.

23 This would not have been the simplified procedure as prescribed by Article 48 TEU today. Instead it would have required unanimity in the European Council and a three quarters maj-ority of the European Congress, a new body composed of members of the European and national parliaments.

http://european-convention.eu.int/docs/Treaty/pdf/28/Art28Duff.pdf

http://european-convention.eu.int/docs/Treaty/pdf/28/Art28Duff.pdf


Lastly, MEP Brok suggested the addition of an Article on agencies following the Articles on the institutions and the two advisory committees. The Article would have read:

1. In order to fulfil its tasks, the Union may, if necessary, establish Union agencies and provide them with legal personality by a Union law which also provides for judicial review of and liability for acts adopted by such bodies. By way of derogation from [the codecision procedure], adoption of such a Union law shall require the European Parliament acting by the majority of its members and the Council acting unanimously.

2. Paragraph 1 shall also apply if Parliament and Council, following a proposal from the European Commission, decide to establish an independent Anti- Trust Authority of the Union.24

Rather than discussing all these proposals in turn, both the academic and polit-ical proposals will be discussed together.

2.2 A legal basis in the Treaties

Geradin’s argument on the impossibility to grant discretionary powers, unless Article 13 TEU is amended, does not appear too convincing. Firstly, it seems to reduce Meroni to the institutional balance, an argument which was rejected in a previous chapter. Secondly, it is not clear why inscribing the agencies in Article 13 TEU would do away with a limit to empowering agencies which would other-wise exist, should a legal basis be provided elsewhere in the Treaties. It should be noted here that the Meroni doctrine is ‘only’ judge- made law. Should the Treaty Authors add a provision in the Treaties saying ‘powers’ may be granted to bodies under secondary legislation, this provision would in itself already redefine the in-stitutional balance and neither the (political) institutions, nor the Member States, nor the Court itself could directly invoke Meroni to argue that such powers could not encompass ‘discretionary powers’. Of course, the same general concern as the one underlying Meroni would remain as relevant, but this could also be addressed in a Treaty amendment.

Above it was noted that the lack of control appeared to be the thorn in the Court’s side in Meroni, which is why the emphasis was also placed on controlling agencies (cf. section V 4.2.3) even in the absence of a clear legal basis. The same problem could be noted in the proposal by the group headed by MEP Duff which proposed to inscribe Meroni in primary law, treating that jurisprudence as sacro-sanct. Similarly, in her proposed legal basis, Scholten suggests barring agencies from exercising policy discretion.25 However, it hardly needs explaining that even

24 Secretariat of the European Convention, Contribution by Mr Elmar Brok, member of the Convention: ‘The Constitution of the European Union’, 7 March 2003, CONV 325/ 2/ 02, p 76.

25 The suggestion made by Rodrigues also suffers from this since he found that the proposed amendment to Article 211 EC would still require the delegation by the Commission to respect the limits set out in Meroni. See Rodrigues, n 9, p 1181.


at the time of the Convention the actual agencification already violated Meroni, something which Scholten also seems to suggest.26 As a result, insisting on the discretionary vs. non- discretionary dichotomy misses the point and in any case does not serve the actual concern expressed in Meroni (ie that of control). The added value of introducing a legal basis in the Treaties would precisely be that it provides the possibility to modernize Meroni (or even do away with it), rather than to codify this outdated jurisprudence in primary law.

As regards the possibility to amend Article 13 TEU, the proposal from Geradin should be juxtaposed with that of Majone and the 2000 IGC. Indeed, if Article 13 TEU were to be amended it would not seem advisable to mention the agen-cies next to the institutions. Instead it should rather be provided that the institu-tions can develop the EU’s institutional architecture without identifying each single one of such bodies by name. However, whether an amendment to Article 13 TEU is required at all remains doubtful. Instead, this again seems inspired by a concern to accommodate institutional balance concerns which in reality are not addressed. The institutional balance prescribes that the institutions’ prerogatives must be safeguarded. Referring to agencies in Article 13 TEU does not ensure these prerogatives are respected. Whether or not Article 13 TEU mentions EU agencies, they would remain subsidiary bodies subordinate to the institutions. A separate clause in Parts 3 or 6 of the TFEU would then in itself alter the insti-tutional balance even without an amendment to Article 13 TEU. After all, Article 13 TEU simply provides that ‘[e] ach institution shall act within the limits of the powers conferred on it in the Treaties’. As the institutional balance results from all the provisions of the Treaties and not from Article 13 TEU, altering the bal-ance does not require an amendment to the latter Article. In addition, amending Article 13 TEU would further bloat that Article which has already been inflated from its origins in Article 4 EEC.

Rather than pursuing a solution through Article 13 TEU, the proposed Article 256a during the 2000 IGC will be returned to. That proposal is akin to Helfritz’ proposal since it would include a new Article in the current Title I (institutional provisions) of Part six (institutional and financial provisions) of the TFEU. Two elements then speak in favour of the IGC’s proposal rather than Helfritz’. Firstly, while this seems inspired by a justified concern to include control elements in the Treaty legal basis, Helfritz’ proposal appears excessively detailed (as discussed previously), containing provisions which do not really belong in a constitutional charter. Secondly, the IGC’s proposed Article 256a made an explicit reference to the implementing powers conferrable by the Council on the Commission. However, even Article 256a was not entirely clear, as it merely provided that the possibility to establish and empower EU agencies was without prejudice to Article 202 EC. It therefore referred to the Treaty provision allowing the Council to grant implementing powers to the Commission, rather than to Article 211 EC, the pro-vision at issue in Romano, allowing the Commission to be empowered by the

26 Scholten, n 11, p 299.


Council. Still, the proposal acknowledged that allowing EU agencies to exercise implementing powers conflicts, or at least sits uneasily, with Article 291 TFEU. Any legal basis for agencies in the Treaties should therefore address this issue as it cannot be bypassed, for instance, by a reference to the agencies in Article 13 TEU.

A Treaty legal basis should then expressly foresee the possibility to (i) establish and (ii) empower EU agencies, setting out (iii) how those powers relate to the Commission’s powers and at the same time ensuring control of (iv) agencies and (v) agencification. One possibility to secure these objectives is to further draw in-spiration from the current Article 291(3) TFEU, which prescribes that the legislator should adopt the Comitology Regulation. A first draft of a legal basis could then be:

1) In so far as necessary to attain the objectives of the policies of the Union, agencies possessing legal personality may be set up pursuant to the applicable legislative procedure and may be granted powers.

2) Should these powers come under Article[s 290 or] 291(2) of the Treaty, the act empowering an agency shall explain why such powers should exceptionally be granted to the agency.

3) For the purposes of this Article, the European Parliament and the Council, acting by means of a single regulation in accordance with the ordinary legislative procedure, shall lay down in advance the general principles governing the func-tioning of such agencies and the exercise of their powers.

The first paragraph makes clear that the need for an EU agency should be clearly motivated, contributing to controlling the process of agencification. While most of the proposals for a legal basis referred to the ordinary legislative procedure to est-ablish an agency, apart from MEP Brok’s proposal, preference is given to the same procedure which applies for the adoption of the material rules, in line with Mok’s argument following the ENISA case.27 Evidently, in most cases this will be the ordinary legislative procedure, and ideally the further democratization of the EU should mean that the ordinary legislative procedure becomes the sole legislative procedure. While it was noted previously that establishing agencies without giving them powers would be meaningless (see section I 1.1), paragraph 1 does make a distinction as it only requires a legislative act to establish an agency. Agencies may then be granted powers by the same legislative act, other legislative acts, or non- legislative acts, similar to what was noted in relation to German agencies (cf. section V 2.2.1). The legislator could for instance establish an agency with a slimmed- down regulation, leaving it to the Commission to work out all the neces-sary arrangements in a delegated act, thereby supplementing the basic act estab-lishing the agency. The major limit to this solution would be that the functioning (in a broad sense) of the agency could not be part of the essential elements of the basic act.28

27 See Chapter IV, n 65.28 A further limit to this technique would be that Article 263(5) provides that ‘[a] cts setting up

bodies, offices and agencies of the Union may lay down specific conditions and arrangements con-cerning actions brought by natural or legal persons’ (emphasis added).


Still, this limit could in turn be overcome pursuant to the third paragraph, which instructs the legislator to adopt a model statute for the agencies. The arrangements adopted by the Commission in its delegated act would then also be covered by the Regulation adopted pursuant to paragraph 3. That Regulation might then also spell out the situations in which the Commission, or— should this be decided as such— only the legislator itself, may deviate from the arrangements prescribed by it. This Regulation would form part of the organic or constitutional laws of the EU which rank above ordinary secondary legislation.29 Accordingly, it would contribute to con-trolling agencification and, depending on the actual provisions of the Regulation, it could (should) contribute to controlling individual agencies. However, unlike Helfritz’ proposal, it is not suggested here to include further detailed provisions on the content of such a regulation in the Treaty legal basis itself. Instead of investing time and energy in working out such details in Conventions and IGCs, it appears best to leave this to the institutions. The regulation could then contain many of the provisions of the Common Approach, making them binding, but ideally the regula-tion would address the critique on the Common Approach (cf. section II 3.6).

Paragraph 2 would then be a correction to the Court’s ruling in Short- selling, where it did not attach enough weight to the Commission’s prerogatives under Articles 290 and 291 TFEU. The paragraph should indeed be seen in this light, since a number of other provisions of the Treaties which bestow prerogatives on the Commission similarly to Articles 290 and 291 TFEU are not specifically men-tioned. Indeed, paragraph 2 should not be read (under an a contrario reasoning) as a permit to the legislator to grant other powers vested in the Commission to agen-cies but rather as a lex specialis, allowing agencies to be empowered where powers are normally exercised by the Commission. Paragraph 2 therefore confirms that empowerments under Articles 290 and 291 TFEU should remain the rule and empowerments to the agencies the reasoned exception. The duty to substan-tiate such exceptional empowerments would apply to both the legislator and the Commission, depending on which authority empowers the agency. The require-ment again strengthens control over agencification and contributes to safeguard-ing the Commission’s prerogatives. In addition, paragraph 2 could also only refer to powers which come under Article 291 TFEU. This would then be a further cor-rection to Short- selling and a codification of the Commission’s argument in that case to the effect that agencies may only adopt implementing- type acts but not delegated- type acts; hence the reference to Article 290 TFEU in square brackets.

Under the current numbering, Helfritz would propose to introduce this provision as Article 309a TFEU under a separate chapter 5 on EU agencies, while the IGC proposal would have been Article 299a TFEU. Since the main requirement would be that the legal basis is inscribed in Title I (institutional provisions) of Part 6 of the TFEU,30 both suggestions would be acceptable.

29 See Chapter IV, n 560.30 A further (theoretical) possibility is to use the simplified Treaty revision procedure for this

by amending Part three of the TFEU. However, even if the Article would not increase the EU’s competences, it could only be introduced as a genuinely horizontal provision (excluding the CFSP agencies) outside Part three of the TFEU.


However, if an IGC could come to a consensus on the nature of the EU ex-ecutive and find that the Commission is indeed the EU’s primary executive actor, the Article could also be introduced as Article 250a TFEU in the section on the Commission. From its position, the Article would make clear that the agencies come under the Commission’s natural authority, and its paragraph 3 could even be redrafted to the effect that mechanisms for control by the Commission would also have to be detailed in the regulation. Granted, this would be rather ambitious, as it goes against the idea that agencies are instru-ments grown from the Member States’ wariness of greater Commission in-volvement. Still it could again contribute to controlling (individual) agencies, although not agencification itself.

A single proposal for a legal basis has not been made here as this would depend on answering questions beyond the scope of this study. Still, a legal basis more congenial to the Commission may be juxtaposed with a legal basis more congenial to the legislator.

Such a legal basis combined with a strong framework would firmly anchor the agencies in the EU institutional framework both legally and politically. It would also give the necessary legitimacy to the agency instrument, allowing control over agencification and the agencies. From this it would follow that the institutions (including the Court) would no longer need to rely on the distinction between discretionary and non- discretionary powers. Delegations (or conferrals) to the

Article 250a TFEU Article 299a TFEU/ Article 309a TFEU

1) In so far as necessary to attain the objectives of the policies of the Union, agencies possessing legal personality may be set up pursuant to the applicable legislative procedure and may be granted powers.

1) In so far as necessary to attain the objectives of the policies of the Union, agencies possess-ing legal personality may be set up pursuant to the applicable legislative procedure and may be granted powers.

2) Should these powers come under Article 291 (2) of the Treaty, the act empowering an agency shall explain why such powers should exception-ally be granted to the agency.

2) Should these powers come under Articles 290 or 291 (2) of the Treaty, the act empower-ing an agency shall explain why such powers should exceptionally be granted to the agency.

3) For the purposes of this Article, the European Parliament and the Council, acting by means of a single regulation in accordance with the ordinary legislative procedure, shall lay down in advance the general principles governing:(a) the mechanisms for control by the

Commission of such agencies;(b) the agencies’ functioning and the exercise of

their powers.

3) For the purposes of this Article, the European Parliament and the Council, acting by means of a single regulation in accordance with the ordinary legislative procedure, shall lay down in advance the general principles governing the functioning of such agencies and the exercise of their powers.


Commission or the agencies would be equally legitimate under the Treaties (albeit that specific reasons would have to be shown to empower agencies). Ultimately the same limits could then apply to the Commission and the agencies, in the sense that the essential elements remain to be reserved to the legislator.31 It could thus allow the institutions, including the Court, to do away with the ‘essence’32 of Meroni altogether.

31 Should agencies be empowered by the Commission through genuine delegations, part of Meroni would of course still apply.

32 See Case C- 270/ 12, UK v. Parliament and Council, ECLI:EU:C:2014:18, para 41.

Selected Bibliography

A RT IC L E S

Adamski, Dariusz, ‘The ESMA Doctrine: A Constitutional Revolution and the Economics of Delegation’, (2014) 39 ELRev 6, pp 812– 34.

Bach, Tobias and Jann, Werner, ‘Des animaux dans un zoo administratif: le change-ment organisationnel et l’autonomie des agences en Allemagne’, (2010) 76 RISA 3, pp 469– 94.

Bast, Jürgen, ‘On the Grammar of EU Law: Legal Instruments’, (2003) Jean Monnet Working Papers 9, pp 1– 52.

Bernard, Elsa, ‘Accord sur les agences européennes: la montagne accouche d’une souris’, (2012) RDUE 3, pp 399– 446.

Chamon, Merijn, ‘EU Agencies: Does the Meroni Doctrine Make Sense?’, (2010) 17 MJECL 3, pp 281– 305.

Chamon, Merijn, ‘EU Agencies between “Meroni” and “Romano” or the Devil and the Deep Blue Sea’, (2011) 48 CMLRev 4, pp 1055– 75.

Chamon, Merijn, ‘The Empowerment of Agencies under the Meroni Doctrine and Article 114 TFEU: Comment on United Kingdom v Parliament and Council (Short- selling) and the Proposed Single Resolution Mechanism’, (2014) 39 ELRev 3, pp 380– 403.

Chiti, Edoardo, ‘An Important Part of the EU’s Institutional Machinery: Features, Problems and Perspectives of European Agencies’, (2009) 46 CMLRev 5, pp 1395– 442.

Chiti, Edoardo, ‘European Agencies’ Rulemaking: Powers, Procedures and Assessment’, (2013) 19 ELJ 1, pp 93– 110.

Daubler, Wolfgang, ‘Die Delegation von Rechtsetzungsbefugnissen im Recht der EWG’, (1966) Deutsches Verwaltungsblatt, pp 660– 4.

de Moor- van Vugt, Adrienne, ‘Netwerken en de europeanisering van het toezicht’, (2011) 59 SEW 3, pp 94– 102.

Egeberg, Morten and Trondal, Jarle, ‘EU- level Agencies: New Executive Centre Formation or Vehicles for National Control?’, (2011) 18 JEPP 6, pp 868– 87.

Ehlermann, Claus- Dieter, ‘Die Errichtung des Europäischen Fonds für Währungspolitische Zusammenarbeit’, (1973) 8 Europarecht 3, pp 193– 208.

Everson, Michelle, ‘Independent Agencies: Hierarchy Beaters?’, (1995) 1 ELJ 2, pp 180– 204.Geradin, Damien, ‘The Development of European Regulatory Agencies: What the EU

Should Learn from the American Experience’, (2004) 11 CJEUL 1, pp 1– 53.Geradin, Damien and Petit, Nicolas, ‘The Development of Agencies at EU and National

Levels: Conceptual Analysis and Proposals for Reform’, (2005) 23 YEL, pp 137– 97.Griller, Stefan and Orator, Andreas, ‘Everything under Control? The “Way Forward”

for European Agencies in the Footsteps of the Meroni Doctrine’, (2010) 35 ELRev 1, pp 3– 35.

Groß, Thomas, ‘Die Kooperation zwischen europäischen Agenturen und nationalen Behörden’, (2005) 39 Europarecht 1, pp 54– 68.

Groß, Thomas, ‘Unabhängige EU- Agenturen— eine Gefahr für die Demokratie?’, (2012) 67 JZ 22, pp 1087– 93.

Jacqué, Jean- Paul, ‘The Principle of Institutional Balance’, (2004) 41 CMLRev 2, pp 383– 91.

Selected Bibliography384

Kelemen, Daniel, ‘The Politics of “Eurocratic” Structure and the New European Agencies’, (2002) 25 WEP 4, pp 93– 118.

Lauwaars, Richard, ‘Auxiliary Organs and Agencies in the E.E.C.’, (1979) 16 CMLRev 3, pp 365– 87.

Lenaerts, Koen, ‘Some Reflections on the Separation of Powers in the European Community’, (1991) 28 CMLRev 1, pp 11– 35.

Lenaerts, Koen, ‘Regulating the Regulatory Process: “Delegation of Powers” in the European Community’, (1993) 18 ELRev 1, pp 23– 49.

Lenaerts, Koen, ‘EMU and the EU’s Constitutional Framework’, (2014) 39 ELRev 6, pp 753– 69.

Louis, Jean- Victor, ‘Le Fonds Européen de Cooperation Monetaire’, (1973) 9 CDE 3, pp 255– 97.

Maas, Herman, ‘Delegatie van bevoegdheden in de Europese Gemeenschappen’, (1967) 15 SEW 1, pp 2– 18.

Neergaard, Anders, ‘European Supervisory Authorities— A New Model for the Exercise of Powers in the European Union?’, (2009) EUREDIA 4, pp 603– 30.

Ohler, Christoph, ‘Rechtsetzungsbefugnisse der Europäischen Wertpapier- und Marktaufsichtsbehörde (ESMA)’, (2014) 69 JZ 5, pp 244– 52.

Pescatore, Pierre, ‘L’Executif Communautaire: Justification du Quadripartisme institue par les Traités de Paris et de Rome’, (1978) 14 CDE 4, pp 387– 406.

Thatcher, Mark, ‘The Creation of European Regulatory Agencies and Its Limits: A Comparative Analysis of European Delegation’, (2011) 18 JEPP 6, pp 790– 809.

Tridimas, Takis, ‘Community Agencies, Competition Law, and ECSB Iniatiatives on Securities Clearing and Settlement’, (2009) 28 YEL, pp 216– 307.

Vos, Ellen, ‘Reforming the European Commission: What Role to Play for EU Agencies?’, (2000) 37 CMLRev 5, pp 1113– 34.

Weiler, Joseph, ‘Does Europe Need a Constitution? Demos, Telos and the German Maastricht Decision’, (1995) 1 ELJ 3, pp 219– 58.

Yataganas, Xenophon, ‘Delegation of Regulatory Authority in the European Union: The Relevance of the American Model of Independent Agencies’, (2001) Jean Monnet Working Papers 3.

BOOK SEC T IONS

Bast, Jürgen, ‘Legal Instruments and Judicial Protection’, in A. Von Bogdandy and J. Bast (eds), Principles of European Constitutional Law, München, Hart Publishing, 2010, pp 345– 97.

Comte, Françoise, ‘Agences décentralisées: vers un statut unifié? Approche commune du Parlement européen, du Conseil de l’Union européenne et de la Commission euro-péenne sur les agences décentralisées’, in I. Govaere and D. Hanf (eds), Scrutinizing Internal and External Dimensions of European Law— Les dimensions internes et externes du droit européen à l’ épreuve, Bruxelles, Peter Lang, 2013, pp 143– 56.

de Witte, Bruno, ‘Institutional Principles: A Special Category of General Principles of EC Law’, in U. Bernitz and J. Nergelius (eds), General Principles of European Community Law, The Hague, Kluwer Law International, 2000, pp 143– 59.

Everling, Ulrich, ‘Zur Errichtung nachgeordneter Behörden der Kommission der Europäischen Wirtschaftsgemeinschaft’, in W. Hallstein and H.- J. Schlochauer (eds), Zur Integration Europas, Karlsruhe, Müller, 1965, pp 33– 49.

Selected Bibliography 385

Halberstam, Daniel, ‘The Promise of Comparative Administrative Law: A Constitutional Perspective on Independent Agencies’, in S. Rose- Ackerman and P. Lindseth (eds), Comparative Administrative Law, Cheltenham, Edward Elgar, 2010, pp 185– 204.

Lenaerts, Koen and Verhoeven, Amaryllis, ‘Institutional Balance as a Guarantee for Democracy in EU Governance’, in C. Joerges and R. Dehousse (eds), Good Governance in Europe’s Integrated Market, Oxford, OUP, 2002, pp 35– 89.

Poiares Maduro, Miguel, ‘From Constitutions to Constitutionalism: A Constitutional Approach for Global Governance’, in D. Lewis (ed), Global Governance and the Quest for Justice: Volume I: International and Regional Organisations, Oxford, Hart Publishing, 2006, pp 227– 52.

Prechal, Sacha, ‘Institutional Balance: A Fragile Principle with Uncertain Contents’, in T. Heukels, N. Blokker, and M. Brus (eds), The European Union after Amsterdam, The Hague, Kluwer Law International, 1998, pp 273– 94.

Türk, Alexander, ‘Case Law in the Area of the Implementation of EC Law’, in R. Pedler and G. Schaefer (eds), Shaping European Law and Policy: The Role of Committees and Comitology in the Political Process, Maastricht, European Institute of Public Administration, 1996, pp 167– 92.

van Ooik, Ronald, ‘The Growing Importance of Agencies in the EU: Shifting Governance and the Institutional Balance’, in D. Curtin and R. Wessel (eds), Good Governance and the European Union: Reflections on Concepts, Institutions and Substance, Antwerpen, Intersentia, 2005, pp 125– 52.

Vos, Ellen, ‘Agencies and the European Union’, in T. Zwart and L. Verhey (eds), Agencies in European and Comparative Perspective, Antwerpen, Intersentia, 2003, pp 113– 47.

Zwart, Tom, ‘La Poursuite du Père Meroni ou pourquoi Les Agences Pourraient Jouer un Rôle plus en Vue dans l’Union Européenne’, in J. Dutheil de la Rochère (ed), L’Exécution du Droit de l’Union, entre Mécanismes Communautaires et Droits Nationaux, Bruxelles, Bruylant, 2009, pp 159– 73.

BOOK S

Barbey, Gunther, Rechtsübertragung und Delegation: eine Auseinandersetzung mit der Delegationslehre Heinrich TRIEPELs, Münster, Rechts- und Staatswissenchlaftliche Fakultät der Westfälischen Wilhelms- Universität, 1962, 137 p.

Berger, Michael, Vertraglich nicht vorgesehene Einrichtungen des Gemeinschaftsrechts mit eigener Rechtspersönlichkeit, Baden- Baden, Nomos, 1999, 187 p.

Bergström, Carl Frederik, Comitology: Delegation of Powers in the European Union and the Committee System, Oxford, OUP, 2006, 391 p.

Busuioc, Madalina, European Agencies: Law and Practices of Accountability, Oxford, OUP, 2013, 307 p.

Couzinet, Jean- François (ed), Les Agences de l’Union européenne: Recherche sur les organis-mes communautaires décentralisés, Toulouse, PUSST, 2002, 306 p.

Craig, Paul, EU Administrative Law, Oxford, OUP, 2012, 777 p.Fischer- Appelt, Dorothee, Agenturen der Europäischen Gemeinschaft: eine Studie zu

Rechtsproblemen, Legitimation und Kontrolle europäischer Agenturen mit interdiszi-plinären und rechtsvergleichenden Bezügen, Berlin, Duncker & Humblot, 1999, 609 p.

Groenleer, Martijn, The Autonomy of European Agencies: A Comparative Study of Institutional Development, Delft, Eburon, 2009, 427 p.

Selected Bibliography386

Hilf, Meinhard, Die Organisationsstruktur der Europäischen Gemeinschaften, Berlin, Springer Verlag, 1982, 442 p.

Hofmann, Herwig, Rowe, Gerard, and Türk, Alexander, Administrative Law and Policy of the European Union, Oxford, OUP, 2011, 977 p.

Kapteyn, Paul, VerLoren Van Themaat, Pieter, and Gormley, Laurence, Introduction to the Law of the European Communities: From Maastricht to Amsterdam, London, Kluwer Law International, 1998, 1447 p.

Lenaerts, Koen, Arts, Dirk, and Maselis, Ignace, Procedural Law of the European Union, London, Sweet & Maxwell, 2006, 790 p.

Maletić, Isidora, The Law and Policy of Harmonisation in Europe’s Internal Market, Cheltenham, Edward Elgar, 2013, 214 p.

Rosas, Allan and Armati, Lorna, EU Constitutional Law, Oxford, Hart Publishing, 2010, 260 p.

Schütze, Robert, From Dual to Cooperative Federalism: The Changing Structure of European Law, Oxford, OUP, 2009, 391 p.

Schütze, Robert, European Constitutional Law, Cambridge, CUP, 2012, 484 p.Triantafyllou, Dimitris, Les compétences d’attribution au domaine de la loi: étude sur les

fondements juridiques de l’action administrative en droit communautaire, Bruxelles, Bruylant, 1997, 432 p.

Tridimas, Takis, The General Principles of EU Law, Oxford, OUP, 2006, 591 p.

U N PU BL ISH E D Ph D T H E SE S

Desomer, Marlies, The Reform of Legal Instruments of the European Union, Leuven, KUL, 2009, PhD Thesis, 629 p.

Driessen, Bart, Interinstitutional Convention as Checks and Balances in EU Law, Leuven, KUL, 2006, PhD Thesis, 445 p.

Gautier, Yves, La délégation en droit communautaire, Strasbourg, Strasbourg 3, 1995, PhD Thesis, 538 p.

Le Bot, Fabien, Le principe de l’ équilibre institutionnel en droit de l’union européenne, Paris, Université Panthéon- Assas, 2012, PhD Thesis, 586 p.

Note: to keep the index surveyable, references to pages 52–100 (The EU Agencies’ Statutes) are limited

Access to documents 64, 100, 211, 239, 341ACER 17, 25, 30, 31, 36, 37– 9, 52, 63,

104– 5, 122, 142, 148, 156, 168, 170– 1, 196, 208– 9, 211, 307, 337– 45

AdministrationDirect 49– 50, 111– 12, 121, 122, 131, 132,

153, 159– 61, 174, 307, 318, 319, 324, 327, 369

EU 46– 51, 61, 106, 116– 17, 119, 157, 162, 234, 291, 296, 297, 298, 307, 327, 336, 354

Indirect 49– 50, 111– 12, 121, 153, 155, 159– 61, 307, 324, 369

National 18, 29, 40, 47, 50– 1, 105, 108, 111, 112, 115– 16, 118, 119, 121, 159, 167, 183, 195, 202, 313

Shared 50– 1Administrative integration 50– 1, 109, 115,

159, 369Agencies

National 3– 4, 45, 61, 112, 116– 17, 163, 306– 15, 321, 360

Self- financed (EU) 97, 130, 319, 325US 6– 7, 161, 210, 266, 299– 306, 326

Amsterdam Treaty 103, 116, 135, 272, 362, 373

Anti- commandeering 49, 160Autonomy

of Agencies 5, 10, 13, 43, 82, 90, 92, 169, 172, 325– 6, 334

National 23, 114, 115, 118, 119, 132, 158, 159, 161, 313, 369

of EU legal order 217, 239, 241, 277

Banking Union 17, 67, 210, 346Barbey G. 232– 6, 238– 40, 292Berec Office 17, 22, 142Board of Appeal 35, 335, 338– 46, 349– 52,

361, 368, 371Budget 18, 45– 6, 64– 5, 83, 100, 112, 120,

122, 126– 7, 129, 130, 132, 156, 172, 317, 319– 21, 325

Bureaucracy 4, 48

Cartel Office 28, 30, 205CdT 16, 22, 121– 2, 129Cedefop 12, 13, 15, 16, 22, 23, 326, 335,

336, 349Centralization (centralize) 41, 47, 51, 108,

110, 115, 140, 160– 2, 263

CEPOL 11, 14, 16, 22, 41, 172, 358Charter of Fundamental Rights 173, 349, 352Chemicals Bureau 104Chiti E. 20– 2, 50, 109, 175, 196, 353Comitology 25, 32, 36, 37, 42, 47, 51, 61,

102– 3, 110, 118– 20, 122, 126, 130, 131– 2, 154, 159, 184, 194, 218, 222, 237, 239, 247, 252, 256, 264, 266, 268, 275, 287, 292, 319, 332, 349, 378

Common Agricultural Policy 41, 118, 126, 155, 159, 170, 221, 234– 5, 275

Common Approach on Decentralised Agencies 7, 20, 52, 61– 3, 110, 121, 123, 125, 129, 168, 316, 319, 320, 322, 345– 6, 370, 371, 372, 379

CompetenceExclusive 29, 42, 158– 9, 164, 165, 167, 238Implied see Implied powersShared 29

Composite procedure 33, 44, 160, 167, 365Constantinesco V. 231, 271, 273, 279Convention on the Future of Europe 129– 30,

239, 291, 347– 8, 352, 363, 368, 375– 6Corine programme 103, 151Court of Auditors 15, 45, 84, 85, 95, 98, 135,

149, 172, 217– 18, 317, 334, 362, 372CPVO 16, 22, 23, 28– 9, 37, 46, 128, 151,

206, 255, 335, 337– 44Craig P. 19, 61, 188, 237Customs 40, 41, 159, 256

Dashwood A. 138, 141Decentralization (decentralize) 7, 41, 50,

109– 10, 126Dehousse R. 45, 113, 192, 193, 195, 213Delegated acts – Article 290 TFEU 25, 27, 29,

33, 34, 36– 7, 39, 54, 124– 5, 131, 194– 5, 203, 234– 6, 238, 241, 252, 257, 258, 265, 286, 289– 96, 311, 332, 378– 80

Democracy 240, 259– 60, 264– 5, 267, 269– 70, 282, 285, 292– 3, 311– 12, 315

Deficit 53– 60, 110, 316de Witte B. 141– 2, 154, 269, 284, 286Discharge (budget) 86, 96, 97, 130, 132, 319,

321, 325

EAR 11, 18, 22, 329– 30, 333Early warning system 75, 125, 320EASA 16, 22, 25, 26, 30– 1, 33, 35, 37, 38, 44,

140, 255, 337, 338– 44, 351, 360

Index

Index388

EASO 17, 23, 41, 42, 105EBA (see also ESA) 17, 24, 31, 34, 37, 307ECB 13, 17, 49, 94, 167, 204, 221– 2, 231, 321,

324, 328, 337, 345, 346, 355, 357, 362, 364, 365, 371

ECDC 16, 23, 121, 140, 335, 349ECHA 17, 22, 24, 32– 3, 104, 142, 148, 153,

196, 206, 337, 338– 46, 349– 52EDA 12, 16, 23, 65, 363EEA 16, 22, 25, 103, 118, 127– 8, 140, 151,

319, 335– 6EEAS 14– 15EIB 8– 9, 13, 94, 231, 328– 9, 355– 7, 371EIGE 17, 23, 121, 162, 164, 169– 71EIOPA (see also ESA) 17, 24, 147EIT 17, 22, 23, 337EFCA 17, 23, 41, 42Effet utile 50, 228, 286EFSA 16, 22, 25, 31– 5, 38, 43, 102, 107, 128,

140, 142– 3, 147, 151, 152, 307, 319, 331– 4, 337, 353– 4, 360

EFSF 11– 12, 118EMA 16, 22, 23, 25, 26, 31– 3, 34– 5, 38,

43, 102, 107, 128, 140, 142– 3, 147, 151– 2, 155, 307, 319, 331– 5, 337, 353, 354, 360

EMCDDA 16, 22, 23, 122, 128, 162– 3, 234, 322, 325, 337

EMCF 17– 18, 94, 155, 185– 6, 346EMSA 16, 23, 33, 43– 4, 110, 121, 122,

320, 325Energy package 104– 5, 208ENISA 10– 11, 18, 142– 6, 153, 166,

211, 307Enlargement 50, 68, 75, 122, 320ERA 16, 19, 22, 23, 34, 36, 121, 122, 307,

325, 338, 360ESA 25, 30– 1, 36– 8, 39, 43, 45, 52, 102– 3,

110, 112, 122, 124, 142, 148, 153, 166, 167, 196, 209, 243, 255, 287, 307, 322, 337– 46, 360, 361, 365

ESDC 14ESMA (see also ESA) 17, 27– 8, 38– 9, 124,

131, 146– 9, 206, 209, 241– 9, 256– 7, 289, 291, 294– 6, 312, 317, 365

ETF 16, 22, 122EUISS 16eu- LISA 17, 22, 168– 71EU- OSHA 16, 22, 23, 104, 118, 122, 128,

335– 6, 349Eurofound 13, 15, 16, 22, 23, 326,

335– 6, 349Eurojust 12, 16, 23, 40, 42, 68, 103, 159, 330,

358, 363European Free Trade Area 88– 9Europol 11, 12, 17, 23, 27, 39, 40, 42, 169,

318, 358, 363EUSC 16European Administrative Space 50, 369

Europeanization 109, 115, 195– 7, 240Everling U. 137, 184, 185Everson M. 5, 59, 190, 262Executive

Agencies (EU) 5, 8, 10, 11, 22, 156– 7, 172, 234, 239

core executive 81, 241, 321, 327Federalism – see Federalismunitary executive 266, 304, 320, 336

Falcone 41Federalism

Co- operative 158, 161Executive (see also Vollzugsföderalismus)

48– 50, 54, 266Fees 37– 8, 195, 319, 342Fiscal Compact Treaty 243Fischer- Appelt D. 9– 10, 136, 162, 272,

334, 338FRA 17, 22, 23, 30– 1, 152, 162– 4,

169– 71, 318Frontex 17, 23, 42– 3, 103– 4, 169

Gautier Y. 216, 231– 9Geradin D. and Petit N. 19, 20– 2, 24– 5, 152,

162, 163, 270Griller S. and Orator A. 9, 19, 24– 5, 27, 190,

203, 204, 221, 254GSA 16, 22, 46Guidelines see soft law

Habermas J. 48, 54, 60Halberstam D. 107, 160, 311Hartley T. 140, 215– 16Heller H. 55– 6Hilf M. 12, 15, 139, 186, 193, 227, 270, 284

Impact assessment 43, 163, 168– 71Implementing acts – Article 291 TFEU 29,

33, 34, 35– 7, 39, 40, 44, 48, 49, 124– 5, 130– 2, 194, 195, 196, 197, 209– 10, 235– 8, 241, 252, 257– 8, 265, 286, 289– 96, 307, 318, 339, 353, 373, 375, 378

Implied powers 140, 141, 144, 168Infringement procedure - Article 258

TFEU 30– 1, 44, 123, 229, 275Institutional balance 23, 28, 62, 68, 188– 91,

194, 196, 197, 200– 7, 209, 211, 221, 226, 229– 30, 240, 243, 248, 249– 298, 318– 22, 329, 346, 362– 3, 365, 370, 373, 376– 7

Inter- institutional agreement 19, 62– 4, 79, 101, 121, 207, 321, 324, 347

Internal market (single market) 23– 4, 30, 50, 118, 141, 144, 147, 149, 291, 332, 372

International agreements 27, 39, 69, 87– 97, 100

Ipsen H. P. 59, 154, 262– 3

Index 389

Joint Investigation Teams 42Joint Research Centre 104Joint Undertakings 10, 23

Kirchhof P. 56– 7

Lamfalussy 102Lauwaars R. 139, 184, 186, 215, 234– 5Le Bot F. 275, 278, 282, 284, 285Legal basis 49, 53, 63– 4, 118, 119– 21, 128,

134, 138, 139– 56, 174, 184, 233, 236, 241, 247, 278, 282, 286, 295, 296, 297, 316, 341, 371– 82

LegitimacyAgencies’ 61, 110, 120, 141, 203, 228,

249, 266, 293, 304, 316, 321, 364, 369– 71, 380– 1

EU 52– 64, 165, 325, 370, 372Input 59, 147Output 59, 61, 110

Lenaerts K. 48, 186, 189, 191, 254, 263, 264, 266, 269, 270, 272– 3, 284– 5, 295, 328– 9, 372

Lisbon Treaty 13, 47, 48, 49, 54, 63, 64, 68, 98, 100, 104, 116, 134– 5, 139, 152, 158, 164, 169, 195, 204, 209, 231, 234, 236, 237, 240, 241, 252, 253, 256, 257, 258, 264– 5, 272, 286, 290, 317, 320, 327– 30, 333, 337, 345, 348, 352, 354, 357– 8, 361, 362, 367– 8, 371

Logic of appropriateness 110, 132Louis J.- V. 94, 185– 6, 346

Maastricht Treaty 57, 103, 127, 138, 151, 164, 362, 364

Maastricht Urteil 55– 6Majone G. 23, 25, 59, 190, 201, 372, 377Mutual recognition 31, 32

Neergaard A. 142, 147, 184Nettesheim M. 137Networks 4, 20, 50, 51, 96, 104– 5, 106, 108,

109, 111– 13, 115– 16, 155, 201– 2, 307New Institutionalism

Historical 114, 132Rational Choice 111, 114, 132

New Public Management 4, 48Nice Treaty 357, 362, 374, 375

Oisin 41OHIM 16, 22, 23, 28– 9, 35, 37, 46, 152,

173, 205, 209– 10, 240, 244, 255, 276, 335– 6, 338– 44

Ohler C. 131, 145, 191, 245, 295, 312

OLAF 14– 15, 219Ombudsman 173, 274, 317, 329, 339, 355Operational cooperation 23, 26, 33, 40– 4,

91, 103– 5

Path dependency 114, 132Pescatore P. 138, 183, 185, 263Prechal S. 258, 269, 286Proportionality 28, 29, 36, 68, 136, 158,

164– 74, 227, 272, 279, 326, 333– 4, 351, 370

Railway (package) 34, 99Regulation 6, 19, 23, 35ReNEUAL Model Rules on EU Administrative

Procedure 33

Scharpf F. 59Schimmelfennig F. 57, 60Schmitt C. 56, 57Schütze R. 158, 160, 189, 239, 255Separation of powers 7, 23, 47, 229, 240,

258– 69, 280, 282, 296, 302, 304– 6, 310– 11, 315

Shapiro M. 4, 107– 9, 305, 306Single European Act 50, 53, 61, 63, 108, 118,

120, 127, 151, 160, 209, 237, 266, 286, 288, 369

Soft law 29, 34– 5, 352– 4Spillover 108, 163Spitzenkandidaten 54SRB 17, 24, 142, 166, 210, 323,

338– 43, 345– 6Subsidiarity 28, 29, 47, 49, 134, 136, 157– 64,

165– 7, 169, 174– 5, 195, 264, 272, 286, 290, 297, 310, 318, 370

Sunset clause 10, 98

Thatcher M. 107, 111– 2, 188Triepel H. 231– 3, 236, 239

Uniformity 40– 1, 49, 50, 155, 236, 253, 289, 295, 321, 354, 369

Union authorization 25, 31– 3, 43, 143, 148, 162, 331– 2, 366

Variable geometry 67, 73Vollzugsföderalismus 48– 50, 160, 238,

266, 306– 7Vetter R. 131, 136, 138– 9, 144, 145, 153Vos E. 5, 20– 22, 142, 190, 191, 200– 2, 272,

277, 372, 375

White Paper on Governance 5, 60, 62, 64, 168